Reconnection by turbulence

Transfers of mass and magnetic flux are known to take place through the magnetopause boundary. When looking for the cause of these transfers, several scenarios have been invoked in the literature: 1) quasi-stationary reconnection; 2) localized reconnection (FTEs) 3) growing reconnection due to a local instability (tearing); 4) impulsive penetration of a magnetosheath inhomogeneity. As the very existence of the transfers implies that the frozen-in condition must be broken at some place, all of the preceding scenarios can be termed "reconnection" scenarios. The larger difference between them is that the two first scenarios pre-suppose the existence of an external electrostatic field; while the two others make use of a self-consistent inductive electric field. The varying magnetic field giving rise to the inductive electric field is due the growth of the tearing mode in the third case, while it is due to the spatial gradients limiting the incident blob in the fourth one. We will present a new scenario, of the fourth type, where the original cause for reconnection is the existence of a magnetic turbulence convecting from the shock region and impinging the magnetopause. We first show that this turbulence is converted onto the Alfven mode in the boundary gradient, where it is trapped and amplified. We also show how it can allow for transfers through the boundary, for both the magnetic flux and the plasma. A non ideal effect is of course mandatory for allowing such transfers: our model is calculated in the frame of Hall MHD, which means that the ion inertia effects are taken into account in the Ohm's law; the finite Larmor radius effects, nevertheless, have not yet been included up to now. Finally, we show that the magnetic flux reconnected per second through a perpendicular elementary surface can be calculated as a function of the local parameters; we are thus able to propose the definition of a local "reconnection rate". Analyzing the numerical results corresponding to the propagation of a linear Hall-MHD wave through the magnetopause, we estimate this reconnection rate and show that the reconnection is localized near the maximum of the trapped Alfven wave. The main predictions of the model are now confronted against the Cluster data.