|
5G - Species sensitivities distribution (TU2/11) Improving the validity of SSD applications. Hart, Andy1, 1 Central Science Laboratory, York, United Kingdom ABSTRACT- The theory and use of species sensitivity distributions (SSDs) has developed greatly in recent years, but there is continuing discussion concerning their validity and suitability for risk assessment. Forbes and Calow (Human and Ecological Risk Assessment, 2002, 8: 473-492) listed 9 assumptions implied by existing applications of SSDs: Interactions between species are unimportant; All species are equally important; Community structure is more important that ecological processes; Tested species are representative of untested species; Test endpoints are ecologically relevant; Chosen level of protection (e.g. HC5) is appropriate; Confidence limits are appropriate; Shape of distribution is appropriate; Number of species tested is adequate. Forbes and Calow conclude these assumptions are rarely satisfied in existing applications of SSDs. I will suggest that most of the assumptions can be avoided, and that the others can be satisfied sufficiently well to make the resulting SSD far superior to the use of conventional risk quotients and uncertainty factors. The key principles are (a) the SSD should be used as part of an appropriate model of cause and effect, (b) the assessment should make full use of the available information, and (c) assumptions and uncertainties should be made explicit and their influence on assessment results should be examined. Key words: Species Sensitivity Distributions, Risk Assessment |
