EXTENSION OF CONVECTION MODELING INTO THE HIGH-LATITUDE IONOSPHERE - SOME THEORETICAL DIFFICULTIES

An attempt has been made to extend the Rice Convection Model (RCM) and to merge it with empirical models, so as to cover the entire high-latitude ionosphere. Specifically, we have modified the RCM in two ways: (1) by a tailward expansion of the region where coupled ionosphere-magnetosphere convection model calculations are done, and (2) by use of ionospheric observation-based models poleward of the region where detailed convection modelling is applicable. We use a scaled and shifted Heppner-Maynard-Rich electric-field model directly for empirical extension of the ionospheric potential distribution into the polar cap and the statistical electron precipitation model of Hardy and co-workers less directly for the poleward extension of the auroral precipitation pattern. One goal in globalizing the RCM is to provide precipitation and electric field inputs for ionosphere and thermosphere modelers. We hope to provide an alternative to purely empirical precipitation and electric field models, by means of a hybrid model that is theoretical and dynamical with regard to the inner and middle plasma sheet, though still empirical with regard to the boundary plasma sheet and polar cap. We wish to avoid the statistical blurring that is a natural characteristic of empirical models and also to produce a model in which the boundaries of the precipitation and electric field patterns maintain physically consistent relationships to each other. Although the first set of runs of this globalized version of the RCM did indeed produce precipitation and electric field patterns with sharp features and with theoretically consisent relationships between boundaries, the results displayed two substantial difficulties. First, the model-predicted latitudinal width of the auroral sunward-flow region tended to be too narrow. Second, to avoid vastly unrealistic model precipitation rates, we were forced to place an artificial floor under the computed precipitation rate from the middle and outer plasma sheet. The computed auroral electron energy flux, plotted as a function of latitude, exhibited an exaggerated two-peak structure: one peak lies poleward of the coupled modeling region and is associated with the region-one Birkeland currents; the other peak lies at lower auroral latitudes within the coupled-modeling region, and is associated with the inner edge of the plasma sheet. When no floor was placed under the precipitation rate, the minimum between the two peaks was much too deep to be consistent with typical observations. The regions of excessively weak precipitation map to equatorial distances of 15-35 R(E) and thus to the regions of the plasma sheet that have not been included in previous self-consistent convection calculations. The most likely origin of the discrepancy is that electrons in the plasma sheet beyond 15 R(E) may not approximately satisfy the simple adiabatic condition p(e)(integral-ds/B)5/3 = constant; there is independent evidence that plasma-sheet ions violate the analogous adiabatic condition. In neither case is it clear what physical mechanism causes the violation.