Jupiter is surrounded by four large moons, discovered by Galileo Galilei (1610) and thus known as the Galilean satellites. In order of increasing distance from Jupiter they are Io, Europa, Ganymede, and Callisto. Table~I lists some of their basic characteristics. Their remarkably different physical natures have become apparent in the past few decades through a variety of groundbased and, more recently, spacecraft observations.

Table I. Some Physical Characteristics of the Galilean Satellites

The unit `RJ' is a Jupiter radius. `Solar day' is the rotation period, in earth days, measured with respect to the sun: it is also the sun-referenced orbital period. The surface composition is determined from the reflectance spectrum.
Satellite       Orbital         Radius
g cm -3
Io 5.91 422,000 1.77  3630 3.57  Sulfur,SO2 , silicates
Europa 9.40 671,000 3.55 3138 2.97 Water ice
Ganymede 15.0 1,071,000 7.17 5262 1.94 Dirty water ice
Callisto 26.4 1,884,000  16.75 4800 1.86 Very dirty water ice

The outer three satellites, unified by the presence of spectroscopically detectable quantities of water ice on their surfaces, are referred to as the `Icy Galilean Satellites'. The presence of ice makes their surfaces very different from the familiar silicate surfaces of the terrestrial planets. This dissertation is concerned with some new observational and theoretical investigations of the surfaces of the three icy satellites, and I will therefore first summarize briefly what we know about these surfaces from previous work.

Previous Work Concerning the Surfaces of the Icy Galilean Satellites

Surface Features Essentially all information comes from the Voyager images, which have been widely publicized. See e.g. Morrison (1982). The three satellites have drastically different surface appearances.
Europa is very strange. It has a young surface with very few impact craters, and very little topography except for numerous low-relief ridges and subdued small hills. The bright surface is traversed by innumerable dark linear features, some global in extent.
Ganymede has two very sharply demarcated terrain types (though transitional areas exist) each occupying about half the surface. Dark cratered terrain, with numerous impact craters and some tectonic features, occurs in polygons up to 2000~km across, surrounded by more lightly cratered, brighter, grooved terrain which shows intense tectonic modification. There are bright polar caps superimposed on both terrains, and very bright recent ray craters. Topography is rougher than on Europa but still subdued compared to the Moon.
Callisto is very heavily cratered and shows no internally-generated features. There are several very large multi-ring impact basins. The surface is generally dark but recent craters and the centers of the basins are brighter. Ray craters exist but are less conspicuous than on Ganymede. Topography is more subdued than the Moon's but may be rougher than on Ganymede.


Most groundbased broadband photometry of the Galilean satellites was obtained prior to the Voyager missions: see Morrison and Morrison (1977) and Veverka (1977a) for a summary. More recent photometric work has concentrated on the Voyager images, and is summarized by Veverka et al (1986), which also includes an update on the groundbased work.
Europa is the brightest icy satellite, with a mean geometric albedo around 0.68 in the V filter, a standard broadband filter centered around 0.56 microns (Morrison and Morrison, 1977, at zero solar phase including opposition effect, adjusted to Voyager diameter). It has a strong lightcurve, the leading hemisphere being 34% brighter than the trailing hemisphere in V, though the curve is very non-sinusoidal. Disk-resolved Voyager photometry (e.g. Johnson et al, 1983), shows a remarkable UV-dark `stain' centered on the trailing hemisphere, probably related to magnetospheric modification of the surface. There is little or no opposition surge, which probably indicates an unusually compacted surface texture.
Ganymede has a zero-phase V geometric albedo of 0.46 (derived as for Europa) and, like Europa, is brightest on its leading hemisphere, which is 15% brighter than the trailing hemisphere in the V filter. The lightcurve may be largely determined by the global distribution of dark cratered and light grooved terrain (Johnson et al, 1983). There is a UV darkening on the trailing hemisphere, like that of Europa but much less pronounced, and a moderate opposition surge.
Disk-resolved Voyager photometry by Squyres and Veverka (1981) gave normal reflectances in blue light of 0.35 for the cratered terrain, 0.44 for the grooved terrain, and up to 0.7 for the bright craters. The most detailed Voyager photometric work has been by Helfenstein (1986), who used Hapke's (1984) equations to obtain the physical characteristics of the various terrains from their photometry. He concluded, among other things, that the dark terrain surfaces were rougher (on a large scale) but more compacted than the bright grooved terrain, and had a smaller and broader opposition surge.
Callisto's zero-phase V geometric albedo is 0.23 (derived as for Europa), and unlike Ganymede and Europa it is brightest on the trailing hemisphere (13% brighter than the leading hemisphere in V). The leading hemisphere is UV­ darkened. The opposition surge is unusual, being twice as large on the darker leading hemisphere as on the brighter trailing hemisphere (where it is still larger than on Ganymede). The difference is too great to be due to albedo differences alone, and probably indicates a more porous surface on the leading hemisphere.
Voyager photometry by Squyres and Veverka (1981) showed a gave a normal reflectance of 0.18 for the dark terrain that forms the bulk of the surface, and values up to 0.7 for bright craters. Color work by Johnson et al (1983) confirmed these values and also showed that even the brightest regions of Callisto are red in color, as is also true on Ganymede and Europa. (`Red' in this context simply describes a surface with a higher albedo at longer wavelengths, which visually could appear red, orange, or more often, yellowish or brownish).

Reflectance Spectroscopy

So far all narrowband reflectance spectroscopy has been obtained from the ground or (in the case of the UV spectra) from Earth orbit. See Sill and Clark (1982) and Clark et al (1986) for summaries of the available data. All three satellites show aborption features due to water ice in the near infrared (1--4 microns), the strength of the major absorption features decreasing between the satellites with increasing distance from Jupiter. All, however, have reddened spectra in the visible region, unlike pure water ice, and must therefore contain another component, relatively dark and red in color, perhaps carbonaceous material or Fe3+ bearing silicates. The major uncertainties are the absolute amounts (areal or weight fractions) of water ice on each surface, and the identity of the dark red component. I discuss the reflectance spectra in detail in Chapter~10 of this dissertation.
Europa's surface appears to be almost pure water ice, with a minor, intimately mixed, red contaminant. Absorption features are broader on the trailing hemisphere, probably indicating a larger ice grain size there. There is a probable signature of sulfur in the UV spectrum, but only on the trailing hemisphere (Lane et al, 1981).
Ice is also abundant on Ganymede's surface but less so than on Europa. Estimates of surface weight percentage range from around 33% (Pollack et al, 1978; Spencer, 1987 and this dissertation) to more than 90% (Clark, 1980), depending on whether a segregated or intimate mixture of the ice and dark components is preferred.
Similarly on Callisto, estimates of ice weight percentage in the surface layers vary from 4% for a segregated surface (Spencer, 1987, and this dissertation) to up to 90% for an intimate mixture of the ice with the dark material (Clark, 1980).
Neither Ganymede or Callisto shows such large variations in spectrum shape with orbital longitude as does Europa.


See Veverka (1977b) for a summary of polarimetric data taken up to that time. Little further work has been performed since 1977. The disk-integrated polarization vs. solar phase curves for Europa and Ganymede are similar, and are consistent with surfaces of low-opacity material, presumably water frost. Callisto shows very different behavior on each hemisphere, with a lunar-like curve on the dark leading hemisphere and a curve intermediate between the Moon and Ganymede on the bright trailing hemisphere. So the leading hemisphere of Callisto looks more `Ganymede like' than the trailing hemisphere in polarization, opposition surge, and albedo. The polarization and opposition effect asymmetries are much more marked than the albedo asymmetry, though.

Thermal Infrared Radiometry

The extensive disk-resolved thermal IR observations of the icy Galilean satellites obtained by the Voyager spacecraft are analyzed in detail for the first time in this dissertation, though preliminary results were published by Hanel et al (1979a,b).
Previous published work has concerned groundbased, disk integrated measurements of the satellites' thermal emission, mostly broadband radiometry in the 10- and 20\dmic\ atmospheric windows. Again, there has been little data published in the last decade, and Morrison (1977) is still a good summary of the groundbased work. There have been two main avenues of exploration, measurement of fluxes from the sunlit hemispheres, and observation of cooling and heating of the surfaces during and after eclipse by Jupiter.
10- and 20\dmic\ brightness temperatures of the Galilean satellites, as obtained by a number of workers, are tabulated by Morrison (1977). Values are close to those expected from equilibrium with sunlight, though 20\dmic\ temperatures are systematically lower than the 10\dmic\ values by up to 10oK. More recent measurements by Matson (pers. comm.) confirm this trend, with 10- and 20\dmic\ brightness temperatures for Ganymede of 146.7oK and 140.4oK respectively (at a solar distance of 5.34 A.U.), and for Callisto 158oK and 153oK respectively. Europa shows large variations in temperature with orbital longitude which correlate with the visible lightcurve (darker orientations being warmer to the expected degree) except that it is anomalously cold for the first quarter-orbit after eclipse by Jupiter. Orbital variations in thermal emission on Ganymede and Callisto are smaller but are consistent with the visible lightcurves.
Eclipse radiometry has been used in conjunction with thermophysical modelling to constrain surface thermal properties (Morrison and Cruikshank, 1973; Hansen, 1973). Europa, Ganymede and Callisto all cool rapidly at eclipse onset and more slowly as the eclipse progresses, repeating this 2-stage behavior during warming after the eclipse. Ganymede and Callisto's thermal behaviors were successfully matched by very similar surfaces in which a thin, very low thermal inertia layer overlies a layer of much higher thermal inertia. Europa's puzzling lack of recovery to its pre-eclipse temperature after eclipse prevented a successful simultaneous model match to both the cooling and heating curves.
The ground-based thermal observations will be discussed in more detail in later chapters, when they are compared to the Voyager data.


Earth-based radar observations, which probe the immediate sub-surface to depths of centimeters to meters, are described by Ostro (1982). The icy satellites have very different radar signatures than rocky objects such as the Moon.
The 12.6-cm albedos of Europa, Ganymede, and Callisto are 0.65, 0.38, and 0.16, compared to around 0.03 for the terrestrial planets. This is presumably due to the icy nature of the subsurface, and the relatively high albedo of Callisto is thus very interesting considering its low visible albedo and possibly small surface coverage of ice.
Even more striking is the fact that most of the reflected radar beam is returned with the same sense of circular polarization as the incident beam, while the opposite is true for the terrestrial planets. Europa and Ganymede have similar circular polarization ratios but that of Callisto is smaller, closer to the terrestrial planets but still highly anomalous. The circular polarizations are difficult to explain, but may result from subsurface scattering of the radar beam in an inhomogeneous medium (Ostro, 1982).

The Present Work

This dissertation describes two separate but related lines of investigation. First, in Chapters 2 thru 7, I describe an analysis of the hitherto little known thermal emission observations of the icy Galilean satellites that were obtained by the Voyager IRIS instrument. This data set, consisting of several hundred disk-resolved thermal emission spectra for each satellite, provides information on global surface temperature distributions. Also, by using the shape of individual emission spectra, I am able to obtain information on local temperature distributions on scales smaller (possibly much smaller) than the several hundred km field of view of the IRIS instrument. I use this information to determine surface thermal properties, including the local distribution of thermal inertias and albedos, and also obtain some information on surface roughness. I find large differences in these properties between the three objects.
Christensen (1986) has applied similar techniques to the Viking IRTM thermal observations of Mars, using differences in brightness temperature at different wavelengths to obtain local temperature distributions. From this he maps regional variations in the abundance of high thermal inertia blocks. However this dissertation describes the first application of the technique to airless or icy surfaces. I show it to be a useful way of getting information on surface properties that is not readily obtainable in other ways.
Secondly, in Chapters 8 thru 11, I describe some theoretical considerations of the movement of water ice on the satellite surfaces, including the likely effects of thermal sublimation, micrometeorite gardening, and ion bombardment. I conclude that ice on the Galilean satellites is likely to concentrate in discrete bright patches separated by dark ice-free areas. This may explain some of the results of the analysis of the IRIS observations of Ganymede. I also discuss critically some of the other evidence concerning the distribution of ice on the satellite surfaces.
Finally, in Chapter 12, I summarize the results of the investigation and consider the opportunities for future work, especially using the data from the forthcoming Galileo Jupiter orbiter mission.
For definitions of many of the terms that I use, and other basic information, refer to the Glossary at the end of the dissertation.
Some of the work in this dissertation is already in press. See Spencer (1987a,b) for more convenient, but less current, versions of Chapters 8 thru 11.