3.1 Motivation

It is clear that the strong latitudinal gradient in electron temperature, measured for the first time by Ulysses, requires a new model of the latitudinal structure of the torus. Figure 2 demonstrates the limitations of using the Voyager-based model [ Bagenal, 1994] to model the densities measured by Ulysses in the region of the torus [ Hoang et al., 1993, Moncuquet et al., 1997].

Figure 2: Electron density measurements from Ulysses (diamonds) compared to profiles extrapolated from Voyager 1 measurements. The two thin curves are the electron density profiles derived from the model of Bagenal [1994] using the $O_6$ magnetic field without current sheet and an isotropic velocity distribution (upper) or with current sheet and anisotropy of 5 for the hot electrons (lower). The dashed curve is a mid-range model ($O_6$ + currentsheet + isotropy) [ NB: the difference between $O_4$ and $O_6$ magnetic field models is small at these distances]. The thick curve, which is limited to the validity range of the polytropic law (i.e. $L\simeq 8{\rm R_J}$) is the density profile deduced from the simplified ``kappa-like'' model of Meyer-Vernet et al., [1995].

One can see that even with fairly extreme ion anisotropies and with a wide range in magnetic field models, the Voyager-based model cannot produce the tight confinement of plasma to the equator as observed by Ulysses, whereas Meyer-Vernet, Moncuquet and Hoang [1995] were able to match the Ulysses electron data with a single kappa velocity distribution for a single ion species and electrons, but this model was admittedly oversimplified.

With data obtained along spacecraft traversals through the torus that are widely separated in time, we are still limited to 2-dimensional models (where plasma properties vary with centrifugal latitude and radial distance but are assumed to be constant in the azimuthal direction). Moreover, it is quite possible that the properties of the torus have changed between the Voyager 1 flyby in 1979 and the Ulysses traversal in 1992. Remote sensing observations show that the torus density and temperature vary on time scales of days to years [ Thomas, 1993]. Nevertheless, the purpose of this paper is to explore whether a single 2-D model of plasma distribution (uniform in longitude), where the number of adjustable parameter is kept to a minimum, is able to match the density profiles observed at very different times and to reproduce the observed variations in temperature with latitude.