Résumé
Degradation of the water quality of superficial waters is now a worldwilde problem of concern. In lakes, this deterioration often leads to a strong development of the phytoplankton biomass. In some favorable conditions, blooms can be dominated by the cyanobacteria, which bloom at the lake surface, strongly reducing water transparency. Furthermore, as most of these cyanobacteria are toxic, any use of the invaded lake for leisure or water resource is compromised. The purpose of this study is to shed light on the most important factors which allow such a strong development of one of these species, Microcystis aeruginosa, in a man-made lake (Villerest, Loire France). A numerical model of the lake was developed, based on an intensive data-base over three years. First of all, it makes possible to quantify the vertical mixing processes that govern not only the thermal structure but also the exchanges of nutrients and, more generally, the distribution of dissolved and particulate matter between the different parts of the lake. The vertical temperature profiles are calculated by solving the one-dimensional heat transfer equation, which takes into account internal heat sources and sinks, advection due to inflow and outflow, molecular diffusion and eddy diffusion. The ecological part of the model takes into account the different dominant phytoplanktonic species observed on the site throughout spring and summer. A particular attention was given to the simulation of Microcystis aeruginosa for which a flottation model has been introduced. The nutrients (nitrogen and phosphorus) cycling are simulated as well. The model includes 3 types of organic matter and the exchanges at the sediment-water interface are deduced from the early diagenesis equations. The resulting oxygen budget is calculated. The model allows to determine the main environmental and physiological factors of the planktonic succession. It appears that the regulation of the buoyancy of the cyanobacteria is a main factor of their dominance on the lake, but that this ability is strongly reduced when the vertical mixing is important in the water column. An ecological model, coupled with an intensive on site data acquisition appears here as a powerful tool to understand the function of the lacustrine ecosystems and helps to define the rules for a better management.