Progress Report of the Ecological Committee on FIFRA Risk Assessment Methods: VII. Terrestrial Risk Assessment

On this Page

Table 1: Uncertainty Associated with Laboratory Toxicity Tests
Identified Uncertainty	Uncertainty Class	Function	Quantifiable	Reducible
Measurement error	Parameter error	Dose, diet preparation Endpoint measurement	Yes (not usually done)	Somewhat
Inter-test variation	Natural stochasticity	Environmental conditions Stock of test animals Condition of animals Inherent sensitivity differences among test animal populations	Maybe (historical data bases of LD50s and slopes)	Somewhat
Lack of fit of statistical model to data	Natural stochasticity	Limitations of test design involving limited number of doses and animals	Yes (confidence intervals)	Possible (not feasible?)
Intra-specific variation (within test population)	Natural stochasticity	Inherent differences in sensitivity among individuals of the test population	Yes (slope of dose-response)	No
Behavioral and physiological responses	Natural stochasticity or model error	Food consumption measurement Repellency Regurgitation	Yes	Somewhat (qualitative observations; quantify residues)

Table 2: Uncertainty Associated with Extrapolations
Identified Uncertainty	Uncertainty Class	Function	Quantifiable	Reducible
Interspecific variability	Natural stochasticity	Inherent differences in sensitivity and behavior Inherent differences in slope of dose-response	Yes (historical data of LD50s and slopes)	Somewhat
Extrapolations to no- or low- effect levels	Model error	Test designed to estimate median lethal dose or concentrations Limitations of design, few doses or animals	Some cases Some cases	Somewhat Somewhat
Extrapolation across taxonomic classes (e.g., mammal for avian)	Natural stochasticity and model error	Inadequacy of animal model	Maybe (if historical data or pharmacokinetic models justify extrapolation)	Possible (not feasible?)
Intra-specific variability	Natural stochasticity	Differences in captive and wild animals	Maybe	Yes?
Extrapolation to field scenarios	Model error	Differences in exposure (e.g., differences in feeding patterns) Differences in sensitivity	Maybe Not likely	Yes No
Extrapolation to populations or communities	Natural stochasticity or model error	Population dynamics (e.g., density dependence) Indirect effects	Maybe via models Limited theoretical and empirical basis	Yes, if measurements occur at each level of organization