PARENT SESSION

5A - Uncertainties and Variabilities
Hall 7
1:45 PM - 3:30 PM, Monday, 28 April 2003
Chair: Scheringer, M.¹, ¹
Co-chair: Matthies, D.M.², Green, J.³, ^{2 3}

(MO7/13) Uncertainty and variability in spatio-temporal probabilistic risk modelling.

Verdonck, Frederik¹, Deksissa, Tolessa¹, De Laender, Frederik¹, De Schamphelaere, Karel², Vandenberghe, Veronique¹, Vincke, Sofie¹, Janssen, Colin², Vanrolleghem, Peter¹, ¹ Ghent University - BIOMATH, Gent, Oost-Vlaanderen, Belgium² Ghent University - Laboratory of Environmental Toxicology and Aquatic Ecology, Gent, Oost-Vlaanderen, Belgium

ABSTRACT- In the current conventional ecological risk assessment, risk quotients of Environmental Concentrations (EC) and Species Sensitivities (SS) (through ecotoxicity test results) are used to estimate the likelihood and the extent of adverse effects occurring to humans and ecological systems due to possible exposure(s) to chemical substances. The major drawback of the current risk quotients is that they insufficiently account for the inherent uncertainty and variability of the EC and the SS. The goal here is to present a suite of environmental modelling tiers, where each time, a component of the variability or uncertainty is explicitly considered and refined. In a probabilistic analysis, the inherent spatial and temporal variability and the uncertainty of EC and SS are quantified and simulated by means of probability distributions. As a result, the risk is a probability instead of >1/<1 result. Some Environmental Concentration Distributions (ECD), Species Sensitivity Distributions (SSD) and their resulting risks are given as examples. Incorporating spatial variability of the EC and SS by combining GIS (Geographical Information Systems) and models can further increase realism. By geo-referencing the EC and SS, the spatial variability is explicitly accounted for and as a result the remaining overall variability of the results can be reduced. GREAT-ER and BLM (Biotic Ligand Model) in combination with GIS are given as examples for geo-referencing the EC and SS respectively. The temporal variability in an environmental analysis can be accounted for by means of a dynamic modelling approach. Some examples of Concentration-Duration-Frequency (CDF) curves will be given.

Key words: Species Sensitivity Distribution, Environmental Concentration Distribution