|
Items in cart [0] |
Questions? Call 888-624-8373 |
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 259
Issues in Risk Assessment
4
Key Scientific Problems Limiting Application of Ecological Risk Assessment
EXTRAPOLATION ACROSS SCALES
The most common scientific limitation exemplified in the case studies is the problem of extrapolating across scales of space, time, and ecological organization. For the most part, scientific data related to a specific stressor are limited to what can be obtained in a controlled laboratory setting or in a limited field study. Observations of environmental contamination and ecological effects of tributyltin were limited to a few marinas. Testing of pesticides even in the best of circumstances is limited to small field plots and carefully controlled applications. Table 4-1 shows, for all the case studies, the scales at which the data used in the assessments were collected and the scales of interest in decision-making. In most cases, the scales of interest in decision-making are substantially larger in space and of longer duration than could be accommodated in any practical assessment effort. Some form of extrapolation, either with explicit mathematical models or with judgment-based decision rules, is necessary to make the risk assessments useful for decision-making. The PCB study discussed by Di Toro (Appendix E) clearly illustrated the value of explicit models for estimating recovery times in response to hypothetical management actions. In the pesticide registration process described by Kendall (Appendix E), extrapolation is based primarily on qualitative evaluation of test data and information on expected use patterns. Kendall argued that models of ecological effects of pesticides are needed to reduce uncertainty and to account for effects
OCR for page 260
Issues in Risk Assessment
TABLE 4-1 Scales of Observation and Management in Case Studies Evaluated by the Committee
Case Study
Observational Scale
Management Scale
Spatial
Temporal
Spatial
Temporal
Tributyltin
< 1m3
< 1 yr
Chesapeake Bay
> 5 yr
~ 1 ha
(laboratory)
< 5 yr (field)
Agricultural chemicals
~ 1 ha
< 1 yr (field test)
Agricultural region
> 5 yr
PCB and TCDD
< 1 L
~ 1-month (laboratory)
Lakes or rivers
> 10 yr
Spotted owl
~ 300 km2
< 6 yr
Pacific northwest
> 100 yr
Species introduction
< 100 m2
~ 1 yr (greenhouse)
Agricultural region
>> 1 yr
Georges Bank
~ 104 km2
last 30 yr
~ 104 km2
next 5 yr
OCR for page 261
Issues in Risk Assessment
that cannot be directly measured in test systems. Extrapolation was not explicitly considered in the paper by Anderson (Appendix E). Other published literature attempts to relate spotted owl abundance to regional patterns of old-growth forest harvesting (Salwasser et al., 1986; Lande, 1988). Bartell et al. (1992) have recently discussed the problem of estimating population and ecosystem-level effects of toxic contaminants from laboratory toxicity tests. A variety of approaches are now being developed for extrapolating local-scale disturbances (e.g., fires or insect outbreaks) to regional-scale changes in landscape patterns (Costanza et al., 1990; Turner and Gardner, 1991; Graham et al., 1991). Extrapolations from spatial and temporal scales suitable for rigorous experiments and observations to scales relevant to environmental management appear to be essential for adequate characterization of ecological risks.
QUANTIFICATION OF UNCERTAINTY
Formal analysis of uncertainty is another major subject for improvement in ecological risk assessments of all types. The ''uncertainties" discussion group at the workshop identified three general categories of uncertainty that affect all types of risk assessments:
Measurement uncertainties, e.g., low statistical power due to insufficient observations, difficulties in making physical measurements, inappropriateness of measurements, and natural variability in organic responses to stress.
Conditions of observation, e.g., spatiotemporal variability in climate and ecosystem structure, differences between natural and laboratory conditions, and differences between tested or observed species and species of interest for risk assessment.
Inadequacies of models, e.g., lack of or knowledge concerning underlying mechanisms, failure to consider multiple stresses and responses, extrapolation beyond the range of observations, and instability of parameter estimates.
Measurement uncertainties can be reduced by making more and better measurements. Uncertainties related to conditions of observation cannot
OCR for page 262
Issues in Risk Assessment
be reduced, but often they can be quantified using empirical regression techniques (Suter et al., 1983), time series analysis (Jassby and Powell, 1990), or formal model uncertainty analysis (Bartell et al., 1992). Di Toro and Fogarty et al. provided examples of model uncertainty analyses in their case study papers (Appendix E). Uncertainties related to inadequacies of models (or scientific ignorance in general) are much more difficult to quantify.
Choices between risk assessment methodologies often involve tradeoffs between different types of uncertainty. For example, decisions about the need for pesticide testing are now based on qualitative evaluation of toxicity and exposure data (Urban and Cook, 1986). Explicit models of the effects of toxicant exposure on the abundance and persistence of bird populations have been developed (Grier, 1980; Tipton et al., 1980; Samuels and Ladino, 1983) and could be used to quantify uncertainties related to variability in exposures or extrapolation from field plots to natural landscapes. Relying on expert judgment avoids the need to postulate particular mechanisms of exposure or complex population dynamics but prevents risk assessors from providing information on the value of collecting additional information to reduce uncertainties or providing information on the ecological costs and benefits of regulatory decisions. Using a model to quantify uncertainties would in principle permit more useful risk assessments, but if the model itself is a poor representation of reality, the results might be totally meaningless.
The committee believes that improvements are needed in techniques for qualitative and quantitative analysis of uncertainty for ecological risk assessment. Techniques for model uncertainty analyses developed by systems engineers have been used by ecologists for more than a decade (Gardner et al., 1981; Bartell et al., 1992; Di Toro, Appendix E). The large and growing technical literature on decision analysis (Raiffa, 1970; Lindley, 1985; Von Winterfeldt and Edwards, 1986) has been much less thoroughly exploited (see Walters (1986) and Reckhow (1990) for examples of ecological applications of Bayesian decision theory) and should be surveyed for potentially useful approaches.
VALIDATION OF PREDICTIVE TOOLS
Improvements in the mathematical models, qualitative and quantitative
OCR for page 263
Issues in Risk Assessment
decision rules, and other predictive tools used in ecological risk assessment still are needed. Although the committee refers to the process of improvement as "validation," we recognize that none of the approaches in question can be proved fully valid in the sense of perfectly predicting natural ecosystem behavior under all circumstances. The purpose of validation is to improve the credibility and reliability of predictive methods. Validation must be viewed as an iterative process in which predictions are tested, models are refined, and then new predictions are tested.
At least three kinds of studies can contribute to validation: improved measurements of specific quantities and tests of assumptions, experimental testing of models under reasonably realistic conditions (e.g., ponds or enclosures), and monitoring of ecological effects of dams, power plants, or other projects to determine the accordance between actual effects and effects predicted before construction or operation of the dams. Each kind of study has its own advantages and disadvantages, and all three should be included in validation programs.
This committee is not the first to note the need for validation studies. Similar recommendations can also be found in at least two previous NRC reports (NRC, 1981; NRC, 1986). In spite of virtually unanimous support within the scientific community for this activity, the resources currently being expended for improvement of predictive tools are much smaller than those devoted to repetitive assessments and routine monitoring of compliance with permit requirements. The importance of enhanced validation programs needs to be recognized by all regulatory and resource management agencies.
VALUATION
Valuation and cost-benefit analyses are recognized as integral components of the risk management process. Such analyses contribute to the regulations that provide the context for risk assessments and to the eventual risk management decisions. Cost-benefit analyses are major parts of the planning and ranking process within and between agencies. Ecological cost-and-benefit analyses have gained acceptance where individual behavior can be used to directly reflect economic preferences, e.g., recreational use and associated travel-cost analysis (Yang et al., 1984; DesVouges and Skahen, 1985). These analyses have also been
OCR for page 264
Issues in Risk Assessment
applied with some success where people's direct use of a resource was the specific issue, e.g., dam construction vs. maintenance of the natural river channel. Methods also have been developed for monetizing ecological values beyond those associated with the use of a resource, but the uncertainties associated with applications of those methods are often quite high.
Ecological values are sometimes described by resource economists as services provided by the environment to humans. Such economists categorize economic values into two segments termed "use" and "nonuse" values. As noted above, reasonably reliable techniques are available for determining use values (e.g., land valuation and recreational use) from the actual behavior of resource users. Several methodologies have been developed to date for estimating nonuse values. For example, contingent valuation uses public surveys to elicit statements of how much an individual hypothetically would be willing to pay for improvements (or to prevent reductions) in the quantity or quality of natural resources. It requires people to assign subjectively economic values for environmental goods. Recent empirical research indicates that the results vary depending on the way the assessments are elicited (Opaluch and Segerson, 1989; Grigalunas and Opaluch, 1991; Hausman, 1991; Rosenthal and Nelson, 1992), and the resulting values must be interpreted with care.
Clearly, a considerable need remains for increased communication and clarification between ecologists and economists to improve the use of valuation methods in ecologic risk management decisions. There is already a substantial literature on the economic value of wetland ecosystems (Scodari, 1990). Valuation of other kinds of ecosystems is being actively discussed (e.g., Orians, 1990), but generally accepted principles for ecosystem valuation do not yet exist.
Representative terms from entire chapter:
risk assessments
