A circuit model for filament eruptions and two-ribbon flares

We derive a circuit model for solar filament eruptions and two-ribbon flares which reproduces the slow energy build up and eruption of the filament, and the energy dissipation in a current sheet at the top of post-flare loops during the two-ribbon flare. In our model the free magnetic energy is concentrated in a current through the filament, another current through an underlying current sheet, and surface return currents. The magnetic field configuration, generated by these currents and a general photospheric background field, has a topology similar to the field topology derived from observations.We consider two circuits, that of the filament and its return current, and that of the current sheet and its return current. These circuits are inductively coupled and free energy stored in the filament in the pre-flare phase is found to be transferred to the sheet during the impulsive phase, and rapidly dissipated there. A comparable amount of magnetic energy is converted into kinetic energy of the ejected filament. The basic equations of the model are the momentum equations for the filament and the current sheet, and the induction equations for the filament and sheet circuits. The derivation of the equations is an extension of previous models by Kuperus and Raadu, Van Tend and Kuperus, Syrovatskii, and Kaastra. The set of equations is closed in the sense that only the initial conditions and a number of parameters, all related to pre-flare observables, are needed to calculate the evolution of the system. The pre-flare observations we need to determine these parameters, are: (1) a magnetogram, (2) an picture, (3) a measurement of the coronal density in the region, and (4) estimates of the photospheric velocity fields in the region.In the solutions for the evolution of the filament current sheet system we distinghuish 4 phases: (1) a slow energy build up, lasting for almost two days, during which the filament evolves quasi-statically, (2) a metastable state, lasting for about three hours, during which the filament is susceptible to flare triggers, and during which a current sheet emerges, (3) the eruptive phase, with strong acceleration of the filament, during which a large current is induced and dissipated in the current sheet, and energy is injected in the post-flare loops, and finally (4) a post-flare phase, in which the filament acceleration declines and the current sheet vanishes.From further numerical work we derive the following conclusions: (1) The magnetic flux input into the filament circuit has to surpass a certain threshold for an eruption to occur. Below that threshold we find solutions representing quiescent filaments. (2)Flare triggers are neither necessary nor sufficient for an eruption, but may set off the eruption during the metastable state. (3) The model reproduces the increase in shear in the filament prior to the eruption, through adecline of the filament current, in contrast to most models for filament eruptions. (4) The ratio of energy lost as kinetic energy of ejecta to the energy radiated away in the post-flare loops is sensitively dependent on the resistance of the current sheet. (5) Flare prediction is possible with this model, but the potential for triggering during the metastable state complicates the prediction of the exact moment of eruption.Former NAS/NRC Resident Research Associate.ST Systems Corporation.