MERCREDI 10 JUIN 2009 14 heures Salle Séminaire 5 Centre de Physique Théorique Marseille-Luminy Freddy Bouchet INLN, Nice Sophia-Antipolis Title: Out of equilibrium statistical mechanics and phase transitions for the large scales of 2D and geophysical turbulence Abstract: We consider two dimensionnal or geophysical flows driven out of equilibrium by stochastic forces. When these are balanced with linear dissipation one reaches non equilibrium statistically stationary states (NESS), without detailed balance. We discuss theses examples from a statistical mechanics point of view: for instance we discuss the existence of out of equilibrium phase transitions, or consider kinetic theory approaches. For the two-dimensional Navier-Stokes equations with weak stochastic forcing and dissipation, the existence of an invariant measure has been mathematically proved recently, together with mixing and ergodic properties. This problem has however never been considered from a physical point of view. We thus address the following issues: when is the measure concentrated on an inertial equilibrium, how are the large scales selected by the forcing, what is the level of the fluctuations? The most striking result is the existence of out of equilibrium phase transitions. One observe transitions from one type of flow (unidirectional) to an other one (dipole), at random time. This is similar to the classical two wheel potential with noise. By contrast, in our case, no such potential exists and the turbulent nature of the flow (infinite number of degrees of freedom) renders the phenomena much richer. Analogies with the Earth magnetic field reversal, and with similar phenomena in experiment of two dimensionnal and geophysical flows will be discussed. The approach of these phenomena using kinetic theory will be discussed. This is the most natural way for a theory for the self organized jets of geophysical turbulence, for instance in Jovian atmospheres. We prove new results for the linearized Euler and Navier-Stokes equations with random forces, and their relations to turbulent fluctuations.