|
(21-07) Uncertainty in the determination of ecological no-effect concentrations. Roelofs, J.J.W.*,1,2, Huijbregts, M.A.J.2, Jager, T.1, 1 National Institute of Public Health and the Environment (RIVM), Bilthoven, The2 Nijmegen University, Nijmegen, The ABSTRACT- A probabilistic method is proposed to predict no-effect concentrations (PNECs) for aquatic ecosystems. Assessment factors will be used to extrapolate acute toxicity data to their chronic equivalents and to extrapolate to a concentration at which the no-effect concentration for 95% of the species is not exceeded. Historical, empirical knowledge will be used to determine prior probability distributions of the assessment factors. These prior distributions are combined with distributions of substance-specific toxicity data, using Bayesian techniques. Both L(E)C50 and NOEC values may be used as input data. An advantage of this method in comparison with currently applied methods is that all toxicity data are used and that the uncertainty of the PNEC is quantified. Furthermore, the method is applicable to a data set consisting of at least 3 L(E)C50 values. Sample calculations show that the proposed method is less conservative than currently applied methods. Uncertainty in the PNEC, defined as the ratio of the 95th and 5th percentile of its distribution, is generally within two orders of a magnitude. A comparison of the proposed method and currently applied probabilistic methods reveals that the use of historical data reduced uncertainty in PNEC calculations with approximately a factor 4. The method seems to be relatively conservative when compared with NOECs derived with multi-species experiments. The results of these sample calculations indicate that the proposed method may provide an alternative for the currently applied methods to derive PNECs. Key words: uncertainty, ecological risk assessment, predicted no-effect concentration, assessment factor |
