We have performed new simulations with realistic potential structures and with an intitial state which had been warmed and accelerated by preexisting aurora. Our earlier simulations of the neutral response to an intense symmetric stable arc showed antisymmetrically distributed counterstreaming zonal winds with the strongest winds outside and on the flanks of the arc. The electric field was taken to be zero in the center of the arc; thus the ion drag acceleration and Joule heating were small within the arc despite tight coupling between ions and neutrals. Also, these simulations did not include atmospheric acceleration or heating associated with the diffuse aurora preceding arc formation. Later theoretical work and observations indicate that the electric field decreases in the arc relative to the field southward of the arc but does not necessarily go to zero or even become small. In our new simulations, strong zonal winds are generated within the arc, and the winds outside of the arc are stronger on the equatorward side by a factor of ~2--3, depending on the model of arc potential structure. In additon, the temperature in the upper part of the model domain (upper E region and lower F region) is cooler by a few degrees within the arc than outside, in qualitative agreement with DE observations.

Our earlier work also did not include the effects of arc-zonal wind feedback or of large-scale cross-arc winds. A subsequent theoretical model of coupled auroral and neutral dynamics showed that feedback causes the electric field to increase so as to remain approximately constant in the frame of the neutrals at altitudes near the peak of the Pedersen conductivity. This maintains an approximately constant ion drag as the neutrals are accelerated. Simulations with the electric field adjusted to maintain such a constant ion drag force indicate that feedback effects are not strong over the duration of a stable arc. Cross-arc winds decrease zonal winds by decreasing the time that air parcels are subject to strong zonal acceleration. Simulations with cross-arc winds representing large-scale flow indicate that strong winds are required to substantially reduce the zonal wind assoicated with wide stable discrete arcs (~50 km in width). ¿American Geophysical Union 1992