The upper atmospheres of the outer gaseous giant planets Jupiter, Saturn, Uranus and Neptune are dominated by hydrogen in atomic and molecular form. These components have strong spectral signatures in the ultraviolet region of the spectrum. Atomic hydrogen has its Lyman series, with the strongest Lyman-alpha line at 1216 Ang (1 Angstrom is 10 nanometers), and the weaker lines spanning down to 911 Ang at the hydrogen ionization continuum. Molecular hydrogen, H2, has the prominent Lyman and Werner bands from ~800 to ~1670 Ang. The reflected sunlight from a planet also shows atmospheric compositional properties, such from the absorption signatures of simple and complex hydrocarbons present in the Jovian planet stratospheres, and also from the scattering properties of molecules and hazes.

UV studies of airglow and auroral emissions have been critical for our understanding of planetary upper atmospheres and their interaction with the magnetospheric plasmas and magnetic fields, directly or indirectly with the impinging solar wind. Major fundamental discoveries on the atmospheres of the Jovian planets were made in the 1980s with the UVS instrument on the Voyager spacecraft (built by researchers at the Univ. of Arizona). During this time and through the mid-1990s the International Ultraviolet Explorer (the IUE satellite) also provided key UV observations of Jupiter, Saturn and Uranus (cf, Ballester, in IUE Final Archive, ESA SP-413, 53,1998). Currently we can make remote-sensing observations of these planets at UV wavelengths in the ~1180-3200 Ang regime with the Hubble Space Telescope (HST). HST orbits the Earth above the UV absorbing, ozone-rich atmospheric layers. It has carried various UV sensitive instruments which have made many planetary discoveries and produced valuable archival datasets. The most useful instrument available at this writting (2003) is the Space Telescope Imaging Spectrograph (STIS). There are many time variable phenomena in Jovian systems which could reveal unique, key parameters of the systems. The Cassini sapcecraft will reach Saturn in the year 2004, and will provide some UV information of this system. However, the need for a remote-sensing UV telescope dedicated to planetary studies, with tailored imaging and spectroscopic capabilities, is becoming more and more pressing.

What are a magnetosphere and an aurora? The sun emits a wind of energetic magnetized plasma that escapes from its hot corona. When the magnetic field and the plasma and ionosphere of a planet interact with the impinging solar wind, a magnetospheric ``cavity'' is formed as the internal and external pressures are balanced. In general, four major aspects characterize a magnetosphere: symmetry of the system, solar wind driven convection, corotation, and plasma sources. In a magnetosphere, current systems and wave-particle interactions can drive energetic particles to travel along the planetary magnetic field lines into the upper atmosphere around the magnetic poles, thereby depositing large amounts of energy and generating aurora. An aurora consists of the emissions that are produced when the incoming energetic particles (mainly electrons, but also protons and heavy ions) collide with the atmospheric gases (atoms, molecules, ions) and excite them. When the gas is de-excited (goes back to its ground state) it then produces an emission. The same thing happens in a neon lamp, when a current of electrons passes through the neon gas.

Curtains of auroral emissions on Earth are seen in this picture taken from the Space Shuttle (see below). Also seen is the "shuttle glow".

Various interesting UV images of auroral emissions on various planets and satellites are compiled in the Aurora Seeker's website.

For centuries, these auroral emissions have fascinated us on Earth. On Earth nitrogen and oxygen species produce white, green and red auroral emissions that vary with the altitude of the auroral energy deposition. Spectroscopic studies of auroral emissions reveal details about the atmospheric species, temperatures of the auroral regions, motions in wind fields, and with adequate planetary atmosphere models one can derive the altitude of energy deposition and thus the energetics of the processes involved. From images of the auroral morphology, and adequate planetary magnetic field models, the emissions can be mapped out along field lines to the magnetospheric sites of precipitation, delineating the global magnetospheric configurations resulting from its interaction with the solar wind. Remote sensing studies of auroral emissions are thus quite useful.


Jupiter's magnetosphere (F. Bagenal, Ann. Rev. Earth Pl. Sci., 20, 289, 1992).

The outer planet most extensively studied has been Jupiter. Its Galilean satellite Io has the most active volcanism in the Solar System, and its atmosphere interacts electromagnetically with the surrounding medium, the Io plasma torus. This interaction is itself the source of the oxygen and sulfur ions of the torus. The plasma is picked-up by Jupiter's magnetic field into corotation with the ionosphere. Eventually, the plasma populates the entire Jovian magnetosphere. Other contributions are from the ionosphere and the solar wind. Jupiter's strong magnetic field, fast 10-hr rotation, and internal plasma sources result in an immense a magnetosphere with a strong corotational character. It is stretched and flattened by an equatorial current/plasma sheet extending from the Io/torus source at 6 Rj out to ~30 Rj where corotation breaks down. (1 Rj ~ 11 Re ~ the size of the Earth's magnetosphere!) The outermost regions are influenced by the solar wind pressure, and so the magnetopause can quickly change size by a factor of 2 following changes in the solar wind. The coupling of this giant, fast-rotating, internally supplied magnetosphere with the solar wind is complex. Even with extended Galileo in-situ measurements since 1995, we still have a lot to learn about this system. The atmosphere-magnetosphere interaction produces 10^14 Watts of auroral power, with emissions spanning from X-ray to radio wavelengths. This is 10,000 times more powerful than the Earth's aurora, and could light all cities around the world!

The figure on the left was published in Science by G.E. Ballester et al. (1996) and consists UV images of Jupiter full disk obtained with the WFPC2 camera on HST. It shows simultaneously the north and south auroral emissions as seen throughout half of a Jovian rotation. (See below.)

The figure on the right is an HST press release by Clarke et al. (1997) and shows a composite figure of Jupiter 's north and south aurorae imaged with the more sensitive STIS instrument on HST. (See below.)

Emissions from both H2 and H are detected in these images. The HST images have shown that the aurora is dominated by main ovals of bright and discrete emission at high latitudes on the north and south polar regions, mapping farther than 15 Rj into the middle magnetosphere. These are accompanied, at least on the north, by emissions that are poleward of the main oval, mapping to the outer magnetosphere. The exact mapping is hindered by uncertainties in the magnetic field models, in the internal non-dipolar components and contributions from the current sheet and magnetopause currents. The STIS image clearly shows emissions at lower latitudes of the main ovals that are at the footprint of the magnetic field lines connected to Io and its extended (wake) tail. Emissions mapping to Europa and Ganymede have also been seen (Clarke et al. 2002).

One major question about this fast-rotating system is to understand the importance of dependencies of the aurora and thus of the magnetosphere on corotational properties which are fixed in magnetic longitude and thus internally controlled, versus dependencies on the magnetic local time which are dominated instead by the solar wind interaction. All the auroral emissions are highly variable in brightness and morphology, and we can use these variations to learn more about the system. The WFPC2 images shown on the left figure show a partial Jovian rotation. Some spots of emission along the ovals were seeing to corotate with the planet. However, other properties were seen to depend on the magnetic local time. In particular, the brightest emissions were seen associated with a strong auroral event that was taking place and remained confined to the dawn regions. There are other, more subtle effects that we can see in the images which we are currently studying, such as the behavior of the emissions as the rotate from the morning to the afternoon.

The Galileo orbiter has toured the system since the mid 1990s, and has provided much valuable information about the Jovian magnetosphere. On December of 2000 the Cassini spacecraft flew by Jupiter on its way to Saturn, and could measure the solar wind impinging on Jupiter and dusk side of the magnetosphere. Galileo was also making its in-situ measurements of the system. HST also took a limited set of auroral images during this period. Some results have been published from this special set of coordinated observations (Science, 2002). But a full study of that STIS dataset still awaits.

Much information is still buried in the extensive HST dataset, which started to be collected in 1994 with WFPC2. This is the basis for our archival program at the Univ. of Arizona. In addition, MHD simulations of the Jovian system and its interaction with the solar wind have only recently been made, and we are aiming at comparing the auroral observations with the MHD simulations, magnetic field models, and simultaneous Galileo measurements made by our colleagues at UCLA.


Sketch of the magnetosphere of Saturn, from Bagenal (1992).

Saturn's magnetosphere is large and fast (10-hr) rotating like that of Jupiter. It also has internal plasma and neutral sources in the icy satellites and the rings, both with ephemeral atmospheres, and also Titan, which has a dense atmosphere. However, unlike Io, Titan orbits in the outskirts of the system, sometimes inside of it and sometimes "outside" of it and directly exposed to the solar wind. The Kronian and Jovian magnetospheres therefore have dominant or at least substantial internal mass and rotational energy sources against an external solar wind control, in contrast to the case of the Earth. In the 1980-1981 Voyager encounters, Saturn's magnetic field was found to be surprisingly aligned with the spin axis, and presents a simpler system than Jupiter's because the field non-dipolar terms are much smaller. Saturn's aurora was also positively identified in UV emission and seemed to emanate from high latitude ovals thought to map to the magnetopause, although, as for Jupiter, the models and mapping are uncertain. Saturn' system may thus be controlled mainly by the solar wind interaction, rather than by internal forces. However, Voyager also revealed kilometric radio emissions (SKR) consisting of sychrotron radiation by precipitating auroral electrons. The SKR and some ring spokes showed some enhancements at a given magnetic longitude, thus indicating that longitudinal (and thus corotational) effects are at play as well. HST imaging will be key for untangling the behavior of Saturn's aurora and magnetosphere.

The figure on the left was the first ever obtained of Saturn in the far-ultraviolet (~1200-2100 Ang) in 1994 with the WFPC2 camera on HST. They were published by John Trauger et al. (1998). Saturn's disk and rings are seen in reflected sunlight (longwards of ~1600 Ang). The auroral emissions are easy to distinguish in the polar regions because those regions are dark in the UV due to UV-absorbing polar hydrocarbon hazes formed in turn by the auroral processes. Both emissions in H2 and H Lyman alpha are detected in these images. A strong auroral event was taking place in Saturn's morning aurora, as shown in the right figure.

The figure on the right, from an HST press release also by Trauger et al., was obtained with the STIS instrument on HST. The STIS far-UV imaging mode has a about 10 times better sensitivity than a typical WFPC2 exposure, and about 4 times the spatial resolution (eg, the Cassini division of the rings is now clearly visible). Details of the aurora can now be better studied. The STIS far-UV MAMA detector covers the 1200-1800 Angstroms, so there is also less reflected sunlight as compared to the WFPC2 images. By 1997 the northern hemisphere of Saturn was barely observable. Brighter morning emission is seen on the outh east limb, while a more diffuse emission is seen in the afternoon. The auroral double structure seen on the south east limb is an artifact since the image is a summation of two STIS exposures. The double structure in turn shows that the aurora has temporal and/or longitudinal variations. The latter explanation implies that Saturn's magnetic field is not at dipolar as previously believed. The emissions also show a not too smooth oval (towards the central portion of the Earth-facing side of the auroral oval).

The figure on the left is a composite of WFPC2 sub-images of Saturn's north aurora reported by Trauger et al. (1998) . They show temporal variations in Saturn's northern aurora. The top 4 panels show data obtained within about 5 hours on 9 October 1994 during the very first imaging of Saturn in the far-UV (of which the full first exposure is depicted on the left figure). Saturn was undergoing an auroral event, where bright emission remained fixed at 8 AM local time throughout the ~5 hours observing period. The auroral storm was dimming with time. Two observations taken on two separate dates in 1995 revealed very little auroral emission.

Saturn's aurora was known to be bursty from both Voyager and contemporaneous IUE observations, and we think that Voyager may have observed on such morning auroral event. Such events seem to be typical of fast-rotating magnetospheres. (On Jupiter the events are fixed at dawn rather than at 8 AM in magnetic local time.) More information on the temporal behavior of Saturn's aurora can be obtained from analysis of the newer STIS data, because these data was obtained in the STIS time-tag mode. Coordinated observations should be made in the years to come with the Cassini mission in-situ solar wind and magnetospheric measurements.

Because the rotational period of Saturn is highly uncertain, we do not know if the bursts of UV aurora enhancements are constrained in magnetic longitude. With HST imaging observations, made over extended periods and combined with Cassini in-situ data, we may be able to resolve this issue. The STIS images can also be used to resolve corotational versus magnetic local time effects from observing, like for Jupiter, throughout a good part of a planet rotation, as the features rotate from the morning to the afternoon.


Sketch of the magnetosphere of Uranus in 1986 at the time of the Voyager encounter, from Bagenal (1992).

Of the planets with atmosphere-magnetosphere interactions, Uranus has been studied very little. This system is quite interesting, however, like that of Neptune (but Neptune is too far to be observed with HST). Uranus has a highly asymmetric magnetosphere because its magnetic dipole is largely tilted with respect to the spin axis, which in turn has a high obliquity. During the Voyager encounter in 1986 the spin axis was basically aligned with the direction of the incoming solar wind. In a 17.24 hr planetary rotation, the magnetic field rotated around this direction, but the magnetic north pole remained facing the wind.

Sketch of Neptune's magnetosphere during the Voyager encounter in 1989, from Bagenal (1992).

Recently, the magnetosphere of Uranus has resembled more that of Neptune in 1989.

Uranus is now in a new season (spring equinox), and the orientation of the magnetic field changes largely during a planetary rotation, so it is quite different to that of 1986. The magnetic field alternates between pointing towards and away from the Sun. For Neptune, the plasma re-organized from a single to a double plasma sheet structure during a single planet rotation. Uranus may have a more consistent double-lobed structure, or could remain more closed.

This figure was published by F. Herbert and B. Sandel (1994) and it is a map of Uranus' H2 emissions (im magnetic coordinates) made from Voyager UVS observations during the encounter of 1986.

In the Voyager 2 flyby of Uranus in 1986 the UVS experiment was able to detect aurora on Uranus in the molecular hydrogen emissions. This is shown in the figure above. The aurora did not show continuous emissions along ovals. Instead, the aurora were seen as enhanced regions of emissions mapping to the magnetotail. They fall along the predicted north and south auroral ovals mapping to 5 Uranus radii (overplotted in the figure). The north/south brightness ratio of the emissions agreed with diffuse precipitation (scattering of particle into the loss cones) as the relevant mechanism.

Since Uranus has been visited only once by spacecraft, remote sensing studies of airglow and auroral emissions are quite critical for this planet. Uranus relatively low gravity and high (750 K) upper atmospheric temperature produces a highly extended atomic hydrogen corona, and solar and interplanetary Lyman alpha emission is scattered by the upper atmosphere. The far distance from Earth, combined with complex rotational and magnetic field geometries, produce an intrinsically weak Lyman-alpha emission that has been difficult to observe and most enigmatic to interpret (cf, Ballester 1998; Herbert and Sandel 1999).

In the 1980's IUE had detected large time variations in Uranus dayside H Lyman-alpha emission, which were interpreted as the first indicators of the planet's auroral activity (and magnetic field). However, Voyager could not resolve auroral H Lyman-alpha in the sunlit hemisphere, while it did find localized H2 aurora (as shown above). The Voyager data were time-averaged throughout the whole encounter, so they could not yield temporal information. Was the time variability observed by IUE real? Due to bursty, localized aurora?

This figure was presented by Ballester et al. at the 1998 DPS meeting. It is an image of Uranus obtained with the HST STIS instrument with a technique that isolated the H Lyman alpha emissions at 1216 Ang from the bright reflected sunlight at longer wavelengths. The disk is seen in scattered solar and interplanetary Lyman alpha. Unfortunately, the aurora was not active in H Lyman-alpha emission during the HST observations.

The first images resolving H Lyman-alpha on Uranus were taken by STIS in June and September of 1998. The average of the latter observations are shown above. Unfortunately, STIS detected weak disk-averaged Lyman-alpha signal that was at the lowest levels observed by IUE. STIS did not detect significant time variations nor localized aurora. New STIS observations of Uranus have been approved for 2010-2011, to search for changes in the FUV spectrum from enhanced hydrocarbon absorption due to strong upwelling during the spring equinox, and also to image H Lyman-alpha.

Return to Gilda E. Ballester Homepage

Last Revised: 21-Mar-2003 (Uranus on 06-Jan-2010)