A set of boundaries was chosen to model the principal observed magnetospheric regions. Those regions which extended out to the distant magnetotail were defined at Xsm = -20 R e. The Tsyganenko (1989) magnetic field model (T89) was used to project the boundaries down to the ionosphere. It was found that all field lines which passed within 3 R e of the magnetopause projected to the dayside ionosphere. Dayside arcs therefore generally map to the low-latitude boundary layer, the cusp/cleft, or the entry layer. The nightside auroral zone, and therefore field lines associated with the substorm current diversion process, primarily traced to portions of the plasma sheet that do not come in direct contact with flowing solar wind plasma. Several mapping problems were encountered. The first involved identifying certain boundaries in the magnetosphere model. The flanks of the magnetotail are not modeled realistically. As a result, we had difficulty in defining a T89 magnetopause in the equatorial plane. Other problems were that some magnetotail boundaries may have no ionospheric signature, and that some boundaries are influenced by cross-field plasma drift. Plasma boundaries are not tangent to field lines when drift is present, but all mapping was done by following field lines. Uncertainties of about 1ø of latitude in the resulting ionospheric projections were found for each 1 R e of drift near midnight at Xsm ------20 R E. The steady state magnetotail then was subdivided out to Xsm ----22 R E according to the expected characteristics of charged particle orbits. Orbits were traced in the modified Harris magnetic field model, with parameters adjusted to approximate the shapes of T89 magnetic field lines. The one-dimensional Harris model was used to eliminate some drift effects. This permitted a detailed study of each separate subregion. Curves defining various orbit types were projected to the ionosphere. It is suggested that low-or middle-altitude satellites may be able to detect regions of quasi-adiabatic and nonadiabatic equatorial orbits by monitoring the loss cones. For these highly field aligned loss cone particles, we found that the net effect of the complex interaction with the current sheet can be explained by an extremely simple model. The model involves shifting all generally field-aligned ions as a block through an energy-dependent "scatter" angle. When viewed microscopically, the resulting pitch angle changes are highly structured. Macroscopically it may be possible to describe the process in terms of pitch angle diffusion. Figures 1 and 2 show how structures map between the ionosphere and the equatorial plane in the K•, = 0 version of T89. In K90 we used color plates to present similar information for the Beard quiet magnetosphere model. Figures 1 and 2 were prepared by mapping field lines down from the equator to the ionosphere. 9307 9308 KAUFMANN ET AL.: MAPPING MAGNETOSPHERIC BOUNDARIES -lORE< x< I0 R E Kp =0 14 / 16 18 60 70 80 90 -60 R E < x < -I0 R E 80 2O 22 70 60 Quiet Magnetosphere