Blooms of Gymnodinium mikimotoi have deleterious effects on marine aquaculture stocks (fish and shellfish), on species recruitment (shellfish and probably fish) and possibly on marine flora and ecosystems. In stratified environments, populations are confined into very thin layers located in the pycnocline. The reasons for this confinement are yet unclear, the pycnocline being the only location either where this species can grow or where it can survive. We demonstrate here that pycnocline layer, when it exists, is the location where the maximum net growth rate is encountered, due to a lower mortality rate.Toxicity of G. mikimotoi is due to a labile exotoxin (20 min. half-lifetime). Synthesis of this exotoxin has allowed to determine the mechanism of action of this toxin: it inhibits in a non-specific way membranes ATPases. These enzymes are the energy source for ions exchanges at membranes. Biological targets are, therefore, incapacitated in their osmotic pressure regulation. The effect of these exotoxins have been studied in terms of stock losses, but never in terms of the effects on the ecology and the development of a bloom. The spatial scale of action in relation to degradation is of the order of few centimeters. Since individual cells have been observed to aggregate during the growth phase of the population, it is very likely that the population creates its own specific environment. Though less sensitive than their competitors, G. mikimotoi cells are sensitive to their own toxins. In still environments, like a culture, cells are able to avoid collision. However, in natural environments, collision rate of cells depend on the shear. A simple formulation of population growth implemented in the 3D-model of the Bay of Biscay has been used to simulate time-series in the Bay of Biscay (France) according to the following equation: with m: growth rate, hn: light intensity and g: shear. The zone of inoculation of the population was defined from different scenarios using analysis of trajectories. We present here the evolution of population growth along different transects, demonstrating the population confinement in the pycnocline.These results are discussed in the context of "species-of-interest" population dynamics approach.