In 1979, the IRIS infrared spectrometers on the two Voyager spacecraft obtained over 1000 disk-resolved thermal emission spectra of Europa, Ganymede, and Callisto, Jupiter's three large icy satellites. This dissertation describes the first detailed analysis of this data set. Ganymede and Callisto subsolar temperatures are 10°K and 5°K respectively below equilibrium values. Equatorial nighttime temperatures are between 100°K and 75°K, Callisto and Europa being colder than Ganymede. The diurnal temperature profiles can be matched by 2-layer surfaces that are also consistent with the eclipse cooling observed from earth, though previous eclipse models underestimated thermal inertias by about 50%. Substrate thermal inertias in the 2-layer models are a factor of several lower than for solid ice. These are 'cold spots' on Ganymede and Callisto that are not high-albedo regions, which may indicate large thermal inertia anomalies. All spectra show a slope of increasing brightness temperature with decreasing wavelength, indicating local temperature contrasts of 10-50°K. Callisto spectra steepen dramatically towards the terminator, a trend largely matched with a laterally-homogeneous model surface having lunar-like roughness, though some lateral variation in albedo and/or thermal inertia may also be required. Subsolar Ganymede spectra are steeper than those on Callisto, but there is no steepening towards the terminator, indicating a much smoother surface than Callisto's. The spectrum slopes on Ganymede may indicate large lateral variations in albedo and thermal inertia. A surface with similar areal coverage of dark, very low thermal inertia material, and bright material with thermal inertia a factor of 2-3 below solid ice, fits the diurnal and eclipse curves, and (less accurately) the IRIS spectrum slopes. Europa spectra have very small slopes, indicating a smooth and homogeneous surface. Modelling of surface water ice migration gives a possible explanation for the inferred lateral inhomogeneities on Ganymede. Dirty ice surfaces at Jupiter are subject to segregation into high-albedo ice-rich cold spots and ice-free regions covered in lag deposits, on decade timescales. Ion sputtering and micrometeorite bombardment are generally insufficient to prevent the segregation. The reflectance spectra of Ganymede and Callisto may be consistent with this type of segregated surface.

The silicate "magma trigger" model of volcanism on Io has been evaluated numerically with finite element methods by considering the one-dimensional heat transfer between hot silicate magma and initially cold sulfur. It is found that for the probable range of initial magma temperatures and sulfur temperatures, the contact between silicate magma and a sulfur crust will be 700 (+OR-) 100 K, or approximately the vapor point of elemental sulfur. A silicate magma sill or laccolith on the order of 10 m thick will yield energetic vapor for a period of several weeks to several months depending on the vapor temperature and the amount of convective cooling of the silicate magma that occurs at the silicate-sulfur interface. This model may account for the origin of plumes and possible sulfur flows, as well as for their observed temperatures ((TURN) 600-700K) and lifetimes (several days to a few months). If the conducted heat flow is similar in high and low latitudes, then the low latitude occurrence of plumes may be explained as a result of lower temperatures at higher latitudes. Because the contact temperature of sulfur and silicate magma depends on the pre-existing sulfur temperature, a system in which sulfur vapor temperature is just reached at the equator would not generate sulfur vapor under lower initial sulfur temperatures existing at high latitudes. If the heat flow is higher in high latitudes, then the sulfur crust must be thinner than it is in low latitudes for the model to work as described above. Most of the heat flow from Io may be moved by convection from the interior to the surface, not by conduction. Heat flow may be modulated by the efficient transfer of silicate melts from 40 to 300 km depth, and emplaced as laccoliths at the sulfur-silicate crustal interfaces at a depth of 5-10 km. Sulfur flows, plumes, calderas and other areas of massive radiant heat dissipation continue the convective cycle to the surface. The temperature at the base of the sulfur crust may be less than the melting point of sulfur, and the silicate magma temperature can be as low as 1200 K. Low silicate magma temperatures will occur if the crust of Io is as differentiated as terrestrial rhyolites and trachytes. High alkalies in the Io plasma torus suggest the possibility that the Ionian crust is a highly differentiated silicate.

This dissertation combines several new observations of the Io sodium cloud to create a consistent picture of the extended Io atmosphere and its interaction with the Jovian plasma torus. I used the LPL echelle spectrograph to obtain three types of high-resolution spectra of the extended sodium cloud at the sodium D-lines (5890, 5896Å). The first class of observations made use of the mutual satellite eclipses of 1985 to probe the density profile of the atmosphere in the range 1.4 to 10 Io radii, a previously unstudied region. The second type of observation examined the sodium emission in Io's immediate vicinity, allowing an accurate measurement of the velocity structure around Io. The final method employed a high-sensitivity detector to study faint jets of high-speed sodium farther out in the extended cloud. The synthesis of these three data sets results in a better understanding of how sodium is distributed about Io as a function of position and velocity. Io's extended atmosphere is composed of many kinematically distinct components. The distribution in space is linked to their characteristic velocities, with low-energy sodium confined near Io and faster atoms (10 to 100 km sec⁻¹) prevalent beyond ∼25 Io radii. The sodium density profile is steep near Io and shallower outside 5.6 Io radii, the effective limit of Io's gravity. The data indicate that the atmosphere is collisionally thick near the surface, but becomes thin by an altitude of ∼700 km. The upper limit of the exobase location is derived from reliable sodium density measurements made during the satellite eclipses. The lower limit is indirectly inferred from the velocity distribution of sodium near Io and the nature of high-speed jets far from Io. The high-speed sodium jets reveal a new type of close interaction between the corotating plasma and Io's atmosphere. The morphology and brightness of the jets require a two-reaction process, in which atmospheric sodium is ionized, accelerated to high speeds, and then charge-exchanges with other sodium atoms. These processes must occur near the atmospheric exobase, indicating that Io's atmosphere is not completely protected from the plasma flow.