Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
The Role of Ecological Risk Assessment in Environmental Decision Making ALAN W. MAK] Exxon Company, USA MICHAEL W. SLIMAK U.S. Environmental Protection Agency PREDICTIVE ASSESSMENT OF ECOLOGICAL RISKS Potential impacts and effects of pollutants on ecosystems are often complex and far-reaching. The simplest counter to a chemical use that produces immediate and obviously deleterious environmental effects is to impose a ban on its use before the effects become irreversible. However, many of the less obvious and subtle environmental effects of pollutants are not readily recognized and therefore are considerably more difficult to handle. Thus, the effect of pollutants on ecosystems is an issue which requires the development and enhancement of methodology to predict the fate and effects of substances in the environment prior to their manufacture, use, or distribution. Procedures for ecosystem risk assessment are currently evolving. An ecosystem nsk assessment is defined as a set of procedures for measuring risk to the environment associated with the use of substances through an objective and probabilistic exercise based on empirical data and scientific judgment. The results of such a risk assessment can then be used to provide a consistent means of estimating the limiting concentrations of substances that will produce no unacceptably negative effects on ecosystems which are potentially exposed to the substance. Thus, ecological risk assessments serve as the scientific basis for decid- ing whether the risks are acceptable or unacceptable for the environment, i.e., a risk management decision. Basically, the risk assessment process consists of two parallel lines of investigation and relates observed biological effects to expected exposure concentrations. Figure 1 represents the two 77
78 ECOLOGICAL RISKS c - In . ~ ~c. ~ Highest test concentration / producing no biological effects. - - <°ofj~e `~/s '~ Flighest expected environmental concentration . 1 5 6 Sequential Tests of Hazard Assessment Procedure FIGURE 1 Diagrammatic representation of a sequential hazard assessment procedure demonstrating increasingly narrow confidence limits for estimates of no biological effect concentration and actual expected environmental concentration. concentrations as parallel lines and demonstrates that increasingly more ac- curate and statistically reliable estimates of these concentrations will result from a sequential series of tests completed along the x axis of the graph. Using the risk assessment process, increasingly more accurate estimates of fate and effects can be made to the point where it becomes possible to state with a high degree of confidence that environmental concentrations and biological effects will likely result in negative environmental conse- quences. It is a matter of judgment to determine just how far into the hazard-evaluation process investigation must proceed to establish accept- able confidence regarding fate and effect concentration predictions (Cairns et al., 1978~.
ENVIRONMENTAL AL4NAGEMENT CONCEPTS DETERMINE THE CHEMICAL AND PHYSICAL PROPERTIES EST1~1ATE ENVIQON~tENTAL CONC£N TRATIONS 79 PREDICT USAGE PATTERNS AND QUANTlilES 1 . ~ TEST FOR TEST FOR ENVIRONMENTAL ~ ENVIRONMENTAL FATE EFFECTS ~ - . DECISIONS DECISIONS ESTIMAtE HUI~AN EXPOSURES _ _ TEST FOR HEALtH EFFECTS my' 0691SE RESTRICTIONS TO DIMINISH EXPOSURES \ MONITOR | DO NOr USE FIGURE 2 Flow chart for the 10 phases of the safeW evaluation program. The process of making and reviewing an environmental risk assessment can be broken down into ten stages (Beck et al., 1981~. These stages and how they are related to each other are shown in Figure 2. The sequence of stages splits into two paths, one for environmental safety and one for human safety, which come together again at the point of decision making. Each stage is described briefly below: 1. The evaluation starts with a consideration of the properties of the substance (e.g., chemicals, materials) ascertained from the literature or determined in the laboratory, if necessary. Estimates may suffice in
80 ECOLOGICAL RISKS the early stages, but more precise determinations will generally be needed later. This is the time for initial development of the analytical methodology that will be required to measure the substance in the environment and in biological tissues. 2. Use of the substance, and quantities involved must be considered, .. since usage patterns (along with such related matters as manufacturing, shipping, and disposal) influence the routes and amounts of environmental exposure. 3. From anticipated usage patterns, the concentrations that can be expected in various environmental compartments are predicted. These estimates can be helpful in projecting exposures, but their chief value is in suggesting how extensively the material should be tested for environmental properties, e.g., the rate of chemical and biological degradation. Then, from the results of the environmental fate tests, more refined predictions of environmental concentrations can be made. These estimates are necessary both for predicting exposures- and for interpreting the results of tests to evaluate environmental effects. 4. Tests for environmental fate suggest what may happen to the sub- stance after it is released into the environment. The nature of both the substance and the potentially affected ecosystem is important in determin- ing fate. This kind of information is necessary for making the refined estimates of environmental concentration referred to above, and this is one of the factors influencing the estimate of exposure. 5. Estimates of exposure can be made from information about man- ufacturing, transport, and usage patterns and from the estimates of envi- ronmental concentration. A number of estimates are needed to cover such diverse situations as direct exposure to pesticides or indirect exposures from a contaminated food source. Exposure may be by oral, respiratory, ocular, or dermal routes. These estimates of exposure are of value in determining which effects tests should be conducted, and are essential in evaluating the results of those tests. 6. Tests for hazard are concerned with the possible effects of the material on non-target health, but most of the tests are conducted with laboratory animals. From information about the kinds of effects produced in these animals, and the concentrations of chemicals necessary to produce them, it is possible to determine the exposure of chemicals that would be acceptably low risk to non-target organisms. 7. Tests for environmental effects indicates whether the concentra- tions expected may cause harm to the environment, particularly to the living organisms in it, the kinds of injury that may occur, and the species or processes most likely to be affected. Further discussion of ecological responses to stress can be found in subsequent chapters by Howell et al. and Breymeyer (Chapters 7 and 8, this volume).
ENVIRONMENTAL MANAGEMENT CONCEPTS 81 8. Decision making occurs repeatedly throughout the process. If the decision is to use the material, field monitoring may be needed to determine whether the resulting environmental concentrations correspond to those that were predicted from laboratory testing and modelling. 9. The first step in decision making is to compare the concentra- tions of a chemical that are predicted to cause unacceptable harm to the environment with the concentrations that will likely result from using the chemical. From this comparison, it is possible to assess the risk of causing adverse effects. Finally, the decision is made whether the risks are soci- etally acceptable or not. It is not necessary to do all the tests that are listed before making this decision; there are provisions at many points in the process for deciding that no further testmg is necessary and that regulation is or is not required. 10. If any of the risks associated with using the chemical as originally planned are judged unacceptable, it may be possible to devise restrictions on the use of the chemical that would diminish the anticipated exposure and therefore lower the risk. The restrictions might be of many sorts, ranging from warning labels to the construction of containment dikes around storage tanks. Once such restrictions have been devised, it will then be necessary to go through parts of the decision-making process again to see whether the risk becomes acceptable. ECOTOXICOLOGY: THE PRACTICE OF RISK ASSESSMENTS Ecotoxicology can be defined as the study of the fate and effects of toxic agents in ecosystems. Ecotoxicology is the study of toxic effects on biota particularly on populations and communitiesand their interactions with processes controlling the functioning of defined ecosystems. The U.S. Environmental Protection Agency (EPA) has primary re- sponsibility for determining the ecological risks associated with the use of pesticides and industrial chemicals (xenobiotics) in the United States. This responsibility comes from two major pieces of legislation: the Toxic Sub- stances Control Act DISCO) and the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA). In practice, both programs assess risks to ecological resources using an ecotoxicological approach: laboratory toxicity bioassays to determine hazard; determination of exposure either from monitoring data or predicted from models; and a comparison of exposure to hazard using the quotient method. In the quotient method, the exposure value is directly compared with a toxicity endpoint (e.g., concentration in water to an LC50 value; 10 ppm/100 ppm). The closer the quotient is to 1 (or greater), the higher the probability that an adverse effect will occur. Interpreting this adverse effect (i.e., the likelihood that what is observed in the lab will actually occur in
82 ECOLOGICAL RISKS the field) is one of the biggest uncertainties in both programs. Although each program derives it differently, the final result is the application of either a safety factor or an assessment factor to account for uncertainty. Although the two programs are similar in their approach to assessing ecological risk, the quantity of data used to make assessments is strik- ingly different: TSCA assessments tend to be data-poor while FIFING assessments are usually data-rich. This difference is due to what is being regulated. TSCA regulates new and existing industrial chemicals, many of which are site-limited intermediates, are produced in low volumes with limited potential for environmental release, and are not manufactured pri- marily for their toxicological (biological) properties. Pesticides, on the other hand, are generally very active biologically, are designed to kill pests and other organisms, and are broadly released into the environment on a regional scale. The Toxic Substances Control Act The Office of Toxic Substances (OTS) at EPA is responsible for im- plementing provisions of TSCN TSCA was enacted in 1976 to protect humans and the environment from unreasonable risks caused by industrial chemicals and mixtures. Section 4 of TSCA is primarily concerned with existing chemicals which were in production before the law was passed. Because of the thousands of chemicals on the inventory, a prioritization scheme had to be established to effectively assess potential risks. Thus, an Inter-Agengy Testing Committee recommends to EPA chemicals it believes should be given priority testing for hazard determination. EPA evaluates these recommendations and can either refute them or proceed with hazard testing. Chloroparaffins are an example of a group of chemicals for which ecological testing has been required pursuant to this section of TSCN OTS used and refined existing ecosystem-level models to address the risks to natural populations posed by indirect and direct toxic effects of these chemicals. The use of these higher level models was necessary since the quotient method of assessment indicated a likelihood of adverse effects. Section 5 of TSCA covers chemicals which are new, i.e., those which were produced since the law was passed. Industrial chemicals regulated in this section are referred to as Pre-Manufacturing Notice chemicals or "PMNs." This section of the law requires a manufacturer to submit a pre- manufacture notice to EPA before it manufactures a particular chemical. EPA then has 90 days (or, with good cause, 180 days) to review the notice for potential risks to the environment. There were over 2,000 PMNs submitted to EPA in 1988. Generally, however, ecotoxicological data is not submitted by the manufacturer and the potential hazard must therefore be determined using predictive toxicological techniques (e.g., structure/activi~
ENF7RONMENTAL MANAGEMENT CONCEPTS 1 83 relationships) techniques. If an unacceptable risk is judged to be a high probability (risk = hazard + exposure), then the manufacturer may be asked to conduct suitable tests. Because of the large numbers of industrial chemicals that must be assessed in this program, a method was devised to insure uniformity and consistency in identifying chemicals for testing to determine ecological haz- ard. Assessment factors are used in conjunction with the hazard assessment to derive concentrations of concern in aquatic media which, if equaled or exceeded, provide a basis for further testing. Assessment factors are num- bers which are used to adjust standard toxicological measurements (e.g., LC50, EC50, etc.) to derive a "concern level." An environmental concen- tration of concern is that concentration at which populations of organisms are adversely affected as found in a field study conducted under simulated or actual conditions of production, use, and disposal. Assessment factors take into account the uncertainties due to such variables as test species sensitivity to acute and chronic exposures, laboratory test conditions, and age-group susceptibility. There are four assessment factors currently being used: 1, 10, 100, and 1000. Able 1, taken from EPA (1984), summarizes the application of assess- ment factors. OTS does not consider assessment factors to be equivalent to safety factors. Safety factors are usually interpreted as being a margin of safety applied to a no-observed-effect level to produce a value below which exposures are presumed to be safe. Assessment factors are used with acute or chronic toxicity values to arrive at a concentration which if equaled or exceeded could cause adverse effects. Assessment factors have been developed solely for the PMN process to identify those chemicals which require ecological testing to fully assess ecological risks. In assessing risks to PMN chemicals, OTS uses the quotient or ratio method. The specific equation used is: Environmental Concentration/Concern Level = Risk If the quotient is equal to or greater than 1, the conclusion is that ad- verse effects are likely to occur to the population of organisms represented by the toxicity data. The quotient method is only used, however, to de- termine if actual testing is necessary. If actual hazard data is obtained, the quotient method is still used; however, more analysis is conducted using dose/response curves in conjunction with the measured or predicted environmental concentration. In addition, consideration is given to un- derstanding the acute to chronic ratios, and inter- and intra-taxa dose relationships. In some instances, simulation models, such as the Standard Water Column Model (SWACOM) developed by Bartell et al. (1988), have been employed where the chemical impact on one trophic level is analyzed relative to the other trophic levels.
84 ECOLOGICAL RISKS TABLE 1 Application of assessment factors to evaluate need for testing. - DATA AVAILABLE ASSESSMENT FACIOR TO BE APPl IFD Structure -Activity Denved LCSo Single LC50 Frown Gh~xnical Analog Single Test LC'o for PMN Two LC';os for Same Analog (e.g., 1 Fish, 1 Algal test) Two LC50s for PMN (e.g., 1 Fish test, 1 Invertebrate) Three Loos for Same Analog (Fish, Algae, Invertebrate) Five LC50s for Same Analog (3 ~vertebrates, 2 Fish) Five LC50s for the PMN (e.g., 3 Algae, 2 Fish) Maximum Acceptable Toxic Concentration for Analog Field Study 1000 1000 1000 1000 1000 100 100 100 10 SOURCE: U.S. EPA, 1984. Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) Under this law, EPA must determine whether a pesticide can be registered for a particular use. FIFRA states that the EPA Administrator shall register a pesticide if he determines that '~when used in accordance with widespread and commonly recognized practice it will not generally cause unreasonable adverse effects on the environment." The term "unreasonable adverse effects on the environment" means any unreasonable risk to humans or the environment, taking into account the economic, social, and environmental costs and benefits of the use of any pesticide. Under FIFRA, the process of determining whether or not a risk is unreasonable (i.e., factoring in benefits along with risks) is a risk management function. For this discussion, the important term used in FIFRA is "risk to the environment." In order for the EPA Administrator to determine if there will be an unreasonable risk to the environment from the use of a pesticide, an ecological risk assessment from a pesticides perspec- tive involves estimating the likelihood or probability that adverse effects (e.g., mortality to single species of organisms; reductions in populations of non-target organisms due to acute, chronic, and reproductive effects; or disruption in community and ecosystem level functions) will occur, are occurring, or have occurred. Ecological risk is a function of toxicological hazard and environmental exposure. Toxicological hazard is the intrinsic quality of a pesticide to cause an adverse effect under a particular set of circumstances. Toxicological
ENVIRONMENTAL MANAGEMENT CONCEPTS TABLE 2 Regulator nskassessm~tcnmna for "sucides. 85 PRESUMPTION OF ~ K PRESUMPTION THAT PLAY BE OBLIGATED PRESUMPTION OF OF NO RUSK BY RESTRICTED USE UNACCEPTABLE RUSK , I. Acute Toxicity 1) Manuals EEC* < 1/5 LC50 EEC 2lJ5 LCso EEC > LC50 mg~day < 1/5 LC50 May> 1/5 LCso 2) B`r~ EEC <1/5 Leo 1/5 LCso~ EEC < LCso EEC 2 LC50 3) Aquatic Organisms EEC <1/10 LCSo EEC 2 1/10 LC II. Chronic Toxicity 1110 LC50~ EEC EEC 21~ LCSo ~ 1/2 LCso EEC < Chronic N/A EEC 2 Chronic Effect Levels No Effect Level Including Reproductive Effects *EEC = Expected Environmental Concentration. This is typically calculated using a series of simple Homographs to complex exposure models. SOURCE: Adapted from Urban and Cook, 1986. hazard data includes, for example, laboratory fish, aquatic invertebrate, or bird LC50 values, and effect levels for fish and avian reproduction tests. Environmental exposure is a function of two data components. The first is the estimated amount of the pesticide residue that will be in the environment and available to non-target organisms. The second consists of the numbers, types, distribution, abundance, dynamics, and natural history of non-target organisms which will be used in contact with these residues. Information on the proposed label use of the pesticide is essential for such exposure estimates. Toxicological hazard is estimated first and environmental exposure separately; then they are compared to each other. In the EPA Pesticides Program, the comparison of exposure with effects data is based on regulatory risk criteria. These criteria are summarized and presented in Able 2. Within the table are risk criteria which contain specific safety factors that were derived from a toxicological model developed by the Program in 1975. The model was designed to provide a safety factor that would allow for differential variability and sensitivity among fish and wildlife species. A detailed explanation of the derivation of these safeW factors is found in Urban and Cook (1986~. Many theoretical questions can be raised about the use of risk criteria and safety factors in general. Currently, the Program does not use the model to predict the probability of the pesticide causing significant acute adverse effects to non-target organisms, since the model does not provide
86 ECOLOGICAL RISKS a mechanism for estimating model uncertainty. Thus, the risk criteria with their safety factors are used as "rough" estimates of potential risk to non-target organisms. Specific information and testing data are necessary in order to conduct an ecological risk assessment for a pesticide. Under FIFRA, EPA is not responsible for producing the data needed to make an ecological risk assessment. That burden is placed upon the applicants for registration. The Office of Pesticides Programs has published regulations which specify the data that are required for registration (40 CFR, Part 158 as cited in Urban and Cook, 1986y, and guidelines which provide recommended testing methods that are needed to produce the required data. The EPA Pesticides Program follows four procedural steps-in assessing ecological risk: 1. review and evaluation of hazard data to identify the nature of the hazards; 2. identification and evaluation of the observed quantitative relation- ship between dose and response; 3. identification of the conditions of exposure (e.g., intensity, fre- quency, and duration of exposure); and 4. combination of dose/response and exposure information to derive estimates of the probability that hazards associated with the use of the chemical will be realized under conditions of exposure experienced by the non-target populations) under consideration. These steps result in the comparison of toxicological hazard data and exposure data using regulatory risk criteria. Typically, toxicological hazard data may consist of acute LD50 and LC50 values, or chronic no-effect levels for the most sensitive indicator species. Exposure data normally consist of model-based estimated environmental concentrations in important media of concern (i.e., water, soil, and non-target organism food items). As the ratio of these input data equals or exceeds the regulatory criteria, a risk is inferred. CONCLUSION Both the toxic substances and pesticides programs at EPA recognize that the ratio method for assessing risk has numerous weaknesses. For example: . · it does not adequately account for effects of incremental dosages; it does not compensate for differences between laboratory tests and field populations; · it cannot be used for estimating indirect effects of toxicants (e.g., food chain interactions);
ENVIRONMENTAL MANAGEMENT CONCEPTS 87 · it has an unknown reliability; · it does not quantify uncertainties; and · it does not adequately account for other ecosystem effects (e.g., predator/prey relationships, community metabolism, structural shifts, etch. Therefore, at the present time, the state-of-the-art does not provide a com- plete characterization of the magnitude of risk or the degree of confidence associated with the characterization. The development of the field of ecotoxicity, like that of risk assessment, has paralleled the increased awareness of the environment during the past two decades. This awareness is especially evident now in Eastern Europe. The science is complex and addresses a broad range of issues that are frequently the focus of public concern and international policy. ~day, many ecological risk assessment protocols are modifications of methods used to characterize risk to public health. Unfortunately, these methods often lack environmental validity and may not effectively measure ecosystem integrity. New directions in this field reflect an increased emphasis on the role of sediments, biomarkers, and ecosystem assessments in controlling environmental contaminants. As these methods and protocols become more refined and are practiced by a larger body of scientists, their role and importance in environmental decision making will become much more important. REFERENCES Bartell, S.M. 1988. Community and ecosystem models for ecological risk assessment. Oak Ridge National Laboratory, Environmental Sciences Division, Oak Ridge, Tennessee. Beck, LOO., A.W. Maki, N.R. Artman, and E.R. Wilson. 1981. Outline and criteria for evaluating the safety of new chemicals. Reg. Tax. and Pharmacology (1~:19-58. Cairns, J. Jr., ILL" Dickson, and A W. Maki, eds. 1978. Estimating the hazard of chemical substances to aquatic life. ASTM STP 667. American Society for Testing and Matenals, Philadelphia, Pennsylvania. Urban, DJ., and NJ. Cook. 1986. Hazard Evaluation Division, Standard Evaluation Procedure, Ecological Risk Assessment 504/9-85 001; NTIS PD86-247457. U.S. Envi- ronmental Protection Agency, Washington, D.C. U.S. EPA. 1984. Estimating concern levels for concentrations of chemical substances in the environment. U.S. Environmental Protection Agency, Washington, D.C.
