SCOPE 53 - Methods to Assess the Effects of Chemicals On Ecosystems

14.1 INTRODUCTION

Most assessments of risks of injury to ecological species caused by chemicals are currently focused on small-scale, local problems. Local assessments may evaluate the ecological hazards of chemically contaminated sites, point-source effluent discharges to rivers, lakes, and bays, chemical air emissions from individual industrial facilities, or other chemical exposures generally confined to a spatial scale of a few metres to a few kilometres. Current ecological risk assessment methods have typically been developed largely to assess chemical effects at this spatial scale.

Exposures that result from the widespread aerial release and transport of chemicals occur at a much larger, regional scale (hundreds to thousands of kilometres). Examples include acid deposition, long-range transport and deposition of ozone and other air contaminants, and broad-scale aerial application of pesticides, such as those used for spruce budworm control in the forests of eastern Canada. Not only do chemical exposures occur at various scales, but ecological processes also operate at a variety of scales in space and time. Therefore, a need exists to identify approaches to estimate risks to ecological receptors and methods useful for larger-scale, regional assessments of the effects of broad-scale chemical exposures.

Suter (1993) identified six reasons for ecological risk assessments at the regional level. First, local sources of chemicals or radionuclides may have regional consequences from their release. The Chernobyl reactor accident is an example of this scenario. Second, the combined releases of multiple individual sources within a region, each within tolerable limits, may be unacceptable, because of the combined toxic effects of the mixture at the regional scale. An example is the combined airborne releases of chemicals in urban industrial areas leading to degradation in regional air quality. Third, regional scale processes may affect the transformation and transport of airborne chemicals in ways that are not observed at local scales. An obvious example is the formation of photochemical smog from numerous independent sources of hydrocarbons and nitrogen oxides. Fourth, emissions may have effects at regional scales that do not occur on local scales.

The depletion of stratospheric ozone by chlorofluorocarbons provides an example of this phenomena. Fifth, regions possess characteristics that do not occur on local scales, and these characteristics warrant protection. These characteristics are largely associated with patterns in the landscape. Finally, the success of various broad scale regulatory and resource management programmes can be adequately assessed only on a regional scale.

This chapter provides a rationale for regional assessments, discusses spatial and temporal scale considerations for regional scale risk assessments, provides a framework and methods for regional scale assessments, and describes a regional ecological risk assessment case study, that seeks to estimate risks to a region exposed to high levels of ozone.

The assessment of risks on a regional scale is not fundamentally different from that on a local scale. Both regional and site-specific assessments of risk can be performed within the framework recently provided by the US Environmental Protection Agency (USEP A, 1992). The paradigms of predictive and retrospective risk assessment recently presented by Suter (1993) in Ecological Risk Assessment are equally applicable to local and regional assessments. However, in contrast to local assessments, regional scale risk assessments require consideration of a separate set of issues that do not enter site-specific assessments. These include understanding the landscape and the relationship of biota to the landscape, characterizing exposures and effects over large spatial scales and sometimes long temporal scales, identifying endpoints and measurement metrics characteristic of the region, combining data collected at very different scales and extrapolating between scales, integrating the inputs of multiple stressors that operate on large spatial scales, and integrating exposures and effects in various terrestrial and aquatic ecosystems within the region to characterize risk for the region as a whole.

14.2 SCALE ISSUES

Several issues are associated with the spatial and temporal scales of chemical hazards and ecological response to chemical exposures that must be considered when undertaking an ecotoxicological assessment of risk at the regional level. These considerations include:

1. the spatial and temporal scales associated with various levels of biological organization for which effects may be assessed;
2. the relation of landscape scales to the distributions of populations in an area and the way in which they use the available resources; and
3. the spatial and temporal scales of potential chemical exposures.

Figure 14.1. Spatial and temporal scales within which individuals, populations, ecosystems and regions respond to environmental stressors (modified from Suter, 1993)

14.2.1 SPATIAL AND TEMPORAL SCALES RELATED TO LEVELS OF BIOLOGICAL ORGANIZATION

Effects of chemicals may be assessed at the level of the individual, population, community, ecosystem, or region, or at combinations of levels of organization. Populations are composed of individual members of a species in the same area; communities are groups of populations interacting with each other; ecosystems are communities together with their physical and chemical environment; regions, in the context of risk assessment, are spatial groupings of contiguous ecosystems.

Several reasons exist to consider the levels of biological organization and their spatiotemporal scales of operation in planning ecotoxicological risk assessments. First, although both measurement and assessment endpoints can be defined at each level of biological organization, these endpoints are not equally important at different spatial and temporal scales (Suter, 1993). This situation is shown in Figure 14.1. At short temporal scales (days to months) and small spatial scales (micrometre to metres), effects of chemicals can be assessed on micro-organism populations and micro-organism biochemistry and physiology. However, little can be estimated about the ultimate effects of chemical exposure on long-lived organisms or on the dynamics of microorganism populations or ecosystems that operate on larger spatial and longer temporal scales.

On a human time scale, the reproducing population is the smallest persistent ecological unit. The effects of chemicals on populations of organisms can be assessed on a temporal scale of months to years and a spatial scale of metres to kilometres. Populations, however, do not live in a vacuum. Direct chemical effects on one or more populations may, in turn, affect other populations in the exposed community. Such effects may occur indirectly through changes in habitat availability or in predator-prey or competitive relationships or other mechanisms. Therefore, a partial overlap exists in spatial and temporal scales for the assessment of chemical effects on population dynamics and community and ecosystem dynamics. However, because additional secondary effects may occur over longer time scales and may spread to larger areas, the range of spatiotemporal scales to evaluate chemical effects on ecosystems is broader than for the assessment of effects on populations alone (i.e., years to hundreds of years and metres to hundreds of kilometres). The spatial scale of regional responses to chemical exposure overlaps with, but extends beyond, that of individual exposed ecosystems (hundreds of kilometres to thousands of kilometres). The temporal scale of regional dynamics also overlaps that of the component ecosystems, but may be longer than that for any one ecosystem in the region (tens of years to hundreds of years). Spatial and temporal scales associated with ecological effects at the various levels of biological organization are further discussed by Sheehan (1984a, 1984b) and Suter (1993).

14.2.2 TOOLS FOR REGIONAL SCALE ASSESSMENTS

A second issue of scale is the practical constraints of assessing ecotoxicological effects at the higher levels of biological organization. Implementation of small scale, short-term population studies is much easier than that for long-term, regional ecotoxicological studies. Field studies to support risk assessment at the ecosystem and regional level may require investigation on a time scale of years and a spatial scale of hundreds of kilometres. Such studies require substantial expenditures of manpower and money, and may not be practical for compliance with regulatory requirements. A recognition of costs of long-term broad-scale field studies has spurred the development of new sampling tools, such as satellite imagery, which are well suited to providing data over extended time periods and large geographical areas, but are not labour intensive. Tucker and Sellers (1986) use satellite remote sensing to assess primary productivity across broad spatial scales. More recently, Simmons et al. (1992) and colleagues used satellite imagery in conjunction with small-scale field studies to evaluate the dispersion of plant cover in semi-arid areas of the state of Washington and the utility of satellite imagery as tools in the assessment of landscape response to physical and chemical stressors.

Geographical information systems (GIS) are also being designed to organize and analyze data on the distribution, accumulation, and effects of chemicals on landscapes (Bartell et al., 1992). In addition, the cost of large-scale field investigation has also prompted the development of models to extrapolate from smaller to larger spatial scales and to simulate regional landscape response to stressors. Solomon (1986) proposed a model to assess the response of forests in eastern North America  to CO₂-induced climate change. Turner (1987) compared the utility of three different models in predicting landscape changes resulting from physical disturbance events. More recently, Graham and colleagues (1991) used a simulation model to predict the effects of ozone on Adirondack forests in the state of New York. This study is described in greater detail as a regional risk assessment case study below. Clearly, regional scale ecotoxicological risk assessments will require extensive use of these types of tools to be cost-effective.

14.2.3 THE SCALE OF LANDSCAPE STRUCTURE AND ITS INFLUENCES ON ANIMAL POPULATIONS AND RESOURCE USE

A second scale consideration for regional ecotoxicological risk assessments is the relationship of scale of landscape patterns (including vegetative structure) with the distribution of animals and the scale of their use of resources. At a regional scale, landscape responses to chemical stressors are often assessed in terms of changes in vegetation and vegetative structure, and few data are provided on the secondary effects of such changes in vegetation on animal populations. Understanding the relationship of animal population dynamics to the structure of the landscape is essential to predicting changes in animal populations resulting from changes in the vegetative structure of their habitat.

Recent studies have evaluated the proposition that a small set of plant, animal, and abiotic processes structure ecosystems across scales of time and space. Based on his studies, Holling (1992) concluded that terrestrial bird and mammal populations are dispersed according to body size by the discontinuous hierarchial structures and textures of the landscape. He found evidence for eight distinct habitat "quanta" defined by distinct textures at a specific range of scales. These eight quanta together cover tens of centimetres to hundreds of kilometres in space and months to hundreds of years in time. All trophic levels of birds and mammals utilize resources in their foraging areas in the same way by measuring the spatial gain of habitat patches defined by their size (i.e., step length or some minimum unit of measurement).

Therefore, large mammals and birds have large home range areas, and these are to some degree limited by the vegetative structure and texture of the landscape. For example, a large wading bird, such as the Great Egret (Casmerodius albus) has a short-term foraging area that is small (a few metres), but a habitat area home range during a year that may extend from tens to hundreds of kilometres. Over several years, this area may extend over thousands of kilometres. The spatial and temporal scales within which wading birds operate are shown in Figure 14.2.

Chemically mediated changes in the vegetative structure of the landscape will influence both the distribution and abundance of these species with large home ranges. Consideration of such secondary effects is essential for regional scale risk assessments. The work of Holling (1992), Gass and Montgomerie (1981), and Orians (1980) on body size and the size of home and foraging ranges provides useful data on the mammal and bird species of interest in regional assessments and the spatial scale at which the dynamics of these populations should be assessed.

Figure 14.2. Temporal and spatial scales within which large wading birds operate daily and over their lifetimes (modified from Holling, 1992)

14.2.4 SPATIAL AND TEMPORAL SCALES OF CHEMICAL HAZARDS

Chemical exposures occur over a range of spatial and temporal scales. A representation of the spatial and temporal scales of selected chemical hazards is shown in Figure 14.3. Single local applications of pesticides often occur on a spatial scale of metres in homes and gardens to perhaps a kilometre in agricultural settings over a temporal scale of hours to a few days. Chemical and petrochemical spill events occur on a somewhat larger spatial scale, but the residual contamination from these spills may pose exposure on a temporal scale of days to years. Chemical releases from hazardous waste sites and liquid effluent discharges occur on a spatial scale similar to spill events, but exposures from these sources occur over longer temporal scales of months to tens of years. In contrast to these small-scale chemical releases, broad-scale applications of pesticides, such as aerial spraying to control grasshoppers in the prairie regions of the United States and Canada (Sheehan et al., 1987) and the spruce budworm in the forests of New Brunswick (Mitchell and Roberts, 1984), cover hundreds to thousands of kilometres in space, and may occur intermittently over years to tens of years. At the extreme, widespread aerial transport of photochemical oxidants and acids have resulted in contamination of large geographical regions (thousands to tens of thousands of kilometres). Cumulative exposures to airborne chemicals also takes place over long time periods (tens to hundreds of years).

 Figure 14.3. Arrangement of the spatial and temporal scales of selected chemical hazards (modified from Suter, 1993)

Clearly, regional exposures to chemicals occur as a result of widespread, long- term deposition of airborne chemicals or broad-scale pesticide applications. In both cases, exposures may cover several individual aquatic and terrestrial ecosystems, and may last for tens to hundreds of years. This scale of exposures points to the need for two types of methods: (1) measurement methods that can be used to assess effects on a regional scale and (2) extrapolation methods that can be used to scale-up local concentration-response data to predict the effects on regional dynamics. A framework and methods for regional scale ecotoxicological risk assessments are discussed in the following sections.

14.3 A FRAMEWORK AND METHODS FOR REGIONAL SCALE ECOLOGICAL RISK ASSESSMENTS

Although the need for ecotoxicological risk assessments at a regional scale is clear, little has been written about methods for such assessments, and few examples are available. As such, this discussion of framework and methods for regional scale ecotoxicological risk assessment is not so much a critique of approaches as it is a review of considerations for such assessments.

The framework for regional ecological risk assessment was first described by Hunsaker and colleagues (1990). They proposed a two-phased approach to regional assessments:

1. The definition phase, in which the regional endpoints, source terms, and reference environments are defined.
2. The solution phase, in which the exposure and effects are assessed, the exposure levels are related to levels of effects to estimate risk, and risks are extrapolated to a regional scale.

An example of the application of this two-phase approach to assess the effects of regional chemical exposures was provided by Graham et al. (1991). They evaluated the probabilities of significant changes in Adirondack forests as a consequence of ozone exposures and related beetle attacks on trees (discussed below). More recently, Suter (1993) described a set of considerations specific to regional scale assessment. These include scaling, landscape description, and integration of several qualitatively dissimilar stresses. In addition, while landscape ecologists have largely ignored chemical exposures, they have developed methods  to assess changes in landscape patterns as the result of physical disturbances, and these methods may find applications in regional ecotoxicological assessments. Examples are provided in the work of Turner (1987), Turner and Gardner (1991), Urban et al. (1987), and others published in the journal Landscape Ecology.

Although our experience with regional scale ecotoxicological risk assessments is limited, potentially a great deal of carry-over exists from the framework and methods for local scale assessments for application to regional scale assessments.

The recently proposed generic EPA framework for ecological risk assessment is applicable to assessments at both local and regional scales (USEPA, 1992). A slightly modified version of the EPA framework that emphasizes considerations for regional scale risk assessments is presented in Figure 14.4. This diagram reinforces the need for understanding of the unique features of the region as well as the terrestrial and aquatic ecosystems within the region. A regional scale risk assessment should include problem formulation, exposure and effects assessments, and risk characterization for the component terrestrial and aquatic ecosystems and the region as a whole. The risks for the component ecosystems must be integrated over the region to characterize risks for the regional unit. The importance of airborne chemical inputs to regional-scale chemical hazards and the hydrological linkage as an important mechanism of chemical transport are emphasized in the diagram.

 Figure 14.4. Generic framework for a regional scale ecotoxicological risk assessment

Ecological risk assessment of chemical exposures and effects at the regional scale may be either predictive or retrospective. Predictive risk assessments begin with a hypothetical chemical release scenario that could contaminate a large geographic area (such as the proposed broad-scale application of the new synthetic pyrethroid insecticides for grasshopper control in the prairie region of North America), and proceeds to the estimation of the risks of ecological effects such as direct toxicity to aquatic organisms and indirect effects on waterfowl populations that may be associated with broad-scale pyrethroid insecticide use (Sheehan et al., 1987). In contrast, retrospective risk assessments at a regional scale generally begin with evidence of chemical contamination (e.g., acid deposition) and proceed to the characterization of risks associated with ecological responses to this contamination such as changes in soil chemistry and subsequent effects on soil fauna and the tolerance of trees to insect pest infestations reported for acid-stressed forests (Loucks et al., in press). 

The sources of information are generally quite limited for a predictive risk assessment; therefore, the evaluation of an exposure scenario is usually based on modelling and effects are characterized from toxicity data extrapolations. The tools of the predictive assessment may also be used in a retrospective assessment along with field data. Regardless of whether an ecotoxicological risk assessment is predictive or retrospective in nature, it will likely contain four components: (1) problem formulation, (2) exposure assessment, (3) effects assessment, and (4) risk characterization.

The first corresponds to the definition phase and the final three components to the solution phase described by Hunsacker et al. (1990). The components of a regional scale ecotoxicological risk assessment are identified in Figure 14.5. The options as to what to assess at a regional scale and how to assess it are discussed for the planning phase in the following sections. An example of exposure assessment, effects assessment, and risk characterization for a regional assessment are provided in a case study below.

14.3.1 PROBLEM FORMULATION

Problem formulation is the first phase of an ecotoxicological risk assessment. This planning phase establishes the assessment objectives and scope and provides a "blueprint" for the risk assessment process. Problem formulation should include the evaluation of existing data, identification of the region and ecosystems at risk, identification of the chemicals of interest, the establishment of risk assessment objectives and scope, selection of measurement and assessment of endpoints, and development of a conceptual risk assessment model to guide the solution phase activities.

14.3.1.1 Evaluation of existing data

Regardless of the type of risk assessment, information generally is available on actual or potential sources of chemicals and the region and types of ecosystems into which the chemicals may be or have been released. Source data may be in the form of chemical inventories and design properties for emission sources at manufacturing facilities such as that produced to meet the requirements of Hazardous Air Pollutant Provisions (Title III) of the US Clean Air Act Amendments of 1990. The modelling of these data to predict airborne chemical concentrations surrounding individual facilities was performed in California to meet  Air Quality Management District Requirements (Conner et al., 1992). The combination of individual facility data could be used to provide an estimate of chemical concentrations in various California air basins. Emissions data for sulphur and nitrogen oxides (Slade, 1990) provided the focus for ecotoxicological risk assessments of the effects of acid deposition on forests in exposed regions of Ohio (Loucks et al., 1993) and Pennsylvania (Nash et al., 1992).

Figure 14.5. Components of an ecotoxicological risk assessment for regional scale evaluations

Data on chemical concentrations in air and other media are also available from various national and regional monitoring programmes. The USEPA monitors sulphur dioxide (SO₂), carbon monoxide (CO), nitrogen dioxide (NO₂), ozone (O₃), and lead (Pb) at over 200 fixed locations throughout the United States (USEPA, 1991). State air monitoring programmes such as that in the South Coast Air Quality Management District in Southern California provide regional data on a wider variety of chemicals commonly released to the atmosphere from mobile and point sources in that region.

Several sources may provide useful data for regional assessments of water quality in the US The US Geological Survey (USGS) and state water resource agencies have monitored the quality of surface water throughout the United States since the 1940s. Monitoring is conducted on a regular basis at fixed location within river basins. Data collected includes concentrations of nutrients, trace metals and pesticides in water. This programme has provided a 20 to 50-year record of concentration trends of selected chemical substances in more than 500 watersheds across the United States.

The USGS more recently established the National Water Quality Assessment Program (NWAQAP) which will eventually investigate about 120 study areas distributed throughout the United States (Hirsch et al., 1988). These watersheds incorporate about 80 percent of the nations water use. These study units will be linked together to form a national network by using a prescribed set of study approaches and protocols for each river basin. The assessment program is perennial to provide data for trends analysis. The programme is focused on water quality conditions that are prevalent or large in scale and persistent in time. Chemical measurements focus on a set of target variables including physical measurements, inorganic constituents, and organic compounds. Biological measurements include plant and animal tissues to help determine the occurrence of trace element and organic compounds, to provide a measure of the bioavailability of these contaminants in water and sediments, and to help understand their environmental fate. Biological measurements also include aquatic toxicity tests and ecological surveys to assess toxicity of water and sediment, document the current status of the biological community, and describe and explain, to the extent possible, the relationships of the biological communities to the physical and chemical characteristics of the drainage river or streams.

A similar long-term monitoring programme providing data on chemical accumulation trends in watersheds is the National Contaminant Biomonitoring Program (formerly called the National Pesticide Monitoring Program). Since 1964, this programme has periodically analysed residues of selected organochlorine chemicals and trace metals in samples of fish and wildlife collected from a nationwide network of over 120 stations (May and McKinney, 1981; Schmitt et al., 1981, 1983, 1985). The National Contaminant Biomonitoring Program data have documented the widespread distribution of DDT and PCBs in fish and the decline in the residue concentrations of these chemicals following discontinuation of their use. This programme has also provided data to identify new sources of persistent chemicals.

Recognizing its lack of an integrated approach to monitoring indicators of ecological conditions and exposures to chemicals in ecosystems, the USEPA has implemented the Environmental Monitoring and Assessment Program (EMAP) (Messer et al., 1991). This programme is focused on documenting changes at the regional level. The EMAP defines stress indicators and economic, social, and engineering data that can be used to determine the most probable sources of physical, chemical, and biological stressors. Exposure indicators are physical, chemical, and biological measurements that can be related to chemical exposure, habitat degradation, or other causes of poor ecosystem conditions. Response indicators are biological measures that quantify the condition of the region and ecosystems, and integrate the effects of various stresses. Examples include evidence of gross pathology, presence or absence of "sentinel" species, effects on keystone species that are important to maintain ecosystem structure, changes in populations in species that are of sports, commercial, or aesthetic interest, and effects on ecosystem structure (e.g., diversity) or function (i.e., primary production).

A goal of EMAP is the identification of a suite of assessment endpoints applicable to physical, chemical, and biological stressors in various types of ecosystems. The EMAP monitoring strategy is to establish a systematic grid of sampling points across the United States, including the continental shelf waters, and to make field measurements on a suite of indicators at grid points of a certain density and at a certain time period (e.g., a four-year average baseline). This programme will provide long-term regional data to assess the status and trends in condition of large-scale ecological systems.

14.3.1.2 Identification of the region and ecosystems potentially at risk

The region and ecosystems within which widespread exposures to chemicals may occur or have occurred is the geographic area of interest for an ecotoxicological risk assessment. However, several complicating factors are present in setting assessment boundaries. For a regional assessment to be effective, the spatial and temporal boundaries must be defined appropriately for both the hazard and assessment endpoints. In this sense, the assessment region may be effectively defined by the projected or actual distribution of chemicals of interest in the environment and/or the distribution of the populations or ecosystems of interest.

As an alternative, the bounds on the region may be based on the distribution or magnitude of the source. For example, a source-defined region includes the Appalachian coal mining area as a region for the assessment of acid mine drainage (Suter, 1993). Regions may also be defined in terms of natural features such an watershed or air basin boundaries, physiographic provinces, or ecoregions. Naturally defined regions have relatively uniform physical and biological properties, and are, therefore, more easily described than more heterogeneous geographic areas. These regions also include the hydrologic and atmospheric processes that, in large part, control chemical transport and exposures. The watershed is an effective unit for the study of effects of physical disturbance (e.g., clear cut logging) on terrestrial and aquatic ecosystems (Bormann et al., 1968), and for the assessment of the effects of chemicals on water quality and aquatic biota. The watershed is the assessment unit for the National Water Quality Assessment Program described earlier (Hirsch et al., 1988). Air basins provide a more appropriate assessment unit for chemicals transported through the atmosphere. The air basins containing coal burning plants in Ohio and Pennsylvania have been used to bound assessments of acid deposition effects on forested regions (Loucks et al., 1993; Nash et al., 1992).

The region may also be defined by the overlap of the chemicals and habitats of the populations of interest. For example, the assessment of potential effects of regional applications of insecticides in the Canadian prairies to control grasshopper infestations was bounded by an analysis of the area of overlap between waterfow nesting habitat and crop areas receiving aerial applications of the specific insecticides of concern (Sheehan et al., 1987).

Historically, regional assessment boundaries also have been defined by political boundaries. As the dispersal of chemicals in the environment is unconstrained by political boundaries, defining an ecotoxicological risk assessment by the political boundary is largely artificial. Where possible, assessments should be done on a region defined by appropriate ecological and anthropogenic factors with political boundaries applied secondarily as an additional overlay.

14.3.1.3 Identification of chemicals of interest

One or more chemicals may be of interest in a regional ecotoxicological risk assessment. For a pesticide application, the chemicals of concern are generally well defined and few in number. For insecticide spraying for spruce budworm control in the forests of New Brunswick, Canada, only DDT, fenitrothion, phosphamidon, trichlorfon, and aminocarb were identified as widely used. These risk assessments focused on DDT and fenitrothion (Brooks, 1974; NRCC, 1977; Mitchell and Roberts, 1984). Similarly, Sheehan et al. (1987) identified 13 insecticides that are widely used and aerially applied for insect pest control on prairie oil seed grain crops in Canada. Again, screening evaluations based on toxicity to aquatic invertebrates showed that only six compounds (permethrin, azinophos, methyl chlorpyrifos, deltamethrin, methoxychlor, and cypermethrin) potentially posed a high risk to aquatic organisms or waterfowl under normal application conditions.

For aqueous effluents and airborne emissions, the number of chemicals released may be tens or hundreds. Although in some cases understanding the risks posed by all of the chemicals released in effluents may be desirable, assessing the exposures and effects of hundreds of chemicals in either a human health or ecotoxicological risk assessment is generally impractical and unnecessary .The USEPA has recognized this issue in their technical risk assessment guidance (USEPA, 1989). Rather than undertake a superficial and unwieldy assessment based on quantitatively evaluating tens or hundreds of chemicals, risks should be assessed only for those chemicals that are likely to pose the greatest percentage of the risk. The USEPA suggests several screening procedures based on mobility, persistence, bioaccumulation, concentration, toxicity, and other site, or area, specific factors that can be used to identify and justify a subset of the chemicals released for risk assessment (USEPA, 1989). The procedures developed for human health assessment are also applicable to ecotoxicological assessments. The list of chemicals of primary ecological concern can be limited by screening chemicals for environmental persistence, bioaccumulation potential, and toxicity to representative species. Laskowski et al. (1982) describe several indices of mobility and persistence based on combinations of physical chemical properties (leaching potential, volatility potential, on-site exposure potential) that can be used to rank chemical persistence. Bioaccumulation and toxicity data for fish are summarized in the USEPA AQUIRE Database (Environmental Research Laboratory Duluth, Minn.). Toxicity data for selected wildlife species are present in various US Fish and Wildlife Service Publications (Contaminant Hazard Reviews Reports 1 to 24). Plant toxicity data are summarized in the PHYTOTOX Database (Department of Botony, University of Oklahoma).

Examples of chemical screening for ecotoxicological assessments are available. A ranking procedure based on chemical persistence (Koc, Vp and T_1/2), bioaccumulation, potential (K_ow and BCF) and toxicity to fish (LC₅₀) and rodents (LD₅₀) has been used to identify 29 of 189 chemicals released from a hazardous waste incinerator as appropriate for a quantitative ecotoxicological assessment of incinerator emissions. Similar, although less well documented, procedures were used by various researchers to identify the chemicals of interest in assessing ecological effects of effluents to the Great Lakes (Evans, 1988).

For some of the more obvious regional contamination issues such as acid deposition and photochemical oxidant exposures, the chemicals of primary interest have been identified. Acid deposition exposures have been described in terms of mass loading of sulphur, nitrogen, and hydrogen ions (Loucks et al., 1993; Nash et al., 1992). Photochemical oxidant exposures have been further quantified using ozone as a surrogate for the oxidant mixture (Skelly, 1980; Graham et al., 1991).

14.3.1.4 Establishment of risk assessment objectives and scope

To be meaningful and effective, an ecotoxicological risk assessment must be scientifically valid and relevant to regulatory needs and public concerns. Therefore, the regulatory or risk management framework within which the assessment is to be used should be considered in identifying the risk assessment objectives. In the US, for example, the Endangered Species Act requires analysis at the level of the individual of a threatened or endangered species. The Comprehensive Environmental Response, Compensation and Liabilities Act (CERCLA) requires that actions selected to remedy hazardous waste sites be protective of human health and the environment. CERCLA assessments generally require analysis of effects at several levels of biological organization and the development of risk-based remediation targets (USEPA, 1989).

Table 14.1. Examples of regulatory and risk management objectives and the scope of regional scale ecotoxicological risk assessments

On the regional scale, ecotoxicological risk assessments may be conducted to meet various regulatory , risk management, and resource management requirements. Selected examples of risk management objectives and the risk assessment scope to address these needs are presented in Table 14.1. Predictive assessments can be used to evaluate the potential ecological effects of the proposed broad-scale aerial application of a new pesticide, the likely positive effects on ecological systems that might result from a reduction in regional concentrations of air contaminants due to regulatory control measures or the long-term effects of projected increases in chemical releases to air or water based on various scenarios of population growth and consumption patterns and industrial production. By contrast, retrospective approaches can be used to evaluate the relative effects of chemical, physical, and biological stressors on resources such as fisheries in the Great Lakes or to assess the effects of pesticides or air contaminants on regional ecological systems.

Although chemical deposition from airborne releases constitutes an obvious problem requiring regional-scale assessments, multiple releases of chemicals in aqueous effluents to aquatic systems may also pose regional scale exposure issues. The investigation of ecological effects in the Great Lakes associated with widespread exposures to persistent organic chemicals and metals is an example of a regional water quality issue which should be addressed with a regional scale risk assessment approach (Evans, 1988).

The scope of a regional ecotoxicological risk assessment is likely to include analyses of both population-level and ecosystem-level exposures and effects as well as larger scale effects on landscape structure of productive capacity of the region. Regional assessments will most certainly require an analysis of multiple stressors due to the wide variety of anthropogenic activities and associated chemical releases, habitat destruction, and biological resource use that occur within large geographic areas. Multiple stressors may include chemical mixtures or combinations of chemical with physical stressors such as habitat alteration or biological stressors such as hunting or fishing pressures. The key to the integration of multiple stressors into an ecotoxicological assessment is the identification of common assessment endpoints. For example, an endpoint such as recruitment abundance (i.e., the number of young added to the population) can be used to integrate the individual effects of chemicals, habitat alteration, hunting, drought, or other stressors on large mammal and bird populations (Sheehan et al., 1987; Barnthouse et al., 1990). The effects of both chemical and physical stressor on plants can be expressed in terms of reductions in primary production (Adams et al., 1985). By contrast, individual toxicological endpoints such as LC₅₀ and no observed effect levels for chemicals have no direct equivalent in the terms generally used to describe system-wide impacts. Suter (1993) suggests that ecotoxicologists look to resource managers who have developed appropriate endpoints to integrate the effects of various stressors on plant and animal populations (e.g., USFWS, 1980; Bovee and Zuboy, 1988).

In large-scale ecotoxicological risk assessments, both the direct toxic effects of chemicals on individuals and populations and indirect effects of chemicals on the environment or biological resources that may subsequently affect other populations, communities, and ecosystems should be considered in quantifying risks. For example, in studies of the effects of acid deposition on forests, researchers have shown that the oxides of sulphur and nitrogen increase soil acidity and enhance the leaching of Ca²⁺ and Mg²⁺ cations essential for plant growth (Loucks et al., 1993).

The acidification of soils can also affect soil invertebrate populations, such as earthworms which are important to the decomposer food web and nutrient recycling. Loucks et al. (1993) reported that the accumulation of undecomposed organic matter in surface soils was positively correlated with high levels of acid deposition. The reduction in essential cations and plant nutrients in acidified soils. places a physiologic stress on trees in these areas and leaves them more susceptible to insect attacks (Haack and Blank, 1991). Thus, the effects of acid deposition on forest production are largely indirect, and are the result of environmental changes that place trees under stress.

Table 14.2. Examples of possible assessment and measurement endpoints for evaluation of the toxicity of insecticides sprayed for control of spruce budworm

Another example of the importance of considering indirect effects is provided in the work of Sheehan and colleagues (1987) on the potential impacts of pesticides on prairie-nesting duck populations. Their evaluation showed that although the new synthetic pyrethroid insecticides pose a low toxicity hazard to ducks at recommended application rates, the widespread aerial application of these insecticides could effect the recruitment of young  in sprayed regions by substantially reducing aquatic invertebrate populations, an essential food resource for ducklings, during the critical growth period following hatching. In years when there was widespread aerial application of these insecticides for grasshopper control during periods of duckling hatching and rearing, insecticide effects on recruitment and abundance of regional duck populations may be equivalent in magnitude to population losses from hunting (Sheehan et al., 1987, 1993).

Table 14.3. Types of potential assessment endpoints for regional ecotoxicological risk assessments (modified from Suter, 1993)

Clearly the accurate characterization of atmospheric and hydrologic transport of chemicals will be important in quantifying chemical exposures at a regional scale. Atmospheric transport plays a key role in the acidification phenomenon noted in North America and Europe. The distance a chemical is transported from its source depends on meteorologic factors, geographic factors, and characteristics of the chemical itself. A useful feature which describes, in general terms, a chemical's atmospheric behaviour is its average residence time in the atmosphere. Pollutants such as sulphur and nitrogen oxides, that have residence times of approximately two days (Rodhe, 1978), are particularly important on a regional scale, because this time period is comparable to the time typically required for atmospheric transport across eastern North America or Western Europe. Hidy et al. (1979) classified meteorologic situations that lead to long-range transport and regional pollution in the Eastern United States. Several atmospheric fate and transport models for chemicals are available, and were recently reviewed by Zannetti (1990).

Precipitation is one of the primary mechanisms removing chemicals for regional air and depositing them into terrestrial and aquatic ecosystems. Precipitation runoff and erosion are also key mechanisms of transport to surface waters of chemicals initially deposited in surface soils. The importance of properly characterizing these hydrologic linkages is again key to predicting regional changes in water quality (Hunsaker et al., 1990). Several models have been developed to predict the hydrologic transport of chemical contaminants. Among these are the Midwest Research Institute Nonpoint Source Loading Function Model, the Environmental Pollution Assessment Erosion Sedimentation and Rural Runoff Model, the US Department of Agriculture Hydrograph Laboratory Model, and the Chemical Runoff and Erosion from Agricultural Management Systems Model.

14.3.1.5 Selection of measurement and assessment endpoints

An endpoint is a characteristic of an ecological component that may be affected by exposure to a chemical. Two types of endpoints should be identified: assessment endpoints and measurement endpoints.

An assessment endpoint is a formal expression of the environmental value to be protected (Suter, 1989). Given the diversity of ecosystems and the various values placed on them by society, no universal list of assessment endpoints exists. However, Suter (1993) has identified five criteria that any assessment endpoint should satisfy: societal and biological relevance, unambiguous operational definition, accessibility to prediction and measurement, and susceptibility to the hazardous agent. Societal relevance implies that the endpoint should be understood and valued by the public. Biologically relevant endpoints reflect important characteristics of the system, and are related to other endpoint up the ecological hierarchy. Unambiguous operational definitions of endpoints are essential to provide direction to testing and modelling within the assessment. Goals such as "ecosystem health" are inadequate for assessments.

Measurement endpoints are the quantifiable responses to a chemical stressor that can be related to the valued characteristic chosen as the assessment endpoint (Suter, 1989). Measurement endpoints are the analyst's input to the risk assessment, and include LC₅₀ and LD₅₀ from toxicity tests and population estimates from field sampling studies. Although assessment and measurement endpoints can be identical, they are frequently defined differently. Several examples of assessment and measurement endpoints that might be used to evaluate the ecological effects of pesticide spraying for spruce budworm control are provided in Table 14.2.

The number of possible assessment endpoints for a region assessment is greater than for a single ecosystem risk assessment. Endpoints are unique to the component populations and ecosystems within the region as well as for the region as a whole.

A list of possible assessment endpoints is presented in Table 14.3. These include traditional ecological endpoints such as population extinction and abundance as well as endpoints characteristic of regions, and anthropocentric endpoints such as sport fisheries values. Traditional ecological endpoints are assessed in the same manner for local and regional ecotoxicological risk assessments, but results must be scaled up for the region.

Representative regional assessment endpoints, however, are less obvious. As Suter (1993) points out, biological productivity is clearly an important regional endpoint, but it is not readily defined. Crop and timber production are largely controlled by economic factors, and are heavily influenced by technology. Therefore, realized production is not a suitable measure of productive capacity or the influence of chemicals on productivity. An approach to overcoming the problem of controlling for factors other than chemicals is to focus on production of species crops or forest types and normalize results for weather, fertilizer, energy, and other inputs. An alternative approach is to assess the proportion of a region devoted to biological production that is lost, or lost from production due to chemical exposures, again controlling for other factors.

Regional assessment endpoints should be defined in terms of observations that .can be made over large geographic areas and long time periods. For terrestrial systems, endpoints might include percent cover of different vegetation types. For aquatic systems, a representative endpoint might be the frequency of lakes in which a valued fish species becomes extinct.

Integrated properties of landscapes described by landscape ecologists may also be representative regional assessment endpoints. Examples of landscape properties include dominance (i.e., the degree to which the landscape is dominated by a particular feature), contagion (the degree to which the landscape is dissected into small patches or aggregated into large patches), fractal dimension (index of complexity of shapes on the landscape), and amount of edge (Krummel et al., 1986; O'Neill et al., 1988; Hunsaker et at., 1990; Graham et al., 1991). These indices can be calculated from remote sensing data, and, therefore, may be particularly useful in large-scale regional assessments.

Table 14.4. Ecological endpoint measures used in the Adirondack demonstration (modified from Graham et al., 1991)

where D is the index of dominance, n is the total number of land-use categories, and P_i is the proportion of the region in land use i;

where C is the index of contagion, n is the total number of land-use categories, and P_u is the probability that a unit (grid cell) of land use i will be found adjacent to a unit of land use j.

14.3.1.6 Development of a conceptual risk assessment model

The conceptual risk assessment model is developed from a series of working hypotheses regarding how the chemical(s) might affect the ecological components of the region and the region as a whole. Although many hypotheses may be developed for regional exposure scenarios, only those that are considered most likely to contribute to risk should be selected for quantitative evaluation. For these hypotheses, the conceptual model describes the approach and methods that are to be used in the analytical phase and the types of data and analytical tools that are needed. An example of the development of a conceptual risk assessment model and the implementation of this model to assess the potential effects of ozone on regional forests is provided in the case study below.

14.4 REGIONAL RISK ASSESSMENT CASE STUDY

Graham and colleagues (1991) provide an example of a predictive regional ecotoxicological risk assessment of the effects of ozone exposures on forests in the Adirondack region of New York (US).

14.4.1 PROBLEM FORMULATION

In forested areas, the initial effect of elevated ozone concentrations is manifested as physiological stress in trees. Tree response to ozone is a function of cumulative uptake and, as such, is related to leaf or needle lifespan. Thus, chronic exposure to elevated ozone is expected to have a greater effect on conifers with long-lived needles than on deciduous trees (Reith, 1987). Stressed trees are more susceptible to bark beetle attack than unstressed trees (Payne, 1980). Once a tree is attacked, bark beetles will then attack neighbouring trees, spreading the infestation. Infested trees frequently die as a consequence of bark beetle attacks due to subsequent infestation by blue stain fungi (Coulson, 1980). Bark beetle-induced tree mortality can alter the amount and type of forest cover in regions with substantial ozone exposures. The conceptual risk assessment model is one in which chemical exposures are related indirectly to tree death and the amount of forest cover through a decrease in tree tolerance to insect and fungal disease infestations.

Figure 14.6. Bark beetle patch-size distribution under the high- and low-ozone scenarios (Graham et al., 1991)

The assessment endpoints for the demonstration are a 10-25 percent deviation from the baseline value of the endpoints identified in Table 14.4, with the landscape indices.

A stochastic spatial simulation model of land-cover change induced by ozone-triggered bark beetle infestations was used to quantify potential ozone impacts on land cover. In the model, a uniform concentration of ozone was assumed to stress trees so that bark beetle infestations would convert coniferous forest grid cells to open-land grid cells, and mixed forest grid cells to deciduous forest grid cells.

14.4.2 EXPOSURE ASSESSMENT

Two scenarios of elevated ambient ozone concentration were used in this analysis. The high-ozone scenario assumes an average daily maximum concentration of 0.20  based on conditions slightly worse than recorded in the San Bernardino Mountains in California (Taylor, 1973). The low-ozone scenario assumes an ozone concentration of 0.04 µL/L, which is slightly higher than currently measured in the Adirondack Mountains (NYSDEC, 1986). Both scenarios assume uniform exposure across the study region. 

14.4.3 EFFECTS ASSESSMENT

The probability of an initial bark beetle attack was assumed to be 0.015 under the low-ozone exposure scenario and 0.04 under the high-ozone exposure scenario. Each attack could result in a patch of dead conifer trees ranging in size from 1 grid cell (4 ha) to 15 grid cells (60 ha). The probability parameters were developed from data on ponderosa pine mortality in the San Bernardino Mountains and bark beetle patch size in the Southeastern US (Taylor, 1973; Coster and Searcy, 1981). The probability of the size distribution of patches is shown in Figure 14.6. Under the low-ozone scenario, the distribution is based on the pattern of beetle infestation patches observed in forest in North Carolina (Coster and Searcy, 1981). For the high-ozone scenario, the tail of the distribution was extended and the patch size increased. The modelled distribution of patch size for the high ozone scenario reflects both the survey data and field observations that, under most circumstances, initial bark beetle attacks do not spread to other trees; however, when conditions are right, the beetle infestation spreads readily (Coulson, 1980).

14.4.4 RISK CHARACTERIZATION

Risks were characterized by first calculating the baseline value for each measurement endpoint from the initial land cover pattern. For each endpoint and ozone exposure scenario, a cumulative frequency diagram of endpoint values was developed from 100 model simulation runs. Using these diagrams and baseline values for measurement endpoints, the frequency of runs in which the value of the endpoint shifted greater than 10 percent (detectable change) and greater than 25 percent (significant change) was found from the baseline value. Risk probability was calculated by dividing the frequency of runs by their total number.

The risk assessment results are summarized in Table 14.5. The analysis showed that detectable, but not significant, changes in coniferous, deciduous, and mixed-forest types were likely under the high-ozone scenario. In contrast, under the low-ozone scenario, zero risk existed for detectable change in any forest cover type. Most of the pattern-sensitive endpoints were affected by the ozone scenarios. Significant changes in forest edge habitat were highly probable under both ozone scenarios. The projected changes in forest edge habitat can be shown diagrammatically.

Table 14.5. The risk to endpoints under the low- and high-ozone scenarios, and the average values for baseline and altered landscapes under both scenarios^a (from Graham et al., 1991)

This case study demonstrates the importance of considering spatial heterogeneity in evaluating the effects of airborne chemical deposition on the original landscape. A spatial model is essential to quantifying risks to the land-cover pattern and indirectly to timber resources and wildlife habitat.

14.5 PROSPECTS FOR REGIONAL SCALE ASSESSMENT

Although few comprehensive regional scale ecotoxicological risk assessments have been conducted, the need for such assessments is obvious. The need is particularly strong with respect to assessing the ecological effects of air contaminants such as ozone and oxides of sulphur and nitrogen.

To encourage the use of regional risk assessment tools to gather data over wide geographic areas, such as satellite imagery, tools to manage large data sets such as GIS, and models to accurately simulate ecological responses to broad-scale and long-duration chemical exposures, will need to be developed further, and incorporated into the process to make it more useful and cost-effective.

14.6 REFERENCES

Adams, R.M., Crocker, T.D., and Katz, R.W. (1985) Yield response in benefit cost analyses of pollution-induced vegetation damage. In: Winner, W.E., Mooney, H.A., and Goldstein, R.A. (Eds.) Sulfur Dioxide and Vegetation. Stanford University Press, Palo Alto, California.

Barnthouse, L.W., Suter II, G.W., and Rosen, A.E. (1990) Risks of toxic contaminants to exploited populations: influence of life history , data uncertainty, and exploitation intensity. Environ. Toxicol. Chem. 9, 297-311.

Bartell, S.M., Gardner, R.H., and O'Neill, R.V. (1992) Ecological Risk Estimation. Lewis, Chelsea, Michigan.

Bormann, F.H., Likens, G.E., Fisher, D.W., and Pierce, R.S. (1968) Nutrient loss accelerated by clear-cutting of a forest ecosystem. Science 159, 882-884.

Bovee, K.D., and Zuboy, J.R. (Eds.) (1988) Proceedings of a Workshop on the Development and Evaluation of Habitat Suitability Criteria. USFWS Biological Report 88(11). US Department of Interior, Fish and Wildlife Service, Fort Collins, Colorado.

Brooks, G.T. (1974) Chlorinated Insecticides, Vol. II. Biological and Environmental Aspects. CRC Press, Cleveland, Ohio.

Conner, K., Holbrow, A.M., Copeland, T.L., and Paustenbach, D.J. (1992) Use of Quantitative Uncertainty Analysis in Air Toxic Risk Assessment. Paper No. 92-166.04 presented at the 85th Annual Meeting of the Air and Waste Management Association, Kansas City, 21-26 June 1992. ChemRisk McLaren/Hart, Irvine, California.

Coster, J.E., and Searcy, J.L. (Eds.) (1981) Site, Stand and Host Characteristics of Southern Pine Beetle Infestation. USFS Technical Bulletin 1612. US Department of Agriculture, Forest Service, Washington, D.C.

Coulson, R.N. (1980) Population dynamics. In: Thatcher, R.C., Searcy, J.L., Coster, J.E., and Hertel, G.D. (Eds.) The Southern Pine Beetle, pp. 71-104. USFS Technical Bulletin 1631. US Department of Agriculture, Forest Service, Washington, D.C.

Evans, M.S. (Ed.) (1988) Toxic Contaminants and Ecosystem Health: A Great Lakes Focus. John Wiley, New York.

Gass, C.L., and Montgomeric, R.D. (1981) Hummingbird foraging behavior: Decision- making and energy regulation. In: Kamil, A.C., and Sargent, T.D. (Eds.) Foraging Behavior: Ecological, Ethological, and Psychological Approaches. Garland STPM Press, New York.

Graham, R.L., Hunsaker, C.T., O'Neill, R.V., and Jackson, B.L. (1991) Ecological risk assessment at the regional scale. Ecol. Appl. 1(2), 196-206.

Haack, R.A., and Blank, R.W. (1991) Incidence of two-lined chestnut borer and Hypoxylona tropunctatum on dead oaks along an acidic deposition gradient from Arkansas to Ohio. Proceedings, Eighth Central Hardwoods Forest Conference, University Park, Pennsylvania. General Technical Report 148. US Department of Agriculture, Forest Service, Northeastern Forest Experiment Station, Radnor, Pennsylvania.

Hidy, G.M. Mueller, P.K., Lavery, T.F., and Warren, K K. (1979) Assessment of regional air pollution over the Eastern United States: results from the Sulfate Regional Experiment (SURE). In: Proceedings of the WMO Symposium on the Long-Range Transport of Pollutants and its Relation to General Circulation Including Stratospheric/tropospheric Exchange Processes. WMO Report No.538. World Meteorological Organization, Geneva.

Hirsch, R.M., Alley, W.M., and Wiber, W.G. (1988) Concepts for a Natural Water-Quality Assessment Program. USGS Circular 1021. US Geological Survey, Federal Center, Denver, Colorado.

Holling, C.S. (1992) Cross-scale morphology, geometry and dynamics of ecosystems. Ecol. Monogr. 62(4), 447-502.

Hunsaker, C.T., Graham, R.L., Barnthouse, L.W., Gardner, R.H., O'Neill, R.V., and Suter II, G.W. (1990) Assessing ecological risk on a regional scale. Environ. Manag. 14, 325-332.

Krummel, J.R., Gardner, R.H., Sugihara, G., and O'Neill, R.V. (1986) Landscape patterns in a disturbed environment. Oikos 48, 321-324.

Laskowski, D.A., Igoring, C.A., McCall, P.J., and Swann, R.L. (1982) Terrestrial environment. In: Conway, R.A. (Ed.) Environmental Risk Analysis for Chemicals. Van Nostrand Reinhold, New York.

Loucks, O.L., Armentano, T.V., Foster, J., Fralish, J., Haack, R., Kuperman, R., LeBlanc, D., McCune, B., Pedorsen, B., and Somers, P. (in press) Pattern of air pollutants and response of oak-hickory ecosystems in the Ohio Corridor.

May, T.W., and McKinney, G.L. (1981) Cadmium, mercury, arsenic and selenium concentrations in freshwater fish. 1976-1977-National Pesticide Monitoring Program. Pest. Monit. J. 15, 14-38.

Mitchell, M.F., and Roberts, A.(1984) A case study of the use of fenitrothium in New Brunswick: The evolution of and ordered approach to ecological monitoring. In: Sheehan, P.J., Muller, D., Butter, G.C., and Bourdeau, P. (Eds.) Effects of Pollutants at the Ecosystem Level. SCOPE 22. John Wiley & Sons, Chichester, UK.

Nash, B.L., Davis, D.O., and Skelly, J.M. (1992) Forest health along a wet sulfate/pH disposition gradient in north-central Pennsylvania. Environ. Toxicol. Chem. 11, 1092-1104.

National Research Council of Canada (NRCC) (1977) Fentrothion: The Long-Term Effects of Its Use in Forest Ecosystems. NRCC No.16073. Associate Committee on Scientific Criteria for Environmental Quality, Ottawa, Canada.
NYSDEC (1986) New York State Air Quality Report Ambient Air Monitoring System. DAR- 87-1. New York State Department of Environmental Conservation, Albany.

O'Neill, R. V., Krummel, J.R., Gardner, R.H., Sugihara, G., Jackson, B., DeAngelis, D.L., Milne, B., Turner, M.G., Zygmutt, B., Christensen, S.W., Dale, V.H., and Graham, R.L. (1988) Indices of landscape pattern. Landscape Ecol. 1, 153-162.

Orians, G.H. (1980) General theory and applications to human behaviors. In: Lockhard, J.S. (Ed.) Evolution of Human Social Behavior. Elsevier, New York.

Otani, J., Dombrowski, F., Sheehan, P., Harman, C., and Brooke, K. (in preparation) A predictive ecological assessment of 30 chemicals emitted from a proposed hazardous waste treatment facility in east-central Mississippi.

Payne, T.L. (1980) Life history and habits. In: Thatcher, R.C., Searcy, J.L., Coster, J.E., and Hertel, G.D. (Eds.) The Southern Pine Beetle. USFS Technical Bulletin 1631. US Department of Agriculture, Forest Service, Washington, D.C.

Reith, P.B. (1987) Quantifying plant response to ozone: A unifying theory. Tree Physiol. 3, 63-91.

Rodhe, H. (1978) Budgets and turnover times of atmospheric sulfur compounds. Atmos. Environ. 12, 671-680.

Schmitt, C.J., Ludke, J.L., and Walsh, D. (1981) Organochlorine residues in fish, 1970-1974: National Pesticide Monitoring Program. Pest. Monit. J. 14, 136-206.

Schmitt C.J., Ribick, M.A., Ludke, J.L., and May, T.W. (1983) Organochlorine Residues in Freshwater Fish. 1976-1979: National Pesticide Monitoring Program. USFWS Resource Publ. No. 152.62. US Department of Interior, Fish and Wildlife Service, Washington, D.C.

Schmitt, C.J., Zajicek, J.L., and Ribick, M.A. (1985) National Pesticide Monitoring Program: residues of organochlorine chemicals in freshwater fish 1980-81. Arch. Environ. Contam. Toxicol. 14, 225-260.

Sheehan, P.J. (1984a) Effects on individuals and populations. In: Sheehan, P.J., Miller, D.R., Butler, G.C., and Bourdeau, P. (Eds.) Effects of Pollutants at the Ecosystem Level. John Wiley & Sons, Chichester, UK.

Sheehan, P.J. (1984b) Effects on community and ecosystem structure and dynamics. In: Sheehan, P.J., Miller, D.R., Butler, G.C., and Bourdeau, P. (Eds.) Effects of Pollutants at the Ecosystem Level. John Wiley & Sons, Chichester, UK.

Sheehan, P.J., Baril, A., Mineau, P., Smith, K.K., Hartenist, A., and Marshall, W.K. (1987) The Impact of Pesticides on the Ecology of Prairie Nesting Ducks .Technical Report Series No.19. Canadian Wildlife Service, Environment Canada, Ottawa, Canada.

Sheehan P.J., Baril, A., Mineau, P., and Paustenbach, D. (1993) Predicting the effects of insecticides on aquatic systems and the water fowl that use them: an ecotoxicological risk assessment. In: Rand, G.M., and Petrocelli, S.R. (Eds.) Fundamentals of Aquatic Toxicology: Methods & Applications. Hemisphere, Baltimore, Maryland.

Simmons, M.A., Cullinan, V.I., and Thomas, J.M. (1992) Satellite imagery as a tool to evaluate ecological scale. Landscape Ecology 7(2), 77-85.

Skelly, J.M. (1980) Photochemical oxidant impact on Mediterranean and temperate forest ecosystems: Real and potential effects. In: Miller, P.R. (Ed.) Effects of Air Pollutants on Mediterranean and Temperate Forest Ecosystems. General Technical Report PSW-43. Pacific Southwest Forest and Range Experimental Station, Berkeley, California.

Slade, J. (1990) Overview of sulfur and nitrogen oxide emissions in Pennsylvania. Proceedings. Atmospheric Deposition in Pennsylvania: A Critical Assessment. Pennsylvania State University, University Park, Pennsylvania.

Solomon, A.M. (1986) Transient repose of forests to CO₂ induced climate change: simulation modeling experiments in eastern North America. Ecologia 68, 567-579.

Suter II, G.W. (1989) Ecological Endpoints. In: Warren-Hicks, W., Parkhurst, B.R., and Bakers, S.S. (Eds.) Ecological Assessment of Hazardous Waste Sites. A Field and Laboratory Reference Document. EPA 600/3-89/013. Environmental Research Laboratory Corvallis, Oregon.

Suter II, G.W. (1993) Ecological Risk Assessment. Lewis Publishers, Boca Raton, Fla.

Swank, W.T., and Grossley, D.A. (Eds.) (1987) Forest Hydrology at Coweeta. Springer-Verlag, New York.

Taylor, O.C. (1973) Oxidant Air Pollution Effects on a Western Coniferous Forest Ecosystem. Statewide Air Pollution Research Center, University of California, Riverside.

Tucker, C.J., and Sellers, P.J. (1986) Satellite remote sensing of primary production. Int. J. Remote Sensing 4, 1395-1416.

Turner, M.G. (1987) Simulation of landscape changes in Georgia: A comparison of 3 transition models. Landscape Ecology 1, 29-36.

Turner, M.G., and Gardner, R.H. (Eds.) (1991) Quantitative Methods in Landscape Ecology. Springer-Verlag, New York.

Urban, D.L., O'Neill, R.V., and Shugart, H.H., Jr (1987) Landscape ecology. BioScience 37, 119-127.

USEPA (1989) Risk Assessment Guidance for Superfund, Vol. I, Human Health Evaluate Manual (Part A). Report No. EPA 510/1-89/002. US Environmental Protection Agency, Office of Emergency and Remedial Response, Washington, D.C.

USEPA (1991) National Air Quality and Emissions. Trends Report 1989. Report No. EPA 450/4-91/003. US Environmental Protection Agency, Office of Air Quality, Planning and Standards, Research Triangle Park, North Carolina.

USEPA (1992) Framework for Ecological Risk Assessment. Report No. EPA 630/12-92/001. US Environmental Protection Agency, Risk Assessment Forum, Washington, D.C.

USFWS (1980) Habitat Evaluation Procedures (HEP). US Department of the Interior, Fish and Wildlife Service, Division of Ecological Services, Washington, D.C.

Zannetti, P. (1990) Air Pollution Modelling. Van Nostrand Reinhold, New York.