Applied Wetlands Science - Chapter 4 pot

Hobson Choosing Biological Endpoints Spatial and Temporal ConsiderationsOther Considerations Analysis Phase Exposure CharacterizationEcological Effects CharacterizationRisk Characterizat

Kent, David J et al “Ecological Risk Assessment of Wetlands”

Applied Wetlands Science and Technology

Editor Donald M Kent

Boca Raton: CRC Press LLC,2001

CHAPTER 4 Ecological Risk Assessment of Wetlands

David J Kent, Kenneth D Jenkins, and James F Hobson

Choosing Biological Endpoints

Spatial and Temporal ConsiderationsOther Considerations

Analysis Phase

Exposure CharacterizationEcological Effects CharacterizationRisk Characterization Phase

Predictive Ecological Assessments

Problem FormulationExposure CharacterizationEcological Effects CharacterizationRisk Characterization

The Quotient Method for Risk CharacterizationRetrospective Ecological Assessments

Problem FormulationExposure CharacterizationEcological Effects CharacterizationRisk Characterization

SummaryReferences

Until recently, the term risk assessment generally was applied to the estimate ofrisk to human health, typically from chemical exposure For example, a cancer riskassessment is an estimate of the risk to humans from carcinogenic compounds.Recently, however, the term risk assessment has been applied to ecological systems.

An ecological risk assessment is an estimate of the adverse effect to an ecosystemfrom chemical, physical, or biological stressors resulting from anthropogenic activ-ity This recent interest in assessing ecological health is evidenced by several pub-lications (Bartell et al., 1992; Cairns et al., 1992; Suter, 1993; Newman and Strojan,1998; Lewis et al., 1999) including two documents produced by the U.S Environ-mental Protection Agency (USEPA, 1992, 1998) The first of these USEPA docu-ments, Framework for Ecological Risk Assessment (1992), was intended as the firststep in a long-term program to develop guidelines for the performance of ecologicalrisk assessments The second document, Guidelines for the Ecological Risk Assess- ment, provided more detailed information and is the current guidance on the subject.The principles of ecological risk assessment can be applied to any ecosystem,although they may be particularly relevant to wetlands The extent and rate of wetlandloss, as well as the biologic, economic, and social importance of wetlands, are welldocumented (Mitsch and Gosselink, 1986) Moreover, the transitional nature ofwetlands may make them especially sensitive to stress Despite the uniform appli-cation of assessment principles to ecological systems, individual wetlands are suf-ficiently different in their spatial, temporal, and physiochemical characters to warrantsite-specific sampling and analysis (Figures 1 and 2) These differences will influencethe design and interpretation of the ecological risk assessment

Figure 1 Risk assessment can be broadly applied to a variety of ecosystems Nevertheless,

individual wetlands are sufficiently different in their spatial, temporal, and ochemical characters to warrant site-specific samping and analysis A New England saltmarsh is shown here.

physic-To understand the use of ecological risk assessment for wetland ecosystems, anintroduction to the principles of risk assessment is necessary The current basis forrisk assessment is derived from the National Research Council Risk Assessmentparadigm (1983) Following this, an in-depth discussion of the USEPA EcologicalRisk Assessment Framework (USEPA, 1992) is presented as is a discussion of thechallenges and strategies associated with carrying out assessments Finally, examples

of specific applications to wetland ecosystems are provided

THE HUMAN HEALTH RISK ASSESSMENT PARADIGM

The risk assessment paradigm has been used for some time to evaluate the chronicimpacts of environmental pollutants on human health This strategy was initiallyconceptualized by the National Research Council Risk Assessment Panel (NRC,1983) and formalized by the USEPA in its 1986 Guidelines (USEPA, 1986a–e) Thisrisk assessment paradigm consists of several components:

Figure 2 An Arkansas riverine system wetland.

• Hazard identification: does a chemical contaminant represent a specific threat to human health? Establishment of cause–effect relationships is central to this com- ponent.

• Defining dose–response: what is the relationship between the magnitude of the exposure and the probability of an adverse health effect?

• Exposure assessment: what is the potential for human exposure to the chemical of concern?

• Risk characterization: what is the potential magnitude of risk to human health given the predicted exposure and dose–response data? What is the uncertainty associated with this risk estimate?

Standard methodologies are employed to evaluate potential threats to humanhealth The methodologies usually involve determining all relevant effects and thensumming those effects to get a total effect value

ECOLOGICAL RISK ASSESSMENT

Ecological risk assessments (ERA) examine the probability that undesirableecological effects are occurring or may occur as a result of exposure to a stressor

or a combination of stressors The term stressors is used here to reflect the broadrange of anthropogenic factors that can result in ecological perturbations Stressorsmay include any chemical, physical, or biological factor resulting from anthropo-genic activities that can cause an ecological disturbance Most often, however, theterm stressor refers to toxic chemicals

ERAs can be used to address a wide range of issues and are generally classified

as predictive or retrospective Predictive ERAs are designed to assess the risksassociated with proposed actions, such as the introduction of new chemicals intothe environment and the establishment of new sources of stressors or hypotheticalaccidents (USEPA, 1991) Predictive risk assessments have usually followed theNRC human health paradigm relatively closely while emphasizing the choice ofbiological endpoints and related stressor–response data

In contrast, retrospective ERAs address the risks associated with stressors releaseddue to current or previous anthropogenic activities Examples of retrospective assess-ments include evaluating the ecological impact of hazardous waste sites and previousreleases or spills The goal of this type of risk assessment is to establish and definethe relationship between the pollution source, the distribution of stressors, the expo-sure of biological endpoints, and the level of effects of this exposure on the ecosystem.Retrospective assessments often take advantage of field data to define contaminantsources and measure adverse biological effect Various levels of data collection orsite-specific assessment may be necessary to provide the information required todesign and conduct the retrospective ecological risk assessment, and to achieve agiven level of confidence The challenge here is to establish cause–effect relationshipsbetween the source of stressors and any observed ecological effects

Some ERAs, such as those used for wetlands, may involve both predictive andretrospective aspects For example, in an assessment of a hazardous waste site, the

current status of the site may require a retrospective evaluation, but the long-termimpact of various repetition scenarios would be addressed in a predictive fashion.

As the type of risk assessment to be conducted is dependent upon the ultimateapplication of the results, a clear understanding of the objectives of the risk assess-ment is essential

The move to ERAs by the regulatory community is driven by a number of factors.From a legal standpoint, many of the underlying statutes require some form ofevaluation of ecological risk For example, the Superfund Amendment and Reautho-rization Act (1987) specifies that the actual and potential risk to public health andthe environment must be assessed for each hazardous waste site The use of a basicrisk assessment paradigm would provide a structural framework for ecologicalassessment and a consistent strategy for managing various types of risk This issue

is particularly important when comparing the sensitivity of ecological endpointsrelative to the human endpoint In some cases, nonhuman endpoints may prove to

be more sensitive than human endpoints, and would, thus, drive the overall riskassessment This type of comparison is facilitated by consistency in the strategiesused to carry out risk assessments for both types of endpoints

Although human health risk assessment strategies provide a useful model, ing ecological risk has proven to be more complex A number of factors contribute

assess-to this complexity

• Multiple biological endpoints: these could include multiple species, and various levels of biological organization (e.g., subcellular, individual, population, commu- nity, and ecosystem).

• More complex exposure pathways: these are determined by the biological endpoints

In spite of these complexities, there are some distinct advantages to estimatingrisk to ecological endpoints Exposures and hazards can often be estimated directly

on the species of concern or a closely related surrogate species This may result in

a more accurate estimate of risk, because there is no need for extrapolation frommore distantly related species as is almost always the case in human risk assessment.This is particularly useful when multiple stressors or complex exposure matrices areinvolved In ERAs, the biological endpoints of interest can often be tested directlyagainst the specific stressor or mixture of stressors of concern, thus eliminating theneed to estimate such factors as stressor interactions, chemical form (speciation),and bioavailability

Despite the aforementioned advantages, the added complexity associated withERAs results in a higher degree of uncertainty than is normally associated withhuman health-based risk assessments This complexity requires more effort in theinitial planning stages so that the final assessment is well focused Furthermore,

some modifications of the basic NRC risk assessment paradigm are required TheRisk Assessment Forum within the USEPA has developed the framework (and nowguidelines) for ecological risk assessment which addresses these issues, yet is con-ceptually consistent with human health risk assessment strategies It is important tonote that slight differences in terminology exist between the various ERA structurescurrently promoted However, the elements are fairly analogous For the sake ofsimplicity, the discussions in this chapter will utilize the USEPA framework termi-nology, although the concepts may be drawn from multiple sources.

The Ecological Risk Assessment Framework

Two major elements that form the basis of the ERA framework are the terization of exposure and the characterization of ecological effects Aspects of thetwo elements are considered in all phases of the framework process While carryingthis common thread throughout the paradigm, the framework is divided into threephases The phases are problem evaluation, analysis, and risk characterization(Figure 3)

charac-Problem Formulation Phase

The first step in this process requires defining the specific purpose of the ERA.Although this may seem trivial, many ERAs suffer from lack of clear focus and, as

a result, may be ambiguous and misleading Once the specific purpose of theassessment is defined, the specific goals that must be met to achieve this purposeare formulated These goals provide a basis for establishing a precise conceptualstudy design In undertaking the study design, a number of questions must first beaddressed

• Is the ERA to be predictive or retrospective?

• Is the ERA to be site-specific or generic?

• What type of ecosystem(s) is at risk?

• What types of stressors are involved?

• What are the potential source(s) for a given stressor or set of stressors?

Addressing these questions requires a rigorous review of available data Whereexisting data are unavailable or incomplete, it may be necessary to carry out apreliminary study, particularly to establish the types of potential stressor Mostimportant, ecological risk assessment is often an iterative process A given level ofinformation is required for developing the design and objectives (i.e., the problemformulation) Additional data may be required for the complete ERA Ultimately,the level of information needed and the extent of new data collection are dependentupon the objectives, such as the level of certainty desired As well, the level ofinformation is dependent upon the outcome of the previous iterations, for example,how much ecological impact has occurred at a given site or how hazardous a newchemical may be based on laboratory toxicity studies

Choosing Biological Endpoints

An important issue at this stage is determining the appropriate suite of biologicalendpoints to be used in the evaluation Biological endpoints should be carefullychosen specifically to address the overall goals of the assessment The parameters

to be considered in choosing these endpoints should include the ecological relevance

of the endpoint and the spatial and temporal occurrence of the endpoint relative tothe distribution of the stressors and potential biological receptors

Figure 3 From the Framework for Ecological Risk Assessment (USEPA, 1992)

In ecological assessments the distinction is often made between assessmentendpoints and measurement endpoints (Warren-Hicks et al., 1989; Suter, 1993).Assessment endpoints represent the ultimate resource(s) or final environmental val-ues that are to be protected They should have social or biological relevance, bequantifiable, and provide useful information for resource management or regulatorydecisions For example, successful reproduction of a species in danger of extinction

is an appropriate assessment endpoint For populations that are valued for cial or sport uses such as anadromous fish (e.g., salmon) or estuarine or marineshrimp, the important assessment endpoints could be growth, reproduction, or overallproductivity Other potential assessment endpoints include yield and productivity,market or sport value, recreational quality, and reproductive capability (see Warren-Hicks et al., 1989) In general, assessment endpoints focus on population and com-munity parameters because ecological risk assessments are usually concerned withprotecting these higher levels of biological organization rather than individual organ-isms However, when there is concern for endangered species, the assessment end-points may indeed focus on the individual organism

commer-In contrast, measurement endpoints represent the specific parameters that are to

be measured in a given assessment They are often chosen based upon practicalconsiderations such as availability and ease of measurement It is often impossiblefrom a practical sense to measure some endpoints (e.g., endangered species) andother more obtainable measurement endpoints are chosen as surrogates for the actualassessment endpoints of interest Such measurement endpoints should be well char-acterized and take into account exposure pathways and temporal factors Ideally,measurement endpoints should be chosen so that the data from these endpoints can

be linked directly or indirectly to appropriate assessment endpoints This latter issue

is often the most difficult to address in ERAs

At the level of the individual, measurement endpoints may include mortality,growth, and fecundity Abundance and reproductive performance are measurementendpoints on a population level Other measurement endpoints include species even-ness and diversity on a community level and biomass and productivity on an eco-system level Assessment of endangered species poses a special problem Becauseexposure for an endangered species cannot usually be directly assessed, the residues

of a contaminant in a principal food item can be a measurement endpoint tively, if exposure to a chemical and its potential impact on an endangered species

Alterna-is the assessment endpoint, then a co-exAlterna-isting species with similar life hAlterna-istory habitsmight be used as a surrogate measurement endpoint

Spatial and Temporal Considerations

Ultimately, the problem formulation phase should result in the establishment ofthe study design which will form the basis of the assessment The design must takeinto account a number of environmental and ecological factors that may affect thestressors or their potential impact on biological systems Many of these factors havespatial or temporal components that must be taken into account in the design Spatial

factors include potential routes of exposure, other sources of stressors, location ofsensitive biological resources, and factors that may modify contaminant mobility oravailability The changing composition of sediments is an example of the latter.Temporal factors may include seasonal changes in physical, chemical, or biologicalaspects of ecosystems that may influence the magnitude or form of the stressor orits potential to cause a biological effect For example, increased surface watermovement in the wet season in riparian habitats can dramatically affect contaminantmigration and the potential for exposure Also, seasonal variation in physical orchemical parameters such as temperature or pH can modify the bioavailability ofcontaminants, thus changing the nature of the exposure.

The biological characteristics of the exposure matrix may also change temporally.Species may be present only seasonally as is typical of migratory waterfowl, anadro-mous fishes, or species that migrate seasonally within a local area Exposure to agiven species, population, or community may vary with seasonal changes in lifehistory habits such as seasonal feeding patterns or reproductive cycles Thus, tem-poral changes in the biological components of an ecosystem may influence thedistribution of the stressor or the availability of the biological endpoint within theecosystem Temporal variation would affect the ultimate risk assessment Therefore,these temporal patterns must be addressed in designing the sampling scheme anddata collection for the risk assessment

Temporal considerations may also include long-term historical or predictedtrends in stressor influence and potential seasonal variation in stressor impact.Historical or predicted trends are very important to understanding the overall impact

of stress on an ecosystem and may be important to the application of risk information

in risk management decisions such as remediation plans, wetland restoration, andregistration of a new pesticide

Other Considerations

Another factor to consider is the presence of biological resources that are eithersensitive to stressor impact or may be of particular economic or social importance.Wetlands are nursery areas for many species that inhabit adjacent terrestrial andaquatic communities Other resources may include populations of economicallyimportant species such as anadromous fish populations or endangered species whichare protected by law

In completing the problem foundation phase of the ERA, these issues must beaddressed in rigorous and systematic fashion The ultimate goal is the development

of a conceptual model that will serve as a basis for the ERA The model shouldaddress the spatial and temporal distributions of potential stressors, appropriatebiological endpoints, as well as probable routes and levels of exposure Finally, themodel should be precisely linked to the goals and purposes of the risk assessment

A well-conceived conceptual model is essential to the effective implementation ofthe subsequent component of the risk assessment

Both the stressor and the ecosystem must be characterized with regard to thedistribution and pattern of change This is accomplished through the use of modeling

or monitoring data, or both, depending on whether the ERA is predictive or spective Stressors can be characterized from direct sampling, laboratory or fieldtesting, or through remote sensing At this stage, it is important to evaluate the means

retro-by which stressors can be modified in the ecosystem through biotransformation orother environmental fate processes such as photolysis, hydrolysis, and sorption.Transport and fate models are often employed, including interactive mass balancemodels like the Exposure Analysis Modeling System (EXAMS) developed at theUSEPA research laboratory in Athens, GA (Burns et al., 1982) Characteristics ofthe ecosystem that can affect exposure also need to be evaluated These can includehabitat requirements, food preferences, reproductive cycles, and seasonally influ-enced activities

After stressor and ecosystem characteristics are defined, the spatial and temporaldistributions are combined to evaluate exposure Finally, the magnitude and distri-bution patterns are quantified for the scenarios developed during the problem for-mulation phase of the ERA All sources of uncertainty should also be quantified tothe degree possible for the input into the ERA Recently, probabilistic techniqueshave been used to characterize uncertainty under various exposure conditions (Bur-master and Anderson, 1994; Power, 1996)

Ecological Effects Characterization

The goal of this portion of the analysis phase is to identify and characterize anyadverse ecological effects that are associated with exposure to a particular stressor

or stressors To the extent possible, these effects should be quantified and any causeand effect relationships evaluated

Initially, an evaluation must be made of all effect data that are relevant to thestressor The types of data that are relevant will depend upon the characteristics ofboth the stressor and the ecological component and also on whether the ERA is to

be predictive or retrospective Commonly used data types include aquatic toxicitytests, computer models, quantitative structure activity relationships (QSARs), micro-and mesocosms, species diversity analyses, artificial substrate comparisons, the

Wetland Evaluation Technique (Adamus et al., 1987), and others (Suter, 1993;USEPA, 1992; McKim et al., 1987; Barnthouse et al., 1986) Data from laboratoryand field observations, as well as from controlled studies, may be used These dataare considered based upon their relevance to the measurement and assessmentendpoints Evaluating the quality of the data, that is, the adequacy of sampling andstatistical design, is an important component of this stage.

The next step is to quantify the ecological response in terms of the sor–response relationship The aim is to describe the relationship between the mag-nitude, frequency, and duration of the stressor, and the magnitude of the response.The dose–response curve in laboratory aquatic toxicity tests is an example of thistype of analysis Determining relative differences in productivity, biomass, speciescomposition, and diversity between contaminated and reference sites is also a com-monly used means of defining this relationship Any extrapolations required betweenthe stressor and measurement endpoints must be evaluated as well as any extrapo-lations between measurement and assessment endpoints The strength of the causalrelationship must be quantified to the extent possible Finally, as with the exposureanalyses, quantitative estimates of uncertainty such as natural variability in ecolog-ical characteristics and responses should be included in the analyses Often thisanalysis of uncertainty can be done probabilistically (Moore, 1996) However, itshould be emphasized that although the goal is to provide quantification of thestressor–response relationship, in many cases, it can only be described qualitativelyand, therefore, professional judgment plays an important role in most ERAs

stres-Risk Characterization Phase

The ultimate purpose of the ERA is to estimate the risk that unacceptable adverseecological effects will occur due to exposure to anthropogenic stressors The riskcharacterization phase provides an estimate of the likelihood that an adverse impacthas occurred or will occur It also should address the relative ecological consequencesthat are associated with the various levels of risk Ideally, risk characterization should

be quantitative and include estimates of uncertainty However, the complexity ofecological risk assessments may often preclude precise quantification Under thesecircumstances, qualitative estimates of risk are often employed

In the risk characterization phase a hazard-exposure matrix is developed Thismatrix represents a fusion of the exposure predictions developed in the exposurecharacterization phase and the estimates of effects developed in the ecological effectscharacterization phase

Ultimately, the risk characterization phase should evaluate the ecological sequences or ecological significance of predicted or observed effects This is sim-pler in some cases for retrospective risk assessments, because the changes can bedocumented and quantified in situ Also, the ecological effects, that is, the resultingchanges at higher levels of biological organization or in the ecosystem as a wholecan be monitored directly with the proper sampling design and appropriate refer-ence sites In predictive ecological risk assessments, these effects or changes inecosystems must be predicted The degree of uncertainty for such predictions isrelatively large in most situations As the historical database for ecotoxicology and

con-ecological risk assessment increases, especially field data sets, these predictionswill be made with higher levels of confidence Especially important is the fieldvalidation of risk assessments based on laboratory-based hazard assessments andmodeled exposure assessments.

As a result of the inherently different goals of predictive and retrospectiveecological risk assessments, the remainder of this chapter will provide illustrations

of how ERAs may be applied for each of these types with regard to wetlandecosystems The first type of ERA to be discussed will be the current tiered assess-ment design, a predictive approach employed by the USEPA in the pesticide regu-lation program The second describes a retrospective approach more common in theassessment of hazardous waste sites or spills

Predictive Ecological Assessments

A practical example of a predictive risk assessment can be seen in evaluatingthe potential ecological impact of the introduction of a new herbicide Under theFederal Insecticide, Fungicide, and Rodenticide Act (FIFRA) and its amendments,all new pesticides and herbicides must be registered for each specific use The Office

of Pesticide Programs of the USEPA prepared the Standard Evaluation Procedurefor Ecological Risk Assessment (SEP) (Urban and Cook, 1986) to define the pro-cedures to be used for the ecological risk assessment process In short, this documentpresents a tiered approach to determining the potential for unreasonable adverseeffects on the environment as a result of the use of pesticides and herbicides.Additional information for the testing of new pesticides and herbicides can be found

in the Pesticide Assessment Guidelines (USEPA, 1982a, b)

To illustrate this example, consider a new herbicide for control of weeds in ricepaddies adjacent to riparian and emergent wetlands The goal here is to assess thepotential risk of the new herbicide on the wetland ecosystem based on the toxicologyand environmental fate of the pesticide under the proposed use pattern

The typical rice field (Figure 4) may be 10 to 60 ha in size and laser-leveled

to control water flow The fields are not allowed to get too dry and, if sufficientrainfall is lacking, they may be flushed with water to depth of 2.5 cm or so on afairly regular basis Once the rice is adequately established, the field is typicallyflooded to a depth of 7.5 to 15 cm, which is then maintained until harvest Thetemporary flush and permanent floodwaters are drained to an adjacent ditch thatruns within 100 m to a small bayou Emergent and riverine wetlands are associated