Effects of dinoseb on the entire life-cycle of Daphnia magna. Part II: Modelling of survival and proposal of an alternative to No-Observed-Effect-Concentration (NOEC)

Effect values, especially for low concentrations, are urgently needed in risk assessment. Often such values are derived from ANOVA analysis or simple regression models. But these methods are not always satisfying, and some authors proposed to use a fully formalized parametric regression model to describe the entire dose-response curves. They allow better evaluation of pollutant effects, including inter- and extrapolation to any other than measured effect values. Following this line, a four-parameter logistic regression model (standard model) has been fitted to survival data of Daphnia magna under a pesticide stress (dinoseb). The heterogeneity of the variance is taken into account with a both-side logarithmic transformation. Besides the standard model, hormesis and a threshold model have been tested too. The three models can all be fitted well to the data, the difference between them decreasing with increasing concentration. However, the standard model gives the most sensitive parameters and we therefore retained this model for our calculations. The use of the two others should be restricted to cases where there is a strong evidence for a hormesis or a threshold. Model calculations permit the quantification of effects on individuals' longevity as well as on the mean survival times of the entire population. Using model calculations we defined a statistically sound no-effect value, the SNEC (Statistical-No-Effect-Concentration), which is less variable and less sensitive to experimental design than the NOEC. The SNEC is an interesting alternative to that classical value.