Etude et modélisation des effets écotoxiques d'un micropolluant organique sur "Daphnia magna" et "Pseudokirchneriella subcapitata"

The aim of this work is the modelling of the effects of a pollutant on an organism including time as a variable. The model developed will permit the evaluation of the toxicity globally, at short or long term, for low or high concentrations. This approach would permit the determination of values for very small effects or values for no-effect which could be utilised in the extrapolation of the laboratory results to the environment. To develop this approach, two organisms were chosen, a microcrustacea (Daphnia magna) and a green algae (Pseudokirchneriella subcapitata). The pollutant tested is a pesticide, dinoseb, which acts as a herbicide and as an insecticide. The measured effects are survival and reproduction for daphnids as well as growth, pH, chlorophyll a, photosynthesis and respiration for algae. To introduce time as a variable for the interpretation of the results, these parameters are measured during the life-cycle of daphnids and during the entire culture of algae. In the first part of this work, daphnids and algae cultures are studied to obtain a better knowledge of the biology of the organisms in laboratory conditions and to evaluate the natural variability of the measured parameters. For daphnids, this work permits us to establish a typical behaviour for survival and reproduction. This behaviour allows the evaluation of the quality of the cultures (and consequently of the test organisms) and the quality of a test, comparing the controls to the cultures. For algae, the study of the cultures permits us to establish the growth curve and the variability of the parameters measured during the test. In the second part, the effects of the pollutant on the different parameters are measured regularly to introduce time as a variable. The results show that dinoseb has an abrupt effect on daphnids and on algae, which could be attributed to the fact that the number of processes affected increases with the concentration. Indeed, dinoseb is a pesticide with a non specific mode of action. The range of concentrations producing effects on daphnids and algae are similar. For daphnids, the reproduction is slightly more sensitive than survival, and for algae, the growth is the most sensitive parameter. These parameters are therefore retained for the modelling. The variable time is directly taken into account in the survival time of daphnids. However, for the reproduction of daphnids and the growth of algae, the time must be introduced as a supplementary axis (longitudinal data). This results in a three dimensional graph with the effects measured as functions of time and concentration. In the third part, a model is adjusted to the survival data of daphnids. The model chosen is a simple logistic model with three interpretable parameters. It enables a global evaluation of dinoseb toxicity on survival. The model points out the effects on life-cycle time and on population for high and low concentrations. Its application permits us to define effects values (ECx, LCx), and especially a statistical no effect value which is more sensible then the NOEC (No Observed Effect Concentration). We propose to call this value SNEC (Statistical No Effect Concentration).The SNEC, as well as the other effect values, could be used for the evaluation of the impact of pollutants on the environment. This constitutes an interesting tool for ecotoxicology.