Uranus and Neptune, were the first planets discovered by telescopic observation. They are Jovian planets of similar size and much more massive than Earth (by factors of 14.5 and 17) but are still far less massive than Jupiter (by factors of 22 and 18). Both have deep atmospheres dominated by hydrogen, but unlike Jupiter and Saturn, most of their mass consists of rocky and icy components, which are in fluid form due to high interior temperatures. Both have ring systems, a diverse system of satellites, unusual tilted and offset magnetic fields, and similar atmospheric circulations. Their blue colors distinguish them from the pale tans of Jupiter and Saturn. But they also differ significantly from each other; they have very different obliquities (spin axis inclinations relative to their orbital planes), vastly different internal heat fluxes, and remarkably different weather patterns. Their orbital and physical parameters are summarized in Table 1.
Discovery
Uranus, at visual magnitude 5.5, is about as bright as Jupiter’s brightest moon and is close to the limit of visual detectability (magnitude 6). Neptune (magnitude 7.85) is far too dim (by a factor of 30) to be seen by a visual observer. Thus, it is not surprising that the discovery of both planets awaited the development of advanced telescopes. Although observed as early as 1690 by Flamsteed (4) and assumed to be a star, it was not until 13 March 1781 that Uranus was discovered by William Herschel, the best telescope maker of his time. Its extended image implied to Herschel that it was not a star, though at first he believed that it was a comet. Herschel was an amateur at the time, but later became the first president of the Royal Astronomical Society. This was the first planet discovery that can be attributed to a specific individual at a particular time. All of the brighter planets were known to the ancients. Neptune’s discovery was even more unusual. It was the first planet discovered with the aid of mathematical predictions. Prediscovery observations and 40 years of post discovery observations of Uranus showed that its motion was being perturbed from its expected elliptical orbital path. John Couch Adams of England and Urbain Jean Joseph Le Verrier of France were independently able to use these perturbations to infer the location of a previously unknown planet (Neptune). Le Verrier was the more successful in publishing his calculations and convincing an observer to look in his predicted direction (5). Following receipt of Le Verrier’s coordinates, John Galle and his assistant, He-inrich D’Arrest of Germany, first located Neptune on 23 September 1846 and confirmed its position and disc-shaped image the following night, more than a year after the first prediction had been made, but only five days after Le Verrier sent his request to Galle. Various prediscovery observations of Neptune were subsequently identified, the earliest by Galileo in 1612 (5)!
Table 1. Orbital and Physical Parameters of Uranus and Neptune
Formation
According to current theories (6,7), the primordial nebula from which the solar system formed is comprised of 98% hydrogen and helium. In the part of the nebula where the outer planets formed, the remaining 2% was dominated by water, ammonia, methane, and rock, all of which were probably in condensed form in the vicinity of Uranus and Neptune. Most of the condensates were ”ices,” a term applied to methane and ammonia, as well as to water, even when they are not actually in solid form. About a quarter of the solids were rock. This very small fraction of condensed solids played the key role of accreting into asteroid-sized planetesimals (of the order of a kilometer in size), which subsequently accreted into planetary cores. When these cores grew to a critical mass, they were then able to attract significant amounts of nebular gas (mostly hydrogen and helium). The process probably proceeded more slowly for Uranus and Neptune than for Jupiter and Saturn due to lower nebular densities and slower orbital velocities, and thus they were able to capture less nebular gas before the nebula was dissipated by the intense solar wind that developed during the Sun’s early evolution. Uranus and Neptune were thus formed by materials of a higher average density than the materials that formed Jupiter and Saturn. The rocky cores of Uranus and Neptune are each about one-sixth of the planet’s radius and are surrounded by icy material out to 75-80% of the radius; the outer 20-25% is primarily a hydrogen and helium gaseous envelope. During accretion of plan-etesimals, a planet will gain angular momentum as well as mass. A planet in an eccentric orbit will gain the greatest angular momentum at the edges of the accretion zone and will tend to accumulate prograde angular momentum (8). During the formation of both Uranus and Neptune, the last accreted objects might have been relatively large and had sufficient angular momentum to shift the spin axis significantly away from its ”natural” spin direction, which is perpendicular to the ecliptic plane. The last large object that hit Uranus may have been the size of Earth and might have played a role in generating debris from which the satellites may have formed (8). The inclined angular momentum of the Neptune system is also a probable result of impacts; the largest impactor was between 0.1 and 0.5 Earth masses (9). When satellites form from debris created by a late large impactor, most of the debris material is from the impactor, which thus determines the satellite composition.
The regular satellites are those that have nearly circular orbits close to the equatorial plane of the planet. These must have formed after the final large impact that shifted the angular momentum of the planet, and they might have formed from the debris generated by that impact. Any existing material in orbit near the planet at the time of the impact would be perturbed into inclined orbits that would result in collisions and breakup, which would then also contribute to accretion into satellite systems. Irregular satellites that have highly inclined, retrograde orbits are indicative of captured objects. Neptune’s largest satellite, Triton, is plausibly a captured object. It could have been captured during a close approach to Neptune if it had impacted a regular satellite of about 1% of its mass, which would then slow Triton sufficiently to keep it in orbit around Neptune (8).
Interior Structure
The planets’ masses, shapes, rotational rates, and gravitational moments place constraints on their interior structures. Models of both planets can be constructed using three shells (10). Some models use an inner core of rocky material, an intermediate shell of icy material, and an outer layer of gas. The ratio of ice to rock in these models is about 15, and they require an atmosphere enhanced in volatiles by about a factor of 20 compared with solar fractions. Neptune’s rock-ice boundary in these models is about 15-20% of the planet’s radius. The ice-atmosphere boundary is about 80% ofthe planet’s radius. Successful models have also been created by using a gradual transition between the ice shell and the hydrogen-rich outer shell. Considerable uncertainty remains concerning the composition of the deep Neptune atmosphere and of the size of the rocky core (it might be about one Earth mass or considerably smaller). It is thought that the icy shells on both Uranus and Neptune are largely chemically homogeneous and consist of a mixture of ice, rock, hydrogen, and helium (10). Compared to Neptune, Uranus is somewhat less dense and somewhat more centrally condensed. As shown in Fig. 1, this structure has three main differences from that of Jupiter. Jupiter has a much larger gaseous molecular envelope, a large region of metallic hydrogen, and a smaller volume of ices.
Neptune loses internal heat at a rate 30 times larger (in power/unit mass, or luminosity) than what could be provided by radioactive decay of trace elements (the Earth’s internal heat source). Neptune’s heat source, like Jupiter’s, is thought to be primordial heat left from the process of formation. The capture of solid and gaseous materials generated heat that raised the interior temperatures to very high levels (thousands K). That interior heat is still being released into space by Neptune, which emits 2.6 times as much heat as it absorbs from the Sun (12). This means that internal heat loss contributes 1.6 times as much as the reradiated heat absorbed from the Sun. It is somewhat of a mystery that no comparable heat emission is measured for Uranus, which is so similar to Neptune in most other respects. The lack of an internal heat source explains why Uranus’ effective temperature (59 K) is the same as Neptune’s even though it is much closer to the Sun (where sunlight is about 2.5 times as intense). Estimates of the heats of formation for Uranus and Neptune and the estimated cooling since formation suggest that the remaining primordial heat is more than sufficient to explain the present luminosity of Neptune. One explanation for the low luminosity of Uranus is that there are compositional gradients in the interior of Uranus that lead to suppressed convection (10), which reduces heat transfer efficiency and thus the temperature of its outer atmospheric layers. Such gradients might have arisen in connection with the giant impact that is presumably responsible for Uranus’ spin axis inclination. An alternate theory (13) suggests that different external solar forcing produced by the high obliquity of Uranus has made the heat transfer process on Uranus much more efficient than on Neptune and resulted in a more rapid loss of internal heat.
Figure 1. Approximate interior structures of Uranus and Neptune compared to Jupiter (11). Rocky and icy material may be mixed on Uranus and Neptune. The gas layer on Uranus is probably thicker than that on Neptune because Neptune is slightly denser. The radius at which metallic hydrogen is formed on Jupiter has been revised recently. Hub-bard (7) now places it at about 0.8 RJ.
The generation of an offset dipole magnetic field, that has a significant quadrupole component requires the existence of a conducting fluid layer in convection. This layer might be bounded at the deepest level by the point below which the interior is stably stratified, and thus not convecting, and the point above which the fluid interior is not electrically conducting. If the entire interior fluid of Neptune exhibited the same angular rotation as that observed by clouds at the top of the atmosphere, then the J4 gravitational expansion coefficient would be positive, whereas the observed value is negative. This implies that the differential rotation of the upper atmosphere is a superficial effect that does not involve a significant fraction of the planet’s mass. However, this analysis does not place a well-defined boundary on how deep that flow could extend.
Atmospheres
Neptune and Uranus have similar compositions, similar tropospheric temperature structures, and similar styles of zonal circulation, but different cloud patterns and great differences in weather activity. The basic atmospheric parameters for Uranus and Neptune are summarized in Table 2. Composition. Hydrogen, helium, and methane are the most prominent components of the atmospheres of Uranus and Neptune. Ammonia and probably H2S and water are present in layers below the region accessible by optical spectroscopy. Hydrocarbons of various types are observed in the upper atmosphere, a result of UV-induced chemistry involving the breakdown of CH4.
The presence of hydrogen in any outer planet atmosphere was established first for Uranus, using the pressure induced 3-0 S(1) line of molecular hydrogen at 825.8 nm, which was first measured by Kuiper (15), but first identified by Herzberg (16). Molecular hydrogen occurs in two forms, the ortho form in which the nuclear spin vectors of the two atoms are parallel, and the para form in which the nuclear spin vectors are antiparallel. At high temperatures, the two forms reach an equilibrium concentration of three ortho molecules to one para molecule. That is called normal hydrogen. Hydrogen convected to lower temperatures will retain the normal mixing ratio for a long time because of the very weak interaction between the two nuclei in the absence of a catalyst. Normal hydrogen was expected on Jovian planets because of rapid mixing that should overwhelm the slow conversion process. With an effective catalyst and sufficient time, the two forms will reach an equilibrium distribution that depends on temperature, and with sufficiently rapid conversion, large effects on specific heat, atmospheric buoyancy, and temperature lapse rate can occur. When conversion is slow, the straints are conflicting (17). The hydrogen quadrupole spectrum and the collision-induced dipole spectrum measured in the infrared are both approximately consistent with thermal equilibrium for the two forms of hydrogen at the temperature at which the spectral lines are formed. But the measured temperature lapse rate is more consistent with that expected for normal hydrogen. This conflict might be resolved with the concept of stratified convection layers (18) in which a given gas parcel resides in a thin layer long enough to reach ortho-para equilibrium, even though the convective overturn time is relatively short. It is suggested that condensation of CH4 might produce a stepwise stratification of mean molecular weight that would increase stability, playing the same role as salinity variations in generating layered convection in terrestrial oceans. An alternative suggestion by Flasar (19) is that the lapse rate for P>700mb is actually stable because of methane condensation that introduces a buoyancy gradient capable of supporting what otherwise would appear to be a superad-iabatic (highly unstable) gradient for equilibrium hydrogen. Flasar suggests that the close agreement between the measured lapse rate and the adiabatic gradient for normal hydrogen is just a coincidence.
Table 2. Atmospheric Composition3
The He/H2 ratio is constrained best from the analysis of far-infrared spectra. The collision-induced dipole absorption of H2-H2 and H2-He proves most of the continuum opacity at long wavelengths and is sufficiently well understood theoretically that it is possible to infer the ratio by fitting the thermal IR spectra that are also consistent with temperature profiles determined by Voyager 2 radio occultation measurements. The inferred helium mole fraction (number or volume fraction rather than mass fraction) in the upper atmosphere of Uranus is 15.2% + 3.3% (20). Neptune’s is a little higher at 19.0%7 3.2% (21). As expected from formation theories, both are within errors of the solar value of 16%. This contrasts with Saturn, which has only about one-fourth of the solar fraction, presumably due to rainout of He in Saturn’s deep interior where it is thought that He becomes insoluble in metallic hydrogen.
Spectroscopic observations indicate that methane (CH4) is enhanced relative to solar abundance values by a factor of 25-30 in regions where it ought to be well mixed (22). However, because methane condenses to form clouds in both atmospheres, methane abundance above the condensation level can be greatly reduced, variable, and difficult to estimate. Initially, it was thought that the tropopause would act as a cold trap, limiting the stratospheric methane mixing ratio to values no greater than the saturation mixing ratio at the tropopause. Yet Voyager Ultra-Violet Spectrometer data for Neptune imply that stratospheric mixing ratios are at least 10 times this limit (23). This excess, termed oversaturation, suggests stronger vertical mixing on Neptune compared to Uranus, where stratospheric oversaturation is not observed. The large tropospheric enhancement of methane relative to the solar mixing ratio for both Uranus and Neptune is a likely consequence of the planetary formation process. The large fraction of icy materials accreted by these planets should also have resulted in similar enhancements of water and ammonia, even though little has so far been observed in the atmospheres.
Microwave atmospheric observations imply a significant depletion of NH3 relative to solar values, by a factor of 100-200 in the 150-200 K region of Uranus’ atmosphere (24). A possible explanation is that NH3 is lost to an extensive NH3-H2O solution cloud or that NH3 is lost to the formation of a cloud of NH4SH (ammonium hydrosulfide). The depletion might be completely accomplished by formation of an NH4SH cloud if the H2S/NH3 ratio is enhanced by a factor of 4 compared to the solar value of 0.2. It seems to require very large enhancement factors of water for the NH3-H2O solution cloud to play a major role in the depletion of NH3. The necessary enhancement of H2S might have resulted from accretion of chondritic meteorites, in which sulfide minerals are found but nitrogen incorporation is not significant. An alternative hypothesis by Lewis and Prinn (25) is that Uranus never acquired much nitrogen because uncondensed N2 and CO were the dominant chemical forms of N and C in the solar nebula (rather than NH3 and CH4) in the region of the terrestrial planets and also, as a result of rapid mixing in the outer parts of the nebula where thermal equilibrium compounds would be NH3 and CH4. So far, there is no direct observation of H2O or NH3 on either Uranus or Neptune, nor of H2S on any Jovian planet except Jupiter. If H2O is present at solar mixing ratios, condensation clouds might be found at pressures of the order of 100 bars. If enhanced 65 times solar, water would condense at a temperature of 647 K, but would not condense at all beyond that enhancement (24).
Spectral Characteristics of Uranus and Neptune. Methane absorption dominates the visible and near-IR disk-averaged spectra of both Uranus and Neptune (upper part of Fig. 2). At wavelengths below 1 micron, there are numerous methane bands of increasingly greater absorption from the green to the red and near IR. The absorption of red light by methane contributes to the blue colors of Uranus and Neptune. Neptune is bluer than Uranus because of increased absorption in the window regions between the methane bands. This might mean that the visible cloud layer that controls the amount of light reflected backward is thicker and brighter on Uranus, whereas the corresponding cloud layer on Neptune is more transparent and allows more of the light to be absorbed by the deeper atmosphere. Alternatively, the cloud itself might provide the extra absorption needed on Neptune (22); if so, the cloud must absorb even more effectively at wavelengths between 1 and 2.5 microns (29).
The basic disk-averaged characteristics of Neptune and Uranus at near-IR wavelengths are illustrated in the upper right panel of Fig. 2, which displays the ground-based observations of Fink and Larsen (27). The relatively low disk-integrated albedos of both planets are due to the significant amount of methane overlying the visible cloud layer. The albedo peaks that occur in windows of relatively weak absorption, for example, 1.3 and 1.6 mm, are narrower for Uranus than for Neptune, indicating that Uranus has a clearer atmosphere with fewer high altitude hazes to reflect photons that would otherwise be absorbed by underlying methane. In the 1-1.8 mm region methane is the dominant absorber, whereas hydrogen collision-induced absorption (CIA) is the major absorber in the 1.85-2.2 mm range.
The effects of these absorbers on the penetration depths of photons into the atmospheres of these planets are illustrated in the middle panel of Fig. 2. This displays the wavelength-dependent pressure levels at which a unit albedo re-
Figure 2. (Top) Reflectivity spectra of Nepture (26-28). (Middle) Penetration depth of photons into Neptune’s atmosphere (30). (Bottom) Appearance of Neptune as a function of wavelength as recorded by HST imaging of WFPC2 and NICMOS (30).
either of the others. (Penetration depths for Uranus are roughly the same as those for Neptune.) In Fig. 2, we see that the 1.27 mm and 1.59-mm windows probe all the way into the putative H2S visible cloud deck near 3.2 bars. Were that cloud composed of high-albedo particles, we would expect albedo values near unity, rather than values in the 0.05-0.08 range. On the basis of the near-IR spectrum alone, the cloud is very dark, more transparent, or deeper than indicated by other methods. The exact levels sensed in the strongly absorbing 2.3-um methane band depend on stratospheric methane mixing ratios, which are not well constrained by current observations. Sample images of Neptune obtained by the Hubble Space Telescope Wide Field/Planetary Camera 2 and Near Infrared
Camera and Multi-Object Spectrometer are shown in the bottom panel of Fig. 2 (30). These illustrate the dominance of Rayleigh scattering at short wavelengths and the effects of methane absorption at long wavelengths.
Neptune has both dark and bright discrete features, but the only discrete features on Uranus seem to be bright features. Dark features on Neptune have the greatest contrast at blue wavelengths (see lower left of Fig. 2), but it is typically a rather small 2-10%. The identity of the blue absorber that creates this contrast is unknown. The contrast between bright clouds and the background atmosphere can also be rather small at short wavelengths, where Rayleigh scattering is important, but it is dramatically enhanced at wavelengths where methane absorption is strong. On Neptune, the contrast can exceed 500:1 at 2 microns (30). The maximum contrast for bright clouds on Uranus is seen at about 1.9 microns and is about 1.8:1 in raw images, but estimated at about 10:1 at high spatial resolution (31). Relatively less contrast is observed on Uranus because its bright cloud features do not reach to pressures as low as those on Neptune. Temperature Structure. The temperature structures of Uranus and Neptune are very similar, as illustrated in Fig. 3 by dotted and solid curves, respectively. There is remarkably little latitudinal variation in this structure. Given the 98° obliquity (pole inclination relative to orbital normal) of Uranus and Neptune’s 29°, the absence of significant pole-to-pole gradients (for Uranus) or equator to pole gradients (for Neptune) implies some heat transport or compensation effect. The effective emission temperatures on both planets are very low (about 59 K), leading to very long radiative time constants of about 5 x 109 seconds at 400500 mb (32). This is the ratio of the thermal energy content of an atmosphere to the radiative cooling rate for one local scale height. This radiative time constant is twice as long as Uranus’ orbital period. Thus, seasonal variations on Uranus are strongly damped and phase shifted (delayed) by close to 1/4 year. In fact, even though the North Pole (IAU convention) of Uranus had been in darkness for 20 years, the Voyager IRIS instrument found no measurable difference in temperature structure between the two polar regions (33). Voyager 2 arrived close to the Southern Hemisphere solstice, a time when hemispheric thermal contrast should be at a minimum because of the phase lag, and thus it was not able to measure the amplitude of the seasonal response. Because of Uranus’ obliquity, the average solar heating for an entire Uranian year is greater at the poles than at the equator. However, seasonal model predictions of a somewhat cooler equator were not confirmed by Voyager observations. The main latitudinal variation in temperature on both Uranus and Neptune is consistent with the thermal wind equation. This equation is a proportionality between vertical wind shear and horizontal temperature gradients, which is valid for vertical hydrostatic balance (between pressure and gravity) and horizontal balance between coriolis forces and horizontal pressure gradients. The sign of the derived vertical wind shear indicates that the zonal winds decay as height increases (discussed later). Cloud Structure. Our current limited understanding of cloud structures on Uranus and Neptune is illustrated in Fig. 3. The saturation vapor pressure curve for a CH4 mixing ratio of 2% suggests a methane cloud base near 1.4 bars on Neptune and 1.2 bars on Uranus, although nucleation at somewhat higher altitudes is likely because it would probably require some degree of supersatura-tion (34). Radio occultation observations of refractivity gradients in this region agree with the 1.2 bars on Uranus (35) but suggest a cloud base near 1.9 bars on Neptune (36). Neptune’s stratospheric hazes (34,37,38), which significantly reduce Neptune’s shortwave albedo, have a relatively low total equatorial optical depth of ~0.02 at 0.75 mm (39). Larger optical depths found for the global average might be due to contributions from unresolved high-altitude, isolated, bright methane clouds. The optical depth of the methane cloud at red wavelengths is relatively small; on Neptune, it ranges from about 0.05 near the equator to 0.3 near 25° S (22), and on Uranus from 0.4 (40) to 1.3-1.5 (41,42); there is evidence for lower values (several tenths) for central disk observations, and higher values are inferred from disk-integrated values (43). Much larger opacities are possible for discrete features. The presence of an H2S cloud is inferred from microwave spectra that probe the deep atmosphere; enhancement by a factor of 30 relative to the solar mixing ratio is not directly measured but is inferred to explain the very low abundance of NH3 (44,45). In this scenario, the formation of a very deep NH4SH cloud consumes the excess NH3. When the corresponding H2S condensation curve is computed, we find that it intersects the temperature profile in the 6-9 bar region, above which we might expect to see a cloud of H2S. The presence of an opaque cloud at pressures deeper than 3.6 to 3.8 bars is inferred from hydrogen quadrupole line widths (40). However, attempts to find direct evidence of H2S in optical spectra have failed (46). Thus, we cannot completely rule out the possibility that both NH3 and H2S are very depleted, as advocated by Romani et al. (37) and that the visible cloud deck is a thin ammonia ice cloud or that no cloud at all is present in the 3-10 bar region, though that would conflict with the quadrupole results. An ammonia cloud, if present, might also form near the 7-bar level. Current models also have difficulty explaining thermal infrared observations of Neptune, which seem to require that the methane cloud be much more opaque at a significantly higher altitude or horizontally heterogeneous (21).
Figure 3. Vertical structure of temperatures and cloud and haze layers. The Neptune temperature profile (solid) differs from the Uranus profile (dotted) mainly in the stratosphere where a larger population of hydrocarbon hazes on Neptune leads to more heat absorption and a warmer stratosphere.
Uranus is clearer than Neptune, perhaps because it lacks internal heat to drive the mixing needed to keep particles suspended for long times. The global average methane haze layer on Neptune is relatively thin (~ 0.1), but the optical depths of discrete bright cloud features are probably much greater. On Neptune, these features reach 50-100 km above the mean cloud level, reaching to pressures of 100-200 mbar. On Uranus, discrete bright clouds do not rise much above 500 mb (31).
Horizontal cloud structures for Uranus and Neptune are illustrated by sample Voyager and HST images in Figs. 4-7. Voyager imaging of Uranus (Fig. 4) in 1986 displayed a rather bland appearance. An approximate true-color image (left image in Fig. 4) displays neither banding nor discrete cloud features. The extremely enhanced false-color image in the middle of Fig. 4 shows that there was a latitudinal variation in the amount of UV-absorbing haze in the atmosphere; the greatest absorption occurred near the visible pole of Uranus (IAU south) where it appears as relatively orange. Discrete cloud features are illustrated in the time sequence of images at the right of Fig. 4, which were made using Voyager’s orange filter.
Hubble Space Telescope images of Uranus made in 1997, 11 years after the Voyager encounter, reveal the first discrete cloud features in its Northern Hemisphere. One feature is barely visible near the right-hand limb in the upper left image, which was made with a 619-nm filter (the wavelength of a weak methane band). Much better contrast between discrete cloud features and the background atmosphere was obtained at near-IR wavelengths using the HST NICMOS camera, as illustrated in the middle and right-hand images of Fig. 5.
![[JPL P29478]. Voyager 2 images of Uranus taken on 17 January 1986 using blue, green, and orange filters to make the true-color composite (left), which displays a virtually featureless disk. UV, violet, and orange filtered images were shown as blue, green, and red components in the extremely enhanced false-color image (middle), which reveals polar bands of UV-absorbing haze particles, centered on the South Pole of Uranus (IAU convention). Even in this view, no discrete cloud features are apparent. [Voyager 2 JPL P29467.] The right-hand time sequence of orange-filtered images from 14 January 1986 shows the motion of two small bright streaky clouds that were the first discrete features ever seen on Uranus. Uranus is rotating counterclockwise in this view, as are the clouds, though more slowly than Uranus' interior, revealing that low latitude winds on Uranus are retrograde, as are the winds on Neptune. This figure is available in full color at http://www.mrw.interscience.wiley.com/esst.](http://lh3.ggpht.com/_1wtadqGaaPs/TF6V7w31ZsI/AAAAAAAAPI8/LYKgnrK5GGM/s1600-h/tmp424_thumb3.jpg "[JPL P29478]. Voyager 2 images of Uranus taken on 17 January 1986 using blue, green, and orange filters to make the true-color composite (left), which displays a virtually featureless disk. UV, violet, and orange filtered images were shown as blue, green, and red components in the extremely enhanced false-color image (middle), which reveals polar bands of UV-absorbing haze particles, centered on the South Pole of Uranus (IAU convention). Even in this view, no discrete cloud features are apparent. [Voyager 2 JPL P29467.] The right-hand time sequence of orange-filtered images from 14 January 1986 shows the motion of two small bright streaky clouds that were the first discrete features ever seen on Uranus. Uranus is rotating counterclockwise in this view, as are the clouds, though more slowly than Uranus' interior, revealing that low latitude winds on Uranus are retrograde, as are the winds on Neptune. This figure is available in full color at http://www.mrw.interscience.wiley.com/esst.")
Figure 4. [JPL P29478]. Voyager 2 images of Uranus taken on 17 January 1986 using blue, green, and orange filters to make the true-color composite (left), which displays a virtually featureless disk. UV, violet, and orange filtered images were shown as blue, green, and red components in the extremely enhanced false-color image (middle), which reveals polar bands of UV-absorbing haze particles, centered on the South Pole of Uranus (IAU convention). Even in this view, no discrete cloud features are apparent. [Voyager 2 JPL P29467.] The right-hand time sequence of orange-filtered images from 14 January 1986 shows the motion of two small bright streaky clouds that were the first discrete features ever seen on Uranus. Uranus is rotating counterclockwise in this view, as are the clouds, though more slowly than Uranus’ interior, revealing that low latitude winds on Uranus are retrograde, as are the winds on Neptune.
![(Left) [STScI PRC97-36b, NASA and H. Hammel]. HST WFPC2 images of Uranus on 31 July and 1 August 1997. Although little contrast is seen at 547 nm (blue), a banded structure and the first discrete Northern Hemisphere cloud are visible at 619 nm (upper image, colored red). (Middle) [Space Telescope Science Institute STSCI-PRC97-36A and E. Karkoschka]. This false-color 1997 image is a composite of near-IR images taken by the Near Infrared Camera and Multi-Object Spectrometer (NICMOS) at wavelengths of 1.1 mm (shown as blue), 1.6 mm (shown as green), and 1.9 mm (shown as red). Absorption by methane gas limits the depth at which reflected sunlight can still be seen at 1.1 and 1.6 mm, and absorption by hydrogen is most significant at 1.9 mm. The blue exposure probes atmospheric levels down to a few bars, responding to scattering by aerosols and by atmospheric molecules above this level. The green component is least sensitive to methane absorption, sensing down to 10 bars or more, but sees much less Raleigh scattering per bar than the blue component, so that a dark absorbing cloud near 3 bars would result in more blue than green in regions that were clear above the cloud, perhaps accounting for the blue color at midlatitudes. The red component can only sense down to about the 2-bar level, and sees the least contribution from Rayleigh scattering, so that very little red is seen in regions that are clear to 3 bars. The green color around the South Pole suggests significant local haze opacity at pressures near 2-3 bars. The red color of the discrete features near the northern (right) limb indicate relatively high-altitude clouds that reflect sunlight before much absorption has taken place. The curved arcs in the central image indicate motions in 90 minutes of cloud features and eight of the 10 small satellites discovered by Voyager 2. The area outside the rings was enhanced to make the satellites more visible. The images also show the bright epsilon ring, which is wider and brighter in the upper part of the image, and two fainter inner rings. (Right) [STScI-PR98-35, NASA and E. Karkoschka]. This Hubble Space Telescope near-IR image of Uranus on 8 August 1998 shows a number of new cloud features in both hemispheres. The false-color image was created in a manner similar to the first, except that the rings and satellites were not separately brightened. This figure is available in full color at http://www.mrw.interscience. wiley.com/esst.](http://lh4.ggpht.com/_1wtadqGaaPs/TF6WCYI81-I/AAAAAAAAPJE/yx5OY5qFW1g/s1600-h/tmp425_thumb3.jpg "(Left) [STScI PRC97-36b, NASA and H. Hammel]. HST WFPC2 images of Uranus on 31 July and 1 August 1997. Although little contrast is seen at 547 nm (blue), a banded structure and the first discrete Northern Hemisphere cloud are visible at 619 nm (upper image, colored red). (Middle) [Space Telescope Science Institute STSCI-PRC97-36A and E. Karkoschka]. This false-color 1997 image is a composite of near-IR images taken by the Near Infrared Camera and Multi-Object Spectrometer (NICMOS) at wavelengths of 1.1 mm (shown as blue), 1.6 mm (shown as green), and 1.9 mm (shown as red). Absorption by methane gas limits the depth at which reflected sunlight can still be seen at 1.1 and 1.6 mm, and absorption by hydrogen is most significant at 1.9 mm. The blue exposure probes atmospheric levels down to a few bars, responding to scattering by aerosols and by atmospheric molecules above this level. The green component is least sensitive to methane absorption, sensing down to 10 bars or more, but sees much less Raleigh scattering per bar than the blue component, so that a dark absorbing cloud near 3 bars would result in more blue than green in regions that were clear above the cloud, perhaps accounting for the blue color at midlatitudes. The red component can only sense down to about the 2-bar level, and sees the least contribution from Rayleigh scattering, so that very little red is seen in regions that are clear to 3 bars. The green color around the South Pole suggests significant local haze opacity at pressures near 2-3 bars. The red color of the discrete features near the northern (right) limb indicate relatively high-altitude clouds that reflect sunlight before much absorption has taken place. The curved arcs in the central image indicate motions in 90 minutes of cloud features and eight of the 10 small satellites discovered by Voyager 2. The area outside the rings was enhanced to make the satellites more visible. The images also show the bright epsilon ring, which is wider and brighter in the upper part of the image, and two fainter inner rings. (Right) [STScI-PR98-35, NASA and E. Karkoschka]. This Hubble Space Telescope near-IR image of Uranus on 8 August 1998 shows a number of new cloud features in both hemispheres. The false-color image was created in a manner similar to the first, except that the rings and satellites were not separately brightened. This figure is available in full color at http://www.mrw.interscience. wiley.com/esst.")
Figure 5. (Left) [STScI PRC97-36b, NASA and H. Hammel]. HST WFPC2 images of Uranus on 31 July and 1 August 1997. Although little contrast is seen at 547 nm (blue), a banded structure and the first discrete Northern Hemisphere cloud are visible at 619 nm (upper image, colored red). (Middle) [Space Telescope Science Institute STSCI-PRC97-36A and E. Karkoschka]. This false-color 1997 image is a composite of near-IR images taken by the Near Infrared Camera and Multi-Object Spectrometer (NICMOS) at wavelengths of 1.1 mm (shown as blue), 1.6 mm (shown as green), and 1.9 mm (shown as red). Absorption by methane gas limits the depth at which reflected sunlight can still be seen at 1.1 and 1.6 mm, and absorption by hydrogen is most significant at 1.9 mm. The blue exposure probes atmospheric levels down to a few bars, responding to scattering by aerosols and by atmospheric molecules above this level. The green component is least sensitive to methane absorption, sensing down to 10 bars or more, but sees much less Raleigh scattering per bar than the blue component, so that a dark absorbing cloud near 3 bars would result in more blue than green in regions that were clear above the cloud, perhaps accounting for the blue color at midlatitudes. The red component can only sense down to about the 2-bar level, and sees the least contribution from Rayleigh scattering, so that very little red is seen in regions that are clear to 3 bars. The green color around the South Pole suggests significant local haze opacity at pressures near 2-3 bars. The red color of the discrete features near the northern (right) limb indicate relatively high-altitude clouds that reflect sunlight before much absorption has taken place. The curved arcs in the central image indicate motions in 90 minutes of cloud features and eight of the 10 small satellites discovered by Voyager 2. The area outside the rings was enhanced to make the satellites more visible. The images also show the bright epsilon ring, which is wider and brighter in the upper part of the image, and two fainter inner rings. (Right) [STScI-PR98-35, NASA and E. Karkoschka]. This Hubble Space Telescope near-IR image of Uranus on 8 August 1998 shows a number of new cloud features in both hemispheres. The false-color image was created in a manner similar to the first, except that the rings and satellites were not separately brightened.
Voyager 2 images of Neptune in 1989 (Fig. 6) provided a rich bounty of detail concerning horizontal cloud structure and revealed many discrete bright and dark cloud features, as well as bright and dark bands. Many of these structures were unique to Neptune and have not been seen before or since. HST imaging of Neptune beginning in 1994 showed some cloud banding similar to the 1989 Voyager images, but many changes were found, including the disappearance of Neptune’s Great Dark Spot (discussed later). Weather Phenomena on Neptune
Dark Spots. Voyager identified two prominent dark oval features in Neptune’s Southern Hemisphere; both are thought to be manifestations of anticy-clonic eddies, as illustrated in Figs. 6 and 7. (Anticyclones rotate counterclockwise in the Southern Hemisphere.) The Great Dark Spot (GDS) was about an Earth diameter long. It was seen for about 8 months during Voyager’s approach to Neptune but has not been seen since. It exhibited several unusual dynamic features; some of them are illustrated in the middle panel of Fig. 6. Its shape and orientation varied cyclically in a period of 193 hours (47). The longitudinal width varied from 31 to 45o, and its latitudinal extent varied from 12.5 to 15.5°. This behavior was successfully reproduced using a Kiva vortex model (49). This model predicts anticyclonic circulation (counterclockwise in the Southern Hemisphere), although no direct measurements have been able to test this prediction. The GDS also drifted toward the equator at a steady rate of 1.24°/ month (47); this would have placed it on the equator in November 1990, although theoretical models imply that it would have dissipated before that point (50). This behavior is in marked contrast to that of Jupiter’s Great Red Spot, which has remained at essentially a fixed latitude for centuries. The GDS was accompanied by a prominent bright companion cloud positioned at its southern boundary (Figs. 6 and 7). In ground-based methane band imaging at the time of the Voyager encounter, the companion was the brightest feature on the planet (51) and seems to be similar to orographic clouds. It is likely caused by vertical deflection of flow around the GDS, producing methane condensation during the upward part of the flow and evaporation on the downward return leg (52). The first Hubble Space Telescope images that had sufficient image quality to show the subtle contrast of the GDS at blue wavelengths and sufficient longitudinal coverage were made in 1994, after the Hubble repair mission. The GDS was not seen at that time, and no other southern dark spot has been seen since.
![(Left) [JPL P34606]. Voyager 2 image of Neptune. Color composite formed from green, blue, and red filtered images taken on 18 August 1989. South is down, and the South Pole is tipped toward Earth. The Great Dark Spot (GDS) seen at the center of the image is about 13,000 km (about the diameter of Earth) by 6,600 km, and at a latitude of 18°S at the time of this image. The bright cloud to the south of the GDS, termed the companion, is at relatively high altitude compared to the blue features, and was very prominent in ground-based images made at methane-band wavelengths (51). The bright clouds at the edge of the dark circumpolar band, called south polar features (SPF), are at a latitude of about 67°S; these are highly variable on short timescales and exhibit vertical relief of about 75-150 km, as inferred from shadows seen in Voyager images. (Right) [JPL P34668]. This image of Neptune's south polar regions near 68°S, made on 23 August 1989, shows the first cloud shadows ever recorded by Voyager on any planet (the Sun is to the left). (Middle) [JPL P34610]. This time sequence of remapped images of Neptune's GDS, from top to bottom, provides views at intervals of about 18 hours, revealing its strange wobble and shape changes, which occur in a period of 193 hours. Its mean dimensions were approximately 38° in longitude and 14° in latitude and had modulation amplitudes of 7.4 and 1.5°, respectively (47). It also drifted toward the equator at a rate of about 1° per month. This figure is available in full color at http://www.mrw.interscience.wiley.com/esst.](http://lh4.ggpht.com/_1wtadqGaaPs/TF6WFMkV9mI/AAAAAAAAPJM/8l3xYbxYKBY/s1600-h/tmp426_thumb3.jpg "(Left) [JPL P34606]. Voyager 2 image of Neptune. Color composite formed from green, blue, and red filtered images taken on 18 August 1989. South is down, and the South Pole is tipped toward Earth. The Great Dark Spot (GDS) seen at the center of the image is about 13,000 km (about the diameter of Earth) by 6,600 km, and at a latitude of 18°S at the time of this image. The bright cloud to the south of the GDS, termed the companion, is at relatively high altitude compared to the blue features, and was very prominent in ground-based images made at methane-band wavelengths (51). The bright clouds at the edge of the dark circumpolar band, called south polar features (SPF), are at a latitude of about 67°S; these are highly variable on short timescales and exhibit vertical relief of about 75-150 km, as inferred from shadows seen in Voyager images. (Right) [JPL P34668]. This image of Neptune's south polar regions near 68°S, made on 23 August 1989, shows the first cloud shadows ever recorded by Voyager on any planet (the Sun is to the left). (Middle) [JPL P34610]. This time sequence of remapped images of Neptune's GDS, from top to bottom, provides views at intervals of about 18 hours, revealing its strange wobble and shape changes, which occur in a period of 193 hours. Its mean dimensions were approximately 38° in longitude and 14° in latitude and had modulation amplitudes of 7.4 and 1.5°, respectively (47). It also drifted toward the equator at a rate of about 1° per month. This figure is available in full color at http://www.mrw.interscience.wiley.com/esst.")
Figure 6. (Left) [JPL P34606]. Voyager 2 image of Neptune. Color composite formed from green, blue, and red filtered images taken on 18 August 1989. South is down, and the South Pole is tipped toward Earth. The Great Dark Spot (GDS) seen at the center of the image is about 13,000 km (about the diameter of Earth) by 6,600 km, and at a latitude of 18°S at the time of this image. The bright cloud to the south of the GDS, termed the companion, is at relatively high altitude compared to the blue features, and was very prominent in ground-based images made at methane-band wavelengths (51). The bright clouds at the edge of the dark circumpolar band, called south polar features (SPF), are at a latitude of about 67°S; these are highly variable on short timescales and exhibit vertical relief of about 75-150 km, as inferred from shadows seen in Voyager images. (Right) [JPL P34668]. This image of Neptune’s south polar regions near 68°S, made on 23 August 1989, shows the first cloud shadows ever recorded by Voyager on any planet (the Sun is to the left). (Middle) [JPL P34610]. This time sequence of remapped images of Neptune’s GDS, from top to bottom, provides views at intervals of about 18 hours, revealing its strange wobble and shape changes, which occur in a period of 193 hours. Its mean dimensions were approximately 38° in longitude and 14° in latitude and had modulation amplitudes of 7.4 and 1.5°, respectively (47). It also drifted toward the equator at a rate of about 1° per month.
![(Left) [Space Telescope Science Institute PRC-98-34 and H. Hammel]. HST composite color image made from images at blue, orange, and near-IR wavelengths and corrected for limb darkening. (Middle) [STScI and L. Sromovsky]. This quartet of false-color images made using HST shows persistent banded structure and an increase in the number of bright clouds between 1996 and 1998. (Right) [STScI and L. Sromovsky]. Views of dark spots seen by Voyager (GDS in top row) and HST (NGDS-32 and NGDS-15 in bottom row). The HST images (48) are composites of several images taken at slightly different times. This figure is available in full color at http://www.mrw.interscience. wiley.com/esst.](http://lh5.ggpht.com/_1wtadqGaaPs/TF6WLtf4imI/AAAAAAAAPJU/0xAI_RT61a8/s1600-h/tmp427_thumb3.jpg "(Left) [Space Telescope Science Institute PRC-98-34 and H. Hammel]. HST composite color image made from images at blue, orange, and near-IR wavelengths and corrected for limb darkening. (Middle) [STScI and L. Sromovsky]. This quartet of false-color images made using HST shows persistent banded structure and an increase in the number of bright clouds between 1996 and 1998. (Right) [STScI and L. Sromovsky]. Views of dark spots seen by Voyager (GDS in top row) and HST (NGDS-32 and NGDS-15 in bottom row). The HST images (48) are composites of several images taken at slightly different times. This figure is available in full color at http://www.mrw.interscience. wiley.com/esst.")
Figure 7. (Left) [Space Telescope Science Institute PRC-98-34 and H. Hammel]. HST composite color image made from images at blue, orange, and near-IR wavelengths and corrected for limb darkening. (Middle) [STScI and L. Sromovsky]. This quartet of false-color images made using HST shows persistent banded structure and an increase in the number of bright clouds between 1996 and 1998. (Right) [STScI and L. Sromovsky]. Views of dark spots seen by Voyager (GDS in top row) and HST (NGDS-32 and NGDS-15 in bottom row). The HST images (48) are composites of several images taken at slightly different times.
The second prominent dark spot observed by Voyager (DS2 in Fig. 7) was much smaller than the GDS (about 20° in longitude by 6° in latitude), and had several unique dynamic characteristics of its own (47). Its latitudinal position oscillated at an amplitude of 2.4° about its mean latitude of 52.5° S in a period of 865 hours. During this oscillation, its zonal speed increased in step with the zonal wind shear, resulting in a 48° longitudinal oscillation amplitude. Bright clouds formed near the center of the dark spot during the northern half of the oscillation and declined during the southern half. Strangest of all, its mean drift rate relative to Neptune’s interior matched to great precision the mean drift rate of the south polar features (discussed later); the two mean positions remained on opposite sides of Neptune and at mean latitudes differing by 20°. The DS2 feature has not been seen since the Voyager encounter.
Two new dark spots were discovered in HST images (Fig. 7); both were in the Northern Hemisphere. NGDS-32 was discovered first in 1994 images (53) at about 32° N latitude, and a bright companion cloud was seen at its equatorward edge. This dark spot is comparable in size to the GDS but is harder to observe because the latitude circle where it is found is close to the northern limb. NGDS-15 was discovered first in 1996 HST images (54) close to latitude 15° N. It is somewhat smaller than NGDS-32 and spans about 20° in longitude and 11° in latitude. It is unusual in having no bright companion cloud. NGDS-15 is also unusual in being close to the latitude at which modeled dark spots tend to dissipate rapidly. Both NGDS-32 and NGDS-15 differ from the Voyager GDS by exhibiting no measurable latitudinal drift, a characteristic that is baffling to modelers because model dark spots always drift equatorward on Neptune. The modeled drift rate depends on the latitudinal gradient in the zonal wind speed. If that gradient is locally perturbed in the vicinity of the dark spot latitudes, it may be possible to get consistency between model and measured behavior.
South Polar Features. Rapidly varying small bright features clustered near 62-70° latitude, termed south polar features, were the first clouds Voyager observed that had shadows (see Fig. 6). These features were observed almost exclusively in a region less than 180° wide, and though the mean position where they were seen drifted only slightly relative to the interior, the individual cloud elements actually moved relatively quickly at a prograde rate of about 200 m/s. Their lifetimes were so short that they could not be tracked during a full rotation of Neptune. At low resolution, the features blend together so that they sometimes appear to form a plume (the southernmost prominent bright cloud in the left image of Fig. 6). But in high-resolution images, they can be seen as small individual elements (see the right image of Fig. 6). The vertical relief implied by shadows is 50-150 km (51).
Waves and Bands. Cloud patterns on Neptune suggest the existence of atmospheric wave motions. Voyager imaging showed that near 20° S latitude, there was a clustering of bright cloud features at two regions spaced approximately 180° apart in longitude; one region was in the vicinity of the GDS. This suggests a wave that has two complete oscillations within 360° of longitude (wave number 2). Neither the pattern nor the GDS have been seen since the Voyager encounter. The phase-locked motions of the region of SPF formation and DS2 suggest a similar wave interaction. A more obvious wave example is the dark band between 55° S and 65° S, visible in blue-filtered images (right group in Fig. 7). When viewed in polar projection, the band appears as a circle offset from Neptune’s South Pole, leading to a 2° sinusoidal modulation in latitude of the wave boundaries as a function of longitude. The wave can be seen as a difference in tilts of the band in the two images shown for August 1998 in the middle group of Fig. 7 where the band appears greenish because of the false-color scheme. Although DS2, which was nestled in one of the northern excursions of the wave in 1989, is no longer visible (at least in HST images), the wave structure appears to have persisted from 1989 through at least 1998.
Secular Variations on Uranus and Neptune. In recent years, Uranus has revealed greater weather activity as its Northern Hemisphere continued its emergence out of decades of darkness. First seen in HST images of Uranus (31,55), bright Northern Hemisphere clouds have now become visible from the ground (56). Over a longer time period, Uranus has exhibited rather strong disk-averaged albedo variations, including a 14% increase from 1963 to 1981 (during which the sub-Earth latitude varied from near zero to 68° S) (57). Additional ground-based observations of a peak blue reflectivity in 1985 and a 7% decline from 1985 to 2001 are reported by Karkoschka (58), who also showed that there is a hemispheric asymmetry in brightness; the Northern Hemisphere is darker, and much of the recent decline in brightness has to do with geometric effects as more of the darker Northern Hemisphere comes into view. He also showed that additional physical effects played a role in earlier brightness increases and suggested that significant physical effects would also appear in the near future.
Recently, Neptune also exhibited a relatively steady brightening at visible wavelengths of 1.4-1.9% from 1996 to 1998 (30) and almost 10% from 1990 to 2000 (59); this dramatically breaks from the previous inverse correlation with solar UV variations. Neptune’s atmosphere also exhibits dynamic activity during decade-long timescales, as established by ground-based imaging of changing distributions of bright features and ground-based photometry of varying light curves. During 1976 and 1987, it appears that much brighter cloud features were present than have been seen during or since Voyager (60). Ground-based observations of bright cloud features at latitudes where none were seen by Voyager, large changes in light curve amplitudes from year to year (61), and factor of 10 changes in near-IR brightness (62) also suggest major developments or large latitudinal excursions during a several year period. Sromovsky et al. (29) showed that the spectral character of the 1977 ”outburst” could be matched by a factor of 7 increase relative to August 1996 in the fractional area covered by high-altitude, bright, cloud features.
Zonal Mean Circulation and Its Stability. Uranus and Neptune provide strong evidence that rotation dominates solar radiation and internal heat flux in determining the form of a planet’s zonal circulation. Given that Uranus has had one pole in sunlight for ~ 20 years, it would have been natural to expect strong meridional circulation between the two hemispheres. Yet, the observed temperature difference between Uranus’ long dark hemisphere and that heated by the Sun is very small and opposite in sign to this expectation. Part of the explanation was already discussed in the section on thermal structure, that is, the long radiative time constant removes most of the interhemispheric thermal contrast. The form of the circulation turned out to be zonal (parallel to lines of constant latitude).
Voyager’s inability to find many cloud features on Uranus in 1989 hampered efforts to characterize fully its mean zonal circulation. Yet, even with sparse sampling, it was clear that the circulation was retrograde near the equator (atmospheric parcels fell behind the planet’s rotation) and prograde at high latitudes. Radio occultation measurements provided one low latitude measurement (Fig. 8) that was critical in confirming this picture. Uranus’ cloud features have remained difficult to observe since that time, until recently. Improved near-IR imaging capabilities (NICMOS and Keck adaptive optics imaging) and the emergence of the Northern (IAU convention) Hemisphere into daylight have brought many new cloud features into view. New measurements, summarized by Hammel et al. (53) and shown in Fig. 8, suggest a possible small difference between hemispheres. To first order, however, the new data are consistent with the Voyager results and suggest a relatively stable and symmetrical circulation that bears a strong resemblance to Neptune’s in form but is considerably weaker in amplitude.
The most detailed definition of Neptune’s zonal mean circulation is provided by Voyager 2 imaging observations (64). A strong retrograde equatorial jet at 400 m/s (895 mph) and a weaker prograde jet at 250 m/s are the main features of the circulation (Fig. 8). The prograde jet on Uranus moves at about 200 m/s and
Figure 8. Zonal mean circulations of Uranus and Neptune. Voyager results (filled circles) and post-voyager results (open circles) are roughly consistent with an unchanging symmetrical circulation for both planets, although somewhat better fits to the Uranus observations are obtained with slightly asymmetrical profiles. Data points are from the Uranus compilation by Hammel et al. (63), and the Neptune Voyager observations of Limaye and Sromovsky (64) and HST observations by Sromovsky et al. (30). The post-Voyager results for Uranus are a combination of HST and Keck imaging from 1997-2000, the post-Voyager Neptune results are entirely from HST imaging from 1994-1998.
From the wind measurements displayed in Fig. 8, it seems that there is much less variability in the measurements of Uranus. This is in part the result of very rapid evolution of cloud features on Neptune, which reduces the typical time interval over which a cloud feature can be tracked and thus reduces the accuracy in speed that can be achieved. It is also true that part of the variability in the Neptune measurements comes from true eddy motions. The cloud features tracked on Uranus, on the other hand, seem to be more stable, allowing longer observation periods. This might be so because the cloud features are less numerous and less visible on Uranus, so that only the more intense and long-lasting features can be seen at all. It is also likely that due to far lower internal energy flux on Uranus, there may be less eddy activity to start with, as indicated by the small number of high altitude clouds and the relatively clear upper troposphere.
The equatorial retrograde winds on both planets might be a consequence of angular momentum conservation in the presence of an axisymmetric meridional circulation. Because distance to the spin axis decreases with latitude, so does the angular momentum of atmospheric parcels at rest with respect to the rotating planet. But if atmospheric mass at midlatitudes were to move equatorward, it would tend to slow its rotational speed to conserve its angular momentum, leading to a retrograde wind near the equator at the level of the equatorward flow. Midlatitude gas parcels moving poleward would tend to speed up as they get closer to the spin access, leading to a prograde circulation at high latitudes. The reverse flow at return levels would tend to produce the opposite effects. This meridional flow may be consistent with the observed zonal circulations, but there are no direct observations that could confirm or deny the existence of the meridional circulation. Such a model would not explain the equatorial prograde jets of Jupiter and Saturn that clearly require transport of angular momentum by eddies.
The solar irradiance at Uranus is about 3.8W/m2, and the planet-wide average is about 0.65 W/m2 in absorbed flux. Neptune, on the other hand, is exposed to a solar irradiance of only 1.5 W/m2 and absorbs an average of about 0.26 W/m2 compared to its internal heat flux of 0.4 W/m2 (65). Thus, the two planets actually have about the same total heat flux available for generating atmospheric motions. Their different atmospheric motions might be due to differences in the latitudinal distributions of the fluxes. Based on its modest spin axis inclination, Neptune’s equator will receive much more solar heat than its poles, even when averaged for a year, whereas the 98° inclination of Uranus results in its poles receiving 50% greater solar input than the equator. Based on calculations for Saturn, which has almost the same obliquity as Neptune, latitudes poleward of 60° should receive only half of the average solar flux received near the equator. Despite these different heat input distributions, there is little latitudinal variation in atmospheric temperatures on either planet, as noted in the discussion of thermal structure. To equilibrate the temperatures across all latitudes would require different magnitudes of horizontal heat transport and (even different directions) on Uranus and Neptune. This may account for some differences in circulation, but the point made by Ingersoll (66) is worth noting: there is not much connection between energy sources and speeds and patterns of outer planet circulations. Jupiter has 20 times the available power to drive circulations, yet it has only one-third the wind speeds. Although baroclinic eddies might provide the needed horizontal heat transport on Uranus, Friedson and Ingersoll (67) suggest that the internal heat flux on Neptune might reduce latitudinal gradients, acting as a sort of thermostatic control, efficiently providing extra heat in regions that are slightly cooler and reducing heat in regions that are slightly warmer. For this mechanism to work, the solar energy absorption must occur in regions within or close to the free convection zone that extends into the interior. But how can the high winds of Uranus and Neptune and Neptune’s highly variable weather phenomena be maintained? The answer seems to be that these atmospheres have very low dissipation and thus take very little power to keep them running, much like a well-lubricated ball bearing.
We don’t know very much about the winds below or above the level of the visible cloud features. We can use Voyager 2 measurements of horizontal temperature gradients to estimate the vertical wind shear by using the thermal wind equation discussed earlier. We find that the winds on both Uranus and Neptune decay as height increases, suggesting frictional dissipation in the stratosphere (33,66). The vertical shear is relatively low and requires about 10 scale heights (the pressure drops by a factor of 1/e for each scale height) to damp to zero velocity. To what depth below the clouds the winds continue to increase is unknown. We do know from studies of Neptune’s gravity field (discussed earlier) that its winds cannot maintain the same speed throughout the interior but must damp to low values within a small fraction of the planet’s radius.
The Voyager measurements of Neptune’s zonal circulation (Fig. 8) are at least roughly consistent with earlier ground-based and subsequent HST observations (53,54). However, it is not clear that this circulation is stable in detail, or whether there is more latitudinal variation than Voyager observations have indicated. According to the theory of LeBeau and Dowling (50), the detailed curvature of the zonal wind latitudinal profile is important because it determines the drift rate and lifetime of Great Dark Spots. Thus, measurements of both zonal wind and discrete feature latitudinal drifts put strong constraints on such theories. Recent HST observations (30) are beginning to indicate a consistent pattern of small deviations from the Voyager profile.
Satellite Systems
Uranus and Neptune, like Jupiter and Saturn, have systems of regular satellites in prograde orbits that lie near the planets’ equatorial planes. The two planets also have irregular satellites. Neptune’s moon Triton is the most significant. Because of their orbital similarities and prograde orbits, regular satellites, it is thought, formed with the planet in a common process, rather than being captured after the planet’s formation. During the early stages of giant planet formation, the local environment becomes hot enough to vaporize the constituents that later cool to form the solids, which subsequently accrete into satellites. The thermal history of each planet and the timing relative to the blowing out of nebular gas by the intense solar wind generated during the Sun’s T Tauri phase are thought to determine the characteristics of the satellite systems. Satellites Of Uranus. Uranus has 20 known satellites; five were known prior to Voyager’s 1986 encounter with the Uranian system (Fig. 9), and 11 were discovered in Voyager images (Table 3). The names of the Uranian satellites are derived from the writings of Shakespeare and Pope. The four largest regular satellites (Ariel, Umbriel, Titania, and Oberon) have low inclination and low eccentricity. Most formation models require that the satellites formed after the aRef. 68: From ssd.jpl compilation on 24 April 2002. bRef. 69.
![[JPL P30054]. Voyager 2 montage of the five largest satellites of Uranus in order of increasing distance from Uranus (Miranda, Ariel, Umbriel, Titania, Oberon) and correct relative sizes and brightness. Similar to Fig. 15. p. 52, Science 233. Imaging Team Report. This figure is available in full color at http://www.mrw.interscience.wiley.com/esst.](http://lh6.ggpht.com/_1wtadqGaaPs/TF6WWd3GtwI/AAAAAAAAPJs/IU7_8Pp0X-A/s1600-h/tmp4210_thumb3.jpg "[JPL P30054]. Voyager 2 montage of the five largest satellites of Uranus in order of increasing distance from Uranus (Miranda, Ariel, Umbriel, Titania, Oberon) and correct relative sizes and brightness. Similar to Fig. 15. p. 52, Science 233. Imaging Team Report. This figure is available in full color at http://www.mrw.interscience.wiley.com/esst.")
Figure 9. [JPL P30054]. Voyager 2 montage of the five largest satellites of Uranus in order of increasing distance from Uranus (Miranda, Ariel, Umbriel, Titania, Oberon) and correct relative sizes and brightness. Similar to Fig. 15. p. 52, Science 233. Imaging Team Report.
Table 3. Satellites of Uranus
“Discoveries are dated by data acquisition date (a = Voyager 2, b = Karkoschka, c = Kuiper, d = Lassell, e = Herschel, f = Gladman, g = Nicholson, h = Holman, i = Kayelaars).
Impact event that tilted Uranus’ spin axis and that they evolved to a state of synchronous rotation in which the satellite always keeps the same side facing the planet. Synchronous rotation has been approximately confirmed for all five of the largest satellites by Voyager 2 imaging (70). These satellites are generally denser and darker than the moons around Saturn and have a rocky fraction of 50% or more. One theory is that a giant impact generated a disk that was ice-poor because the shock energy converted methane and ammonia into largely uncon-densable CO and N2. An alternative theory is that the nebular gas in the vicinity of Uranus was more primitive and unprocessed, so that C and N were more tied up in CO and N2 and less in CH4 and NH3. The impact theory has the advantage of possibly generating amorphous solid carbon, which might help to explain the relatively dark color of the satellites. Oriel and Umbriel have a dense population of large impact craters; the population is especially dense for diameters of 50 to 100 km, similar to that observed for many of the oldest and most heavily cratered objects in the solar system. Titania and Ariel have very different crater populations; there are far fewer large craters, and numerical frequency increases in smaller crater sizes, indicative of a secondary impact population and younger surfaces. Miranda’s crater distribution looks like that of Oberon and Umbriel but has an average factor of 3 greater number of any given size (70).
Umbriel and Ariel have similar size and mass but dramatically different surface characteristics, indicating differences in evolution or composition. Umbriel (see enlarged view in Fig. 10) is by far the darker, displays a weaker spectral signature for water ice, exhibits very little albedo contrast across most of its surface, and shows no evidence of crater rays at Voyager resolution. The global albedo pattern suggests a young fresh surface, but the crater population suggests a very ancient surface. The surface of Ariel, on the other hand, is much brighter and seems to be geologically younger. Old, large population I craters have been lost, presumably by some combination of viscous relaxation, as indicated by the slumped configuration of its largest existing crater, and extrusion of material over the surface (indicated by smooth plains). Flows on Ariel (and also Titania) are more likely to be a mixture of ammonia and water, methane, or CO clathrates, due to the lower melting points of these mixtures.
![(Left) [USGS P30230]. South polar view of Miranda, produced by a mosaic of nine images obtained by Voyager 2 on 24 January 1986. Older, heavily cratered terrain appears to have low albedo contrast, whereas the younger complex terrain is marked by bright and dark bands, scarps, and ridges. (Right) [JPL P29521]. Southern Hemisphere of Umbriel imaged by Voyager 2 on 24 January 1986. This darkest of the five large moons of Uranus has an ancient, heavily cratered surface and little albedo contrast except for the strange bright ring near the top of the image, which may be a frost deposit associated with an impact crater.](http://lh4.ggpht.com/_1wtadqGaaPs/TF6WfeBEHoI/AAAAAAAAPKA/wfflLb2u27k/s1600-h/tmp4212_thumb3.jpg "(Left) [USGS P30230]. South polar view of Miranda, produced by a mosaic of nine images obtained by Voyager 2 on 24 January 1986. Older, heavily cratered terrain appears to have low albedo contrast, whereas the younger complex terrain is marked by bright and dark bands, scarps, and ridges. (Right) [JPL P29521]. Southern Hemisphere of Umbriel imaged by Voyager 2 on 24 January 1986. This darkest of the five large moons of Uranus has an ancient, heavily cratered surface and little albedo contrast except for the strange bright ring near the top of the image, which may be a frost deposit associated with an impact crater.")
Figure 10. (Left) [USGS P30230]. South polar view of Miranda, produced by a mosaic of nine images obtained by Voyager 2 on 24 January 1986. Older, heavily cratered terrain appears to have low albedo contrast, whereas the younger complex terrain is marked by bright and dark bands, scarps, and ridges. (Right) [JPL P29521]. Southern Hemisphere of Umbriel imaged by Voyager 2 on 24 January 1986. This darkest of the five large moons of Uranus has an ancient, heavily cratered surface and little albedo contrast except for the strange bright ring near the top of the image, which may be a frost deposit associated with an impact crater.
Miranda (see enlarged view in Fig. 10) is the smallest of Uranus’ large satellites, the closest to Uranus, and is thus most affected by gravitational focusing of external impactors. Its surface is composed of two very different terrain types: an old, heavily cratered terrain without much albedo contrast and a young, complex terrain that has scarps, ridges, and bright and dark bands.
The 11 inner satellites are small and very dark (Table 3). The increasing ice content (inferred from lower density) of satellites closer to Uranus might be due to the higher temperatures closer to Uranus that promote the conversion of CO and N2 to CH4 and NH3. Most of the satellites have nearly circular orbits close to the equatorial plane of Uranus, but the outer four are much more elliptical. Additional small satellites are probably present near the rings and control the narrow rings.
Small Satellites of Neptune. Neptune has eight known satellites (Table 4). Voyager was able to resolve surface features clearly on three of the four largest: Triton, Proteus, and Larissa (Fig. 11), but never got close enough to Nereid to obtain a detailed image. The six inner satellites, all discovered by Voyager in 1989, range in size from 29 to 208 km in radius. They all have low geometric albedos of 0.06-0.08 at visible wavelengths and are gray in color. These are darker than Nereid, which has an albedo of 0.155 (71). The inner satellites are in nearly circular orbits within five planetary radii, wheras Nereid is on a distant and very eccentric orbit that extends from 57 to 385 Neptune radii, suggesting aRef. 68. bRef. 3. cRef. 69.
Table 4. Satellites of Neptune
![Voyager imaging of Triton and newly discovered satellites 1989N1 and 1989N2. (Left) [JPL P34687]. This Voyager 2 image has a resolution of 10 km. The South Pole, which is sunlit throughout the current season, is at bottom left. The absence of large impact craters suggests that Triton's surface has been renewed within the last billion years. (Upper right) [JPL P34727]. Voyager 2 image of Neptune's satellite 1989N1 (Proteus) at a resolution of 2.7 km. Its average diameter is 208 km. Its albedo is only 6%, compared to Triton's 76%, and its color is gray. (Lower right) [JPL P34698]. Voyager 2 image of 1989N2 (Larissa), Neptune's fourth largest satellite (mean radius 95 km), at a resolution of 4.2 km. It also has a low albedo (about 5%) and seems to have craters 3050 km in diameter. This figure is available in full color at http://www.mrw. interscience. wiley.com/esst.](http://lh6.ggpht.com/_1wtadqGaaPs/TF6W4v2Rj_I/AAAAAAAAPKU/ydDsnYBSFmg/s1600-h/tmp4214_thumb3.jpg "Voyager imaging of Triton and newly discovered satellites 1989N1 and 1989N2. (Left) [JPL P34687]. This Voyager 2 image has a resolution of 10 km. The South Pole, which is sunlit throughout the current season, is at bottom left. The absence of large impact craters suggests that Triton's surface has been renewed within the last billion years. (Upper right) [JPL P34727]. Voyager 2 image of Neptune's satellite 1989N1 (Proteus) at a resolution of 2.7 km. Its average diameter is 208 km. Its albedo is only 6%, compared to Triton's 76%, and its color is gray. (Lower right) [JPL P34698]. Voyager 2 image of 1989N2 (Larissa), Neptune's fourth largest satellite (mean radius 95 km), at a resolution of 4.2 km. It also has a low albedo (about 5%) and seems to have craters 3050 km in diameter. This figure is available in full color at http://www.mrw. interscience. wiley.com/esst.")
Figure 11. Voyager imaging of Triton and newly discovered satellites 1989N1 and 1989N2. (Left) [JPL P34687]. This Voyager 2 image has a resolution of 10 km. The South Pole, which is sunlit throughout the current season, is at bottom left. The absence of large impact craters suggests that Triton’s surface has been renewed within the last billion years. (Upper right) [JPL P34727]. Voyager 2 image of Neptune’s satellite 1989N1 (Proteus) at a resolution of 2.7 km. Its average diameter is 208 km. Its albedo is only 6%, compared to Triton’s 76%, and its color is gray. (Lower right) [JPL P34698]. Voyager 2 image of 1989N2 (Larissa), Neptune’s fourth largest satellite (mean radius 95 km), at a resolution of 4.2 km. It also has a low albedo (about 5%) and seems to have craters 3050 km in diameter.
Discoveries are dated by data acquisition date (a = Voyager2, b = Lassel, c = Kuiper) that it may be a captured object rather than having formed in place. Nereid, though smaller than Proteus, is somewhat brighter. It was discovered by Kuiper in 1949. Nereid’s photometric properties suggest that it has a surface of dirty frost. Of these comparably sized satellites, only Proteus has been imaged well enough to permit even a crude map of surface features. Its most prominent feature is an impact basin (72) about 210 km in diameter (larger than the 208-km radius of Proteus). The capture of Triton (see next section) and its evolution would have greatly perturbed any inner satellites, inducing mutual collisions and breakup. Catastrophic disruption would also be likely from external sources. Thus, the inner satellites are likely to be reaccreted debris. It remains unexplained why Nereid is brighter than the inner satellites. One possibility is that its greater distance prevented it from acquiring a veneer of dark particles lost from the rings. Among the known satellites, only Galatea seems to have any dynamic influence on Neptune’s rings, which is to confine the ring arcs azi-muthally (73), as discussed in a later section.
Triton. Neptune’s satellite, Triton (Figs. 11 and 12), is one of the most peculiar of all satellites. It is the only satellite that has a retrograde orbit (it orbits opposite to the direction of Neptune’s rotation). It was discovered by William Lassel, an amateur astronomer, less than a month after the discovery of Neptune — no mean feat given that it is 200 times fainter than Neptune! Prior to the Voyager encounter with Neptune, there was considerable speculation about the size of Triton and whether or not it had an atmosphere. From its deflection of Voyager’s orbit and imaging of its surface, the size and mass were finally established in 1989. Its diameter of 2710 km is larger than Pluto’s 2300 km, though smaller than Titan’s 3150 km. Its surface albedo of 72% at visible wavelengths (74) was that of an icy surface, but its density of 2.05 grams/cm3 implied that rocky material also had to be a significant component. This density is comparable to that of the Pluto/Charon system. Only a very tenuous atmosphere of mainly nitrogen was observed and a surface pressure of only about 15 microbars.
![Close-up views of Triton. (Left) [JPL P34714]. Voyager 2 1989 image of the south polar terrain of Triton, showing 50 dark plume deposits or ''wind streaks'' on the icy surface. The plumes originate at dark spots a few miles in diameter, and some deposits stretch for more than 100 miles. A few active plumes were observed during the Voyager encounter. (Lower right) [JPL P34690]. Voyager 2 image of irregular dark patches on Triton's surface. (Upper right) [JPL P34722]. Voyager 2 image of Triton's cantaloupe-like terrain at a resolution of about 750 m. This terrain form of roughly circular depressions separated by rugged ridges is unique to Triton and covers large areas in its Northern Hemisphere.](http://lh4.ggpht.com/_1wtadqGaaPs/TF6W_rhXoPI/AAAAAAAAPKc/1QjsU69FJpk/s1600-h/tmp4215_thumb3.jpg "Close-up views of Triton. (Left) [JPL P34714]. Voyager 2 1989 image of the south polar terrain of Triton, showing 50 dark plume deposits or ''wind streaks'' on the icy surface. The plumes originate at dark spots a few miles in diameter, and some deposits stretch for more than 100 miles. A few active plumes were observed during the Voyager encounter. (Lower right) [JPL P34690]. Voyager 2 image of irregular dark patches on Triton's surface. (Upper right) [JPL P34722]. Voyager 2 image of Triton's cantaloupe-like terrain at a resolution of about 750 m. This terrain form of roughly circular depressions separated by rugged ridges is unique to Triton and covers large areas in its Northern Hemisphere.")
Figure 12. Close-up views of Triton. (Left) [JPL P34714]. Voyager 2 1989 image of the south polar terrain of Triton, showing 50 dark plume deposits or ”wind streaks” on the icy surface. The plumes originate at dark spots a few miles in diameter, and some deposits stretch for more than 100 miles. A few active plumes were observed during the Voyager encounter. (Lower right) [JPL P34690]. Voyager 2 image of irregular dark patches on Triton’s surface. (Upper right) [JPL P34722]. Voyager 2 image of Triton’s cantaloupe-like terrain at a resolution of about 750 m. This terrain form of roughly circular depressions separated by rugged ridges is unique to Triton and covers large areas in its Northern Hemisphere.
Voyager observations revealed slightly pinkish bright regions and slightly bluish to gray dark regions, but the composition of the surface materials could be obtained only from ground-based spectroscopic observations. In 1978 observations of Triton, Cruikshank and Silvaggio (75) identified the spectral signature of methane gas. Later observations identified spectral features indicating that frozen methane (NH4), solid molecular nitrogen (N2), carbon monoxide (CO), carbon dioxide (CO2), and water (H2O) were present on the surface. Because of their extreme volatility, the existence of N2, CO, and CH4 in solid form on Triton’s surface means that Triton must be extremely cold. In fact, Triton’s high reflectivity and great distance from the Sun make it one of the coldest places in the solar system; it has a surface temperature of only 38 K. The small amount of vapor sublimed into the atmosphere at these temperatures (mainly N2) is balanced by condensation from the atmosphere. The equilibrium point at which condensation equals sublimation occurs at a pressure of only 15 microbars (15 millionths the pressure at the surface of Earth). Slightly warmer and slightly colder regions upset this local balance and produce atmospheric pressure variations, winds, and a redistribution of surface materials. As Triton’s seasons lead to variations in the distribution of solar heating, the distribution of Triton’s surface materials is likely to change over time, along with its mean albedo, its light curve (its brightness variation as a function of rotational angle), and perhaps its color. In fact, Triton’s color in 1989 seems to have been considerably less red than it was in 1979 (51).
Voyager images of Triton’s limb revealed discrete clouds and thin hazes at altitudes up to 30 km (51). The haze particles are possibly CH4 ice and/or more complex organics created by interactions with UV radiation. Organic compounds may also be responsible for some of the slightly darker material on the surface, although rocky materials are also plausible contributors to surface features. Some of the most remarkable images of Triton were those of the active geyserlike plumes. The geysers were about 8 km tall, rising vertically from a small spot on a surface then abruptly ending in a dark cloud that extended in a narrow (about 5 km wide) plume downwind for as much as 150 km before disappearing (51). The cloud form suggests that horizontal wind speed increases abruptly at 8 km. The plumes are a plausible cause of the numerous dark streaks seen in the south polar region (Fig. 12). Venting of nitrogen gas, or perhaps methane, induced by solar heating (most active plumes are near subsolar latitudes) or localized geothermal energy, seems to entrain small dark particles that form the plume. The solar driven model involves dark material deposited under a layer of transparent frozen nitrogen that becomes heated by the Sun to a point at which the nitrogen in contact with the dark material is vaporized, builds in pressure, and eventually is released through a crack or rupture in the surface.
Triton’s seasons are extreme as a result of its inclined and precessing retrograde orbit. Neptune’s spin axis is tipped 29° relative to its orbital plane normal (compared to Earth’s 23.5° tilt relative to its orbit normal). If Triton orbited within Neptune’s equatorial plane, it would have seasonal forcing similar to Neptune’s (and similar to Earth’s as well, though much reduced in amplitude because of its great distance from the Sun). Triton’s orbit, however, is inclined 21° from Neptune’s equatorial plane, and because of the torque exerted by Neptune’s oblate mass distribution, it precesses in a period of 688 years (76). That inclination can either add to or subtract from Neptune’s. At one extreme, Triton’s rotational (and orbital) axis can be tipped by 29° + 21° = 50° away from the Neptune orbit normal, putting much of one of Triton’s hemisphere in constant darkness throughout its 5.9-day orbital period and leading to heating at latitudes constantly exposed to the Sun and cooling at latitudes that are in constant darkness. This results in a transfer of volatile materials from the heated region to the cooled region, forming a polar cap of fresh ice at one pole and eroding the cap that had previously formed at the other pole. At another extreme, Triton’s rotational axis can precess to a point at which it has only an 8° (29° — 21°) inclination to the Neptune orbit normal and thus would receive a much more uniform exposure to solar heating in Northern and Southern Hemispheres. As Neptune orbits the Sun every 164.8 years, the subsolar latitude on Triton will be further modulated between positive and negative values of the current orbital inclination angle, leading to a two-component modulation of seasons on Triton; the shortest period of modulation is the 164.8 year orbital period of Neptune, which has an amplitude envelope modulated by the 688 year period of precession of Triton’s orbital plane about Neptune. The summer solstice of 2000 will be followed by an equinox in 2041, during which Triton’s surface should undergo considerable change.
Triton’s surface is unlike that of any other satellite. Much of it appears roughened like the surface of a cantaloupe (upper right image of Fig. 12), formed by multitudes of circular dimples called cavi, which are about 25-30 km in diameter. The largest impact crater is only 27 km in diameter. From the small number of craters observed, Triton’s surface appears to be relatively young geologically. It has large plains, apparently flooded by cryovolcanic fluids. Linear ridges 12-15 km wide indicate cracking of the surface and upwelling of material within the cracks (Figs. 11 and 12). Peculiar irregular blotches that have bright aureoles (Fig. 12, lower right) are of unknown origin.
Ring Systems
Uranus has a much more extensive and massive ring system than Neptune, perhaps a consequence of the additional debris generated by the larger impact event that presumably tilted Uranus’ spin axis.
Rings of Uranus. The basic characteristics of the rings of Uranus are summarized in Table 5. Rings were first discovered by stellar occultation (dimming of a star’s brightness as rings pass between a star and an observer). By measuring the star’s brightness as a function of time, it is possible to determine ring aRef. 69. Distance is from Uranus’ center to the ring’s inner edge.
Table 5. Rings of Uranusa
In 1977, nine rings were discovered and characterized using stellar occultation (77). These are unofficially named 6, 5, 4, alpha, beta, eta, gamma, delta, and epsilon, in order of increasing distance from Uranus. Samples of the 1986 Voyager imaging of the Uranian rings are displayed in Fig. 13. The left image compares the image in reflected light on approach to Neptune as the image is in forward scattered light, taken after Voyager passed Uranus. Rings that have a significant component of small particles (of the order ofa micron) appear much brighter in forward scattering (much like dust on a car windshield). Large particles dominate the appearance in backscattered light. (See the figure caption for further details of this comparison.)
The Uranian rings are narrow and sharp-edged and have optical depths of 0.3 or more. Most are very narrow (no more than 10 km wide), inclined to the equatorial plane of Uranus, and eccentric. Exceptions include the gamma and epsilon rings, which are not inclined, and the eta ring, which is not inclined and nearly circular. The epsilon ring is the widest and brightest (the most prominent ring in Fig. 5) and is also the most eccentric. It varies in distance from Uranus by about 800 km, and varies in width from 20 km where it is closest to Uranus, to 100 km where it is furthest. Its variation in radial distance is five times that of the next most eccentric ring (ring 5). It is somewhat of a mystery why orbital speed differences across the epsilon ring do not spread the ring material radially, though it is likely that some satellite resonances are responsible. Two shepherd satellites have been identified for the epsilon ring (right image of Fig. 13). These provide forces tending to confine the ring particles radially. The alpha and beta rings also vary systematically in width, from 5 km to 12 km; extremes are offset by about 30° in orbital longitude from the closest and most distant positions (periapsis and apoapsis). The most inclined rings (6, 5, and 4) deviate from Uranus’ equatorial plane by only 24 to 46 km. A much more complex series of rings is seen in forward scattered light. In Fig. 13, there is an obvious lack of correlation between regions of high dust density (bright in forward scattering due to wavelength-sized particles) and regions of large particles (seen in backscatter and occultation observations). The structure and lack of correlation between dust features and large-particle features are reminiscent of Saturn’s D ring. Rings of Neptune. Neptune’s rings (Table 6 and Fig. 14) are the least understood. Following the discovery of rings around Uranus, stellar occultations were also used to search for rings around Neptune. After several failures, the first detection in 1984 puzzled astronomers by showing rings on only one side of the planet. Voyager imaging in 1989 revealed that though Neptune’s rings did completely encircle the planet, Neptune’s ring particles were not uniformly distributed along the rings. Instead, much of the ring mass was clumped in restricted ring arcs. The outermost (Adams) ring, though continuous, contained three main arcs of much higher particle density of the order of 10° wide in longitude. The confinement of material in the ring arcs, is thought to be the result of gravitational interactions with the moon Galatea (80). One resonance interaction confines the material radially, producing narrow rings, and a second resonance interaction produces clumping of the ring material longitudinally, although the theory predicts regular spacing of clumps. It is not clear why the material is not periodically clumped. Only two rings are prominent in images taken on the sunlit side.
![(Left) (Fig. 16-14 of Ref. 78). Voyager 2 images of the Uranian ring system first in backscattered light (upper half) one day before passing by Uranus in Janaury 1986 and second, in forward scattered light (lower half), taken after passing by Uranus and looking backward. The nine labeled rings (upper half) are those discovered by stellar occultation measurements. The forward scattering view dramatically enhances the visibility of micron-sized particles, revealing structures not otherwise visible. Also note the mismatch of the two views of the epsilon ring, a consequence of its significant eccentricity. (Right) [JPL P29466]. Voyager 2 discovered two moons (1986U7, named Cordelia, and 1986U8, named Ophelia) that are shepherd satellites. The inner moon pushes ring particles outward, and the outer moon pushes them inward. This prevents the narrow rings from spreading out.](http://lh4.ggpht.com/_1wtadqGaaPs/TF6XVdUAH_I/AAAAAAAAPKs/LStkY3SvG9s/s1600-h/tmp4217_thumb3.jpg "(Left) (Fig. 16-14 of Ref. 78). Voyager 2 images of the Uranian ring system first in backscattered light (upper half) one day before passing by Uranus in Janaury 1986 and second, in forward scattered light (lower half), taken after passing by Uranus and looking backward. The nine labeled rings (upper half) are those discovered by stellar occultation measurements. The forward scattering view dramatically enhances the visibility of micron-sized particles, revealing structures not otherwise visible. Also note the mismatch of the two views of the epsilon ring, a consequence of its significant eccentricity. (Right) [JPL P29466]. Voyager 2 discovered two moons (1986U7, named Cordelia, and 1986U8, named Ophelia) that are shepherd satellites. The inner moon pushes ring particles outward, and the outer moon pushes them inward. This prevents the narrow rings from spreading out.")
Figure 13. (Left) (Fig. 16-14 of Ref. 78). Voyager 2 images of the Uranian ring system first in backscattered light (upper half) one day before passing by Uranus in Janaury 1986 and second, in forward scattered light (lower half), taken after passing by Uranus and looking backward. The nine labeled rings (upper half) are those discovered by stellar occultation measurements. The forward scattering view dramatically enhances the visibility of micron-sized particles, revealing structures not otherwise visible. Also note the mismatch of the two views of the epsilon ring, a consequence of its significant eccentricity. (Right) [JPL P29466]. Voyager 2 discovered two moons (1986U7, named Cordelia, and 1986U8, named Ophelia) that are shepherd satellites. The inner moon pushes ring particles outward, and the outer moon pushes them inward. This prevents the narrow rings from spreading out.
The ring material is very dark and probably red (80). The Neptune ring particles are as dark as those in the rings of Uranus, and have a single scattering albedo of about 0.04. A plausible composition is ice mixed with silicates and/or some carbon-bearing material. The Adams and Le Verrier rings contain a significant fraction of dust, comparable to the fraction in Saturn’s F ring or Jupiter’s ring; both are significantly dustier than the main rings of Saturn or Uranus.
Table 6. Rings of Neptune
![(Left) [JPL P35060]. Three clumps, or ring arcs, are visible in this view of Neptune's outermost Adams ring, imaged by Voyager 2 in August 1989. (Right) [JPL P35023]. This pair of Voyager 2 clear-filter images shows the ring system at the highest sensitivity and at a phase angle of 134°. The brighter, narrow rings are the Adams and Le Verrier rings. Extending out from the Le Verrier ring is the diffuse Lassell ring. The inner medium width ring is the Galle ring. The ring arcs seen in the left-hand image were in the blacked out region between these more sensitive images.](http://lh3.ggpht.com/_1wtadqGaaPs/TF6X6OcQM-I/AAAAAAAAPK8/g-wENkMHGdw/s1600-h/tmp4219_thumb3.jpg "(Left) [JPL P35060]. Three clumps, or ring arcs, are visible in this view of Neptune's outermost Adams ring, imaged by Voyager 2 in August 1989. (Right) [JPL P35023]. This pair of Voyager 2 clear-filter images shows the ring system at the highest sensitivity and at a phase angle of 134°. The brighter, narrow rings are the Adams and Le Verrier rings. Extending out from the Le Verrier ring is the diffuse Lassell ring. The inner medium width ring is the Galle ring. The ring arcs seen in the left-hand image were in the blacked out region between these more sensitive images.")
Figure 14. (Left) [JPL P35060]. Three clumps, or ring arcs, are visible in this view of Neptune’s outermost Adams ring, imaged by Voyager 2 in August 1989. (Right) [JPL P35023]. This pair of Voyager 2 clear-filter images shows the ring system at the highest sensitivity and at a phase angle of 134°. The brighter, narrow rings are the Adams and Le Verrier rings. Extending out from the Le Verrier ring is the diffuse Lassell ring. The inner medium width ring is the Galle ring. The ring arcs seen in the left-hand image were in the blacked out region between these more sensitive images.
Satellites disrupted by collisions provide a plausible model for the source of Neptune’s ring material.
Meteorite collisions seem to be sufficient to explain the dust in the main Uranian rings (81), and in the diffuse Galle and Lassell rings (impacting some unseen parent body), but a more prolific source (perhaps in-terparticle collisions) is needed to account for the high abundance of dust in the Le Verrier and Adams rings.
Magnetic Fields and Magnetospheres
Jupiter’s magnetic field was obvious from the ground because of the synchrotron radiation that it generated, but the magnetic fields of Uranus and Neptune were substantiated only by Voyager observations during close approaches to the planets. The supersonic solar wind particles interact with planetary fields to form a bow shock, and inside the bow shock, the solar wind encounters and is deflected by the planetary field, forming a boundary called the magnetopause. For Uranus, the detached bow shock was observed as a sudden increase in magnetic field intensity upstream at 23.7 Uranus radii (or 23.7 RU), and the magnetopause boundary was seen at 18 RU (82). For Neptune, the corresponding distances (in Neptune radii) were 34.9 RN and 26.5 RN. The magnetospheres are not complete barriers to solar wind particles, however, especially in the polar regions where particles are able to impact the atmosphere and generate an aurora. The UV spectrometer of Voyager 2 found auroral emissions from both Uranus and Neptune. The large offsets and tilts of the magnetic fields (discussed later) lead to auroral zones far from the poles of the rotational axes. The solar wind particles also contribute charged particles that form radiation belts. The Uranian rings and moons are embedded deep within the magnetosphere and thus play a role as absorbers of trapped radiation belt particles, as confirmed for Uranus by depressions in electron counts at magnetic latitudes swept by Miranda, Ariel, and Umbriel (83). A similar situation exists for Neptune’s rings and satellites, except for Nereid, which is outside Neptune’s magnetosphere when it is on the sunward side of the planet.
The magnetic fields of Uranus and Neptune have unusually large inclinations relative to their rotational poles: the north dipole inclination is 58.6° for Uranus (84) and 47° for Neptune (85). The dipole moments are 48 and 25 times that of Earth (which is 7.9 x 1025 gauss cm3). For comparison, Jupiter’s moment is 20,000 times that of Earth, Saturn’s 600 times, and respective inclinations are 9.6° and less than 1°, compared to Earth’s 11°. Besides inclination relative to the spin axis, these magnetic dipoles are also unusual in having large offsets from the planet centers (Fig. 15) by 0.3 Ru for Uranus and 0.55 RN for Neptune. However, the magnetic field of Uranus is not purely that of a dipole. It also has a strong quadrupole component, which is most significant close to the planet and contributes to the large variability of the magnetic field near the cloud tops that ranges from 0.1 gauss on the 1986 sunlit hemisphere to 1 gauss at a point on the 1986 dark hemisphere (Earth’s field is about 0.3 gauss near the equator). The rotational periods of the magnetic fields are used to define the rotational periods of the interiors, which are given in Table 1.
The primary mechanism for producing magnetic fields in major planets, it is thought, is the dynamo mechanism, which appears to have three basic requirements (86): (1) planetary rotation, (2) a fluid electrically conducting region of the interior, and (3) convection within the conducting fluid. For Jupiter and Saturn, the rotation is rapid, the conducting fluid is metallic hydrogen, and primordial
heat of formation drives the interior convection. For Uranus and Neptune, the rotations are somewhat slower, and the conducting fluid appears to be hot “ices” composed of water, methane, ammonia, and compounds derived from them. Neptune has a measurable internal heat flux exceeding the solar input that can drive interior convection, but Uranus has no measurable internal heat flux. Yet, Uranus actually has the larger dipole moment!
Figure 15. Orientations and offsets of outer planet magnetic fields. The north pole of the rotational axis is at the upper end of each indicated rotational axis, and the orbital planes are horizontal. The magnetic dipole locations are indicated by rectangles; the magnetic south end is indicated by S.
