ECOFRAM Update: Probabilistic Risk Assessment Tools for Pesticides
On this Page
Development of new probabilistic assessment tools for use in the pesticide regulatory process will be the focus of a poster-discussion session at the 1998 Annual Meeting in Charlotte. This session will present near final recommendations of the ECOFRAM (Ecological Committee On FIFRA Risk Assessment Methods) for improving ecological risk assessments that the US EPA conducts under FIFRA.
In May 1996 the Environmental Fate and Effects Division (EFED) of the Office of Pesticide Programs (OPP) presented two pesticide risk assessment case studies to EPA's Scientific Advisory Panel (SAP) and asked them to address the Agency's current pesticide risk assessment methodology. The SAP commented that while the current process is believed to be cautious and protective in terms of adverse environmental effects, it best serves as a screen because it provides little information on the likelihood of damage. The SAP recommended that the pesticide risk assessment process be expanded to include probabilistic assessments of risk and to identify the uncertainties associated with the assessment.
The ECOFRAM was formed in June 1997 in response to recommendations from the SAP. The underlying charge to the ECOFRAM was to develop methods and tools, within the context of FIFRA regulatory perspective, for predicting the magnitude and probabilities of adverse effects to non-target aquatic and terrestrial species resulting from the introduction of pesticides into their environment. The ECOFRAM was instructed to first address direct acute and chronic effects of pesticides on individuals and populations of terrestrial vertebrates and aquatic vertebrates and invertebrates. Probabilistic techniques should use existing fate and effects data when possible, but ECOFRAM was free to make recommendations for improving data quality.
To insure adequate representation of all interested parties, EFED invited representatives from the US EPA, US Department of Agriculture (USDA), Environment Canada, American Crop Protection Association (ACPA), non-government environmental organizations, academia, and environmental consulting groups to participate in ECOFRAM. A Workshop Organizing Committee structured the ECOFRAM such that its 48 members are evenly divided into aquatic and terrestrial areas of emphasis. The two workgroups are each subdivided into areas of exposure and effects characterization. These workgroups have been meeting monthly for the last year and expect to have a draft report completed by the end of 1998.
The terrestrial workgroups began their efforts by discussing the basic problem addressed in a pesticide risk assessment. Extensive discussions centered on defining the sources of uncertainty associated with assessing effects and exposure, the effects of phylogenetic relationships among species sensitivities, the effects of behavior on exposure, the incorporation of non-lethal endpoints, statistical considerations such as the number of species represented or safety factors, the effect of landscape, and the suitability of current test designs.
The exposure workgroup is focusing upon two general areas: characterizing routes of exposure and pesticide residues. Monte Carlo simulation modeling approaches are being developed for estimation of exposure through the diet (food) and ingestion of pesticide granules. Additionally, a random walk model is being developed for simulating exposure resulting from avian foraging scenarios in an agricultural landscape that includes a mosaic of pesticide-treated and untreated habitats. Both models will provide estimates of the probability of occurrence for incorporation into effects characterization models. In conjunction with the foraging models, additional work is being undertaken to develop models for predicting pesticide residues in terrestrial compartments which utilize dissipation kinetics to a greater extent than current models.
The effects workgroup is evaluating a tiered approach to combining exposure and effects information in risk assessments. This tiered approach begins with distribution-based quotients, and becomes more complex as the situation demands using single distribution and point estimates, joint distributions, single distribution and effect response function and population level assessments. Single distribution and effect response function models being evaluated include hazard analysis, logistic function of dose, and a modified probit model for estimating the confidence intervals around mortality observed in acute toxicity tests. Population models being evaluated include age class structure models, stage and size structured models, combined age and size structured models and individual-based models.
The aquatic workgroups began their efforts by discussing the basic problem addressed in a pesticide risk assessment and then began defining the sources of uncertainty for assessing effects and exposure. A conceptual framework was developed that incorporated characteristics of chemicals, agricultural landscapes, and aquatic ecosystems that interact to influence exposure. From this framework, the workgroups designed a tiered system for the risk assessment process in which the assessment progresses from a deterministic assessment to probabilistic assessments of increasing complexity. Each tier includes several experimental and analytical options that reduce uncertainty and provide more complete descriptions of the aquatic environment.
The exposure characterization consists of tiers that progress from a simple farm pond to a watershed analysis that incorporates farm ponds and streams, wetlands, or estuaries where appropriate. This tiered structure will provide outputs that progress from a single point estimate of exposure to a multi-dimensional probability surface showing regional and temporal variation in aquatic exposure.
The effects characterization at the lower tiers consists of the standard battery of acute and chronic ecotoxicology tests. For higher tiers, the effects workgroup evaluated a number of analytical and experimental tools for their suitability in probabilistic risk assessments and as a means to reduce uncertainty in the risk assessment. The tools include time-to-event analysis, analysis of time-varying exposures, analysis of species sensitivity distributions, population models, sediment toxicity, chronic and behavioral tests, and model ecosystems (microcosms or mesocosms). Initial efforts of aquatic effects characterization will concentrate on the population and community, emphasizing endpoints such as magnitude of population effect and number of fish and invertebrate species affected. The proposed tiered aquatic risk assessment approach includes several probabilistic techniques to integrate the exposure and effects characterizations. The process is structured to include discussions with risk managers and analysis of potential mitigation decisions after the risk is characterized. The decision may be to carry the risk assessment process to the higher tiers, with more complex analyses, for the endpoints for which risk remains high, with the objective to reduce uncertainty and lessen the need for use of conservative assumptions.