Azimuthal variation of ion density and electron temperature in the Io plasma torus

We analyzed about 47 hours of Voyager 1 ultraviolet spectrometer observatio ns of the Io plasma torus in order to determine the distributions of ion an d electron density and electron temperature as five-dimensional functions o f (1) radius (L), (2) System III longitude (lambda(III)), (3) System IV lon gitude (lambda(IV)), (4) Jovian local solar time (phi(circle dot)), and (5) azimuth relative to Io (phi(Io)). We present averaged profiles of electron density (n(e)) and temperature (T-e) as one-dimensional functions of each of these five variables in turn. Each of these profiles is am average of th e full five-dimensional distribution over the other four variables. Our T-e (L) estimate is substantially the same as that determined from in situ obse rvations by the Voyager 1 plasma science investigation. As for the azimutha l profiles, we find that the n(e)(lambda(III)) profile has a peak near lamb da(III)=180 degrees, as expected from ground-based observations, while the T-e(lambda(III)) profile peaks around lambda(III)=300 degrees; The Jovian s urface magnetic field reaches its lowest magnitude on the Io footprints nea r the T-e(lambda(III)) maximum, suggesting that part of the torus electron heating might be associated with Birkeland currents or some other magnetosp here-ionosphere coupling. In the Io frame, we find an apparent n(e)(phi(Io) ) maximum about 90 degrees downstream from Io while the highest T-e(phi(Io) ) occurs near Io, suggesting that electrons are locally heated by the Io at mosphere-torus MHD interaction. These two T-e profiles suggest that field-a ligned currents might comprise the long-sought additional (i.e., besides fr esh ion pickup) electron-heating power source for torus emissions needed to resolve the torus "energy crisis". In local solar time, we find that the T -e(phi(circle dot)) and n(e)(phi(circle dot)) profiles peak between dusk an d local midnight, with the former maximum nearer dusk and the latter nearer midnight. This combination suggests that the variation is due to an offset of the torus toward local midmorning, and perhaps that the cooling time is short. We also describe a general method for constraining solutions to be nonnegative when fitting data using the singular value decomposition least squares technique.