Acknowledgement. We thank Science for their permission to use an excerpt from:
Smith, B. A., et al. 1982. A New Look at the Saturn System: The Voyager 2 Images Science 215 (4532), 504-537. (Excerpt from pp. 530-537.)
In response to these discoveries, substantial revisions were made in the Voyager 2 sequences to increase ring coverage. The Voyager 2 trajectory was more suitable for ring observations, particularly near closest approach, where the spacecraft was able to view the lit side under favorable geometries. The Voyager 2 imaging sequences included several time-lapse "movies" of the B ring spokes during approach, an extensive search for possible small embedded satellites that seemed likely candidates for clearing ring gaps (77), special efforts to image the rings near the time the spacecraft crossed the ring plane, and a series of high-resolution views of the F ring from above and below, including stereo pairs of frames. Most of these studies were successfully carried out, although seizure of the spacecraft scan platform shortly after ring plane crossing resulted in loss of most of the special F ring observations as well as all the high resolution images of the dark side of the rings.
In describing the Voyager 2 observations of the rings, we use the classical terminology for A, B, and C rings and for the Cassini division between the A and B rings. We apply the term D ring to the faint structure inside the C ring, and adopt the names F, G, and E for the three rings lying outside the classically known rings. No effort is made to name the thousands of individual ringlets that can be resolved in the high-resolution images; indeed, at scales of tens of kilometers or less it is probable that the structure is transitory. There has been interest in applying names to a few of the main gaps, and proposals to this effect are under consideration by the International Astronomical Union. The most prominent of these gaps lies about 3000 km inside the outer edge of the A ring and has sometimes been called the Encke division, although it was apparently first clearly seen in 1888 by J. E. Keeler of Lick Observatory (82). Pending a formal decision on gap nomenclature, we will simply call this the A ring gap.
General structure of the rings. The trajectory of Voyager 2 carried the spacecraft much closer to the rings than Voyager 1. In addition, the higher elevation angle of the sun produced brighter rings in the days and weeks before encounter. For 65 hours before closest approach, Voyager 2 observed the lit face of Saturn's rings at higher resolution than attained by Voyager 1. Complete radial coverage was obtained at about 10 km/lp, and selected coverage, at a resolution of a few kilometers per line pair in the wide- angle camera and about 600 m/lp in the narrow-angle camera, was obtained near closest approach. The highest resolution observations were obtained with the aid of image motion compensation. Because the scan platform did not operate for nearly 3 days after ring plane crossing, the highest resolution obtained on the unlit face was only 470 km/lp, taken then with the wide-angle camera.
Figure 32 is an enhanced color composite of the rings made from images obtained in clear, violet, and ultraviolet filters. The resolution is 165 km/lp. The "blueness" of the C ring and the Cassini division is immediately apparent when they are compared with the "redder" or less neutral-colored A and B rings, in agreement with Voyager 1 results (4). With the higher resolution and greater signal-to- noise ratio of the Voyager 2 images, subtle shadings in color are also observed in the A and B rings. The B ring color grades from nearly neutral to more reddish between its outer and inner edges; A ring color seems to vary in the opposite sense. Such subtle differences, which are greatly exaggerated in Fig. 32, may be due to variations in intrinsic particle properties, as we believe is the case for the blueness of the C ring and Cassini division, or merely to differences in scattering phase function, owing perhaps to variations in a complement of tiny particles. Further analysis will be necessary to distinguish between these possibilities.
Voyager 2 was able to extend color coverage to a resolution of about 50 km/lp with nearly complete radial coverage. Figures 33 and 34 show enhanced color composites of the A and C rings, and a line scan through part of the C ring. Of great interest is the existence in both the C ring and the Cassini division of sharply defined color differences. Specifically, two narrow, opaque, eccentric ringlets discovered in these regions by Voyager 1 are significantly different in color from their surroundings; they are more yellow than the relatively neutral C ring and Cassini division material, hence more similar to the A and B ring material. The C ring eccentric ringlet is also slightly darker than the surrounding C ring. Some of this color difference may be due to optical depth; however, other color differences exist between regions of comparable optical depth such as the outer part of the C ring, which is yellowish, and the outer central regions of the C ring, which are less red. Thus both gradual and abrupt color differences exist at a variety of scales in the rings. The maintenance of color differences on such small scales, in the face of continuous erosion and redistribution of ring material, is surely of significance.
Images of the faint G and E rings were obtained by Voyager 2 in forward- scattering geometry. In Fig. 35 the G ring is shown lying about 30,000 km beyond the F ring. This ring is quite optically thin, with optical depth for conservative scatterers of 1e-4 to 1e-5 depending on the phase function of the particles; however, outlying G ring material may have been sufficient to affect the Voyager 2 spacecraft, which passed about 2000 km outside it at ring plane crossing. A Voyager 2 image of the E ring was obtained at a high phase angle (~ 166 degrees). Because this image was obtained at a low elevation angle, enormous foreshortening allows the ring plane to be covered out to very great distances. Any radial concentrations of ring material comparable in optical depth to the E ring (1e-6 to 1e-7) would have been detected well beyond the orbit of Titan. The good agreement of the radial extent of the E ring with ground-based observations (83) gives us confidence that no rings of comparable optical depth exist farther out in the Saturn system.
In spite of the loss of high-resolution observations of the unlit side of the rings, photometric and color information was obtained in abundance at 500-km resolution. Figure 36 shows a wide-angle color composite obtained shortly after observations resumed, and a photometric scan through the image. The high brightness of the outer A ring and inner B ring results from moderate slant optical depth in these regions of bright particles. The other bright regions, in the outer Cassini division and outer C ring, are brighter than the inner Cassini division and inner C ring by an amount that may not be accounted for by the differences in optical depth alone (4) and indicate brighter particles in these regions. Moderate brightness of the outer B ring shown in the scan is due primarily to "Saturn-shine" on the rings. The large azimuthal variation in the brightness of the B ring seen in Fig. 36 results from preferential backscattering of illumination on the rings from the bright lit face of Saturn. Further analysis of this variation will be useful in constraining ring particle scattering behavior.
Little information was previously available on the structure of the B ring at high resolution, as the best Voyager 1 images had only 70 km/lp resolution on its lit face and its unlit face was too dark for meaningful observations by Voyager 1. Features seen by Voyager 1 at 100-km resolution were everywhere resolved by Voyager 2 into further fine-scale ringlets, down to the limit of resolution of 10 km/lp (Fig. 37).
Several wavelike features were observed by Voyager 2. The wavelike patterns in the A ring gap vicinity (4) were observed with improved sensitivity and resolution from the lit face; some of these features appear to exhibit azimuthal asymmetry as well. Similar wavelike features were observed in the outer A ring near closest approach. None of these patterns has yet been satisfactorily explained. The wavelike pattern observed in the outer Cassini division by Voyager 1 (84) was observed only very faintly by Voyager 2: a combination of unexpectedly extreme darkening of the region due to particle scattering properties and loss of several important unlit face observations prevented immediate detection of these features. However, further processing is under way.
Resonances. Gravitational resonances must play a role in driving the rings, as we will show from the observed 2:1 resonance with Mimas of the outer edge of ring B. The tentative list of resonances and suggested correlations for early Voyager 1 observations offered by Collins et al. (81) must be considered out of date, if only because of a more recent search by Franklin and Cook (85) for possible resonances with Saturn's figure. Five such resonances are tentatively suggested along with the possibility that the structure within Cassini's division may be affected on a coarse scale (seven such resonances). Franklin and Cook quote a suggestion by R. Smoluchowski that these harmonics could be generated by deep-seated meteorological activity associated with a helium phase change interacting with Saturn's core of metallic hydrogen. In that case the phases and amplitudes would be variable and so would any associated structure in the rings.
Asymmetric ring features. Several different kinds of asymmetric ring features were found in the Voyager 2 images. Several asymmetries can be seen in Fig. 38, which compares four images of the outer B ring and inner Cassini division taken on both ansae over a period of about 6 hours. Two images were taken on the west ansa and two on the east ansa. These images were scaled independently of each other and then registered in order to match up most of the large features in the Cassini division. This procedure assumes that the Cassini division features are not all eccentric by the same amount and locked in phase over a distance > 3000 km. The fine structure in the B ring is found to be highly variable when comparisons are made between opposite ansae or between observations of a single ansa during several hours. Features with radial widths of about 500 km remain correlated, but there is no systematic correlation of the 20-km-wide features. If the B ring fine structure actually reflects wave phenomena or diffusion instabilities (86), these phenomena are probably driven by many different sources with different forcing periods which give rise to the apparently chaotic and uncorrelated morphology.
Also evident inFig. 38 are large variations in the width of the gap just outside the outer edge of the B ring. The inner edge of this feature varies in radial position by over 140 km on time scales of several hours. This effect was also seen in the sequence of images taken for a spoke movie, which cover this region continuously for about 16 hours. Analysis of these data showed that the B ring edge is not simply eccentric, but has two maxima and two minima, as if the ellipse describing its edge were displaced so that the planet lies at the center of the ellipse instead of a focus. Furthermore, the radial minima of the ellipse follow the motion of Mimas. This behavior strongly suggests distortion of streamline orbits of ring particles by the classical Mimas 2:1 resonance. It has long been known that the outer edge of the B ring is near this resonance and probably related to it in some way. The larger than expected amplitude and the phase of the B ring radial asymmetry suggest that the Mimas 2:1 resonance is essentially coincident with the outer edge of the B ring.
An eccentric ringlet is also shown in Fig. 38 which was found by Voyager 1 in the inner Cassini division gap. This ringlet precesses under the influence of Saturn's oblateness and does not appear to be directly tied to an external resonance. The radial position of this feature varies by at least 40 km, with the ringlet being slightly wider at its greatest radius. This is the same as the behavior of the other eccentric ringlets found in the C ring at about 1.44 R_S (4) and 1.25 R_S. The last feature is close to the expected position of the Titan apsidal resonance. Ring particle orbits at this radius will precess at the same angular rate as Titan's orbital rate of 17.8 deg/day.
Kinky rings. Voyager 1 showed the F ring to be composed of at least three separate components, the outer two displaying a kinked and twisted appearance (4). To better understand these phenomena, several regions were selected in the F ring for repeated observations at a resolution high enough to resolve the separate components. The areas near the shepherding satellites (1980S26 and 1980S27) were chosen in order to look for changes in the ring after the close passage of these objects. Another region selected was the part of the F ring that was between the shepherds at the time of their conjunction, about 10 days before encounter. Two other regions were chosen at different orbital longitudes in order to cover other parts of the ring and to have a convenient place where the spacecraft could perform an image motion compensation maneuver, to allow subkilometer resolution of a small part of the F ring.
Voyager 2 found the same clumpy and occasionally kinked appearance of the F ring, but surprisingly found only one small region where the rings appeared twisted or braided. Evidently, the twisted appearance of the F ring is either a time-variable phenomenon (possibly modulated by the precession of the eccentricities of the ring and shepherds) or one that occurs over a restricted region of the ring. In total, Voyager 2 imaged about 15 percent of the F ring with a resolution high enough to see any twisted appearance. The highest resolution showed at least five separate components within the F ring. Figure 39 shows a wide-angle view of the F ring, taken shortly after ring plane crossing, that covers about 20 degrees near 1980S26. With a resolution of about 15 km/lp, one bright and four faint components can be seen. Individual concentrations within these components have also been detected (Fig. 40) and may be responsible for the multiple structure of the ring. Perturbations of large ring particles by the shepherding satellites will take much longer to damp out than those of smaller particles. Goldreich (87) has suggested that the motions of large particles within the F ring could induce the twisted and clumped morphology observed by Voyager 1.
The clumps in the F ring were observed from about 4 weeks before encounter through closest approach, and then about 3 days after encounter, when the spacecraft scan platform was repointed to Saturn. Clumps could be tracked and were recognizable over periods of at least 15 orbits. No apparent variation was found in the longitudinal distribution of clumps. They appear to be fairly uniformly distributed around the ring at intervals of 5000 to 13,000 km (typically about 9000 km). A similar spacing of about 7000 km was observed on Voyager 1, but erroneously reported as 700 km (4). This spacing is comparable to the relative motion of the F ring particles with respect to the shepherding satellites over one orbital period. During this time ring particles will experience a pulse in the gravitational pull from the satellites because of the relative eccentricities between the ring and satellites.
High-resolution images (10 km/lp) of the major A ring gap at about 2.21 R_S reveal two apparently discontinuous rings inside the division. Both rings, shown in Fig. 41, appear to vary with longitude in optical depth; they fade in and out and have not been observed together on a single Voyager 2 image. From the Voyager 2 observations it is not clear whether there are two separate rings or one eccentric ring; however, the existence of two rings was established by Voyager 1, which clearly imaged both rings at different radii from Saturn at the same orbital longitude. The dumpiness of these rings can also be observed to orbit as a pattern over times of several minutes and at the orbital period of the rings at that radius. One ring is about 50 km from the inner edge of the 325-km wide gap and appears to vary in brightness with a length scale of about 3000 km. There are other regions where this ring is not visible for distances over 20,000 km while the other ring is clearly visible. The other ring is found very near the center of the division and also varies in brightness over distances of about 3000 km. This ring was found to have a kinked morphology when observed at the highest resolution. The kinks are tens of kilometers in amplitude and about 1000 km apart. It has been shown (88) that similar kinked morphologies can be induced from perturbations by eccentric satellites orbiting very close to a ring. No objects except the clumps in the rings have been found orbiting inside the A ring gap, but because of our restricted sampling of longitudes we cannot put strong observational limits on the existence or sizes of such bodies.
Ring spokes. The radial features or spokes discovered in the B ring have proved to be among the most interesting features in the ring system (4, 81). They appear dark in backscattering and bright in forward-scattering illumination conditions. Spokes were apparent in the rings as soon as Voyager 2 began the observatory phase 82 days before encounter, when the narrow-angle camera resolution was approximately 2000 km/lp. A longer time was afforded for observing the spokes than was available on Voyager 1.
Most of the spokes are confined to the central B ring with an inner boundary at 1.72 +/- 0.01 R_S (103,900 km) and an outer boundary at approximately the outer edge of the B ring. Figure 42 shows a typical spoke pattern in the B ring. Spokes can be generally classified as (i) radial or having radial components and (ii) nonradial. It has been suggested that spokes form radially in a reference frame rotating with a period close to the corotational rate of Saturn's magnetic field and then follow the differential orbital motion of the individual ring particles (89). In this model, spokes in the process of formation will have radial components, and the leading and trailing edges of forming spokes should have different periods. Precise rotational velocities of the various portions of spokes have been measured from Voyager 2 images at the JPL Image Processing Laboratory with the same software used for atmospheric feature tracking on Jupiter and Saturn. Features are compared in two or more images with time separations of 10 to 20 minutes. In several cases where radial spoke edges were observed, the leading and trailing edges show distinctly different angular rotational rates. Furthermore, the older or more tilted edges have a Keplerian rate, while the younger or radial edges are closer to the corotational rate of the magnetosphere.
Spokes have been observed in the process of formation in the B ring (Fig. 43). In particular, a new radial spoke appeared in the middle of an existing pattern in a long sequence of pictures taken 5 minutes apart. The spoke did not appear at all in one image, was very faint in the next, and then was clearly visible in every image thereafter. The time scale for formation of the 6000-km feature appears to be at least as short as 5 minutes.
As in the Voyager 1 data, the Voyager 2 images reveal the spokes more easily on the morning ansa (just rotated out of Saturn's shadow). Some spokes can be tracked as they rotate through 360 degrees or more; however, it is not clear whether we really see the identical spoke pattern or a new one that is "reprinted" on top of the old one. If spokes are created as radial features, then mapping their orbital motion backward in time should allow a determination of the point of origin. Spokes mapped backward in this way were found not to have a common orbital longitude of creation. Places of origin also appear unrelated to the shadow of the planet on the rings, although a few spokes appear to have formed within the shadow.
After Voyager 2 passed through the ring plane, the image shown in Fig. 44 was taken of the unilluminated side of the rings, showing spokes for the first time with this viewing perspective. In this image the B ring is illuminated predominantly by sunlight scattered off Saturn's atmosphere and appears brightest in backscattered light. Bright spokes are visible in the region with Saturn-shine phase angles of about 80 degrees to 120 degrees. The fact that these features are only seen in this region of favorable Saturn illumination indicates that they are dark-side phenomena and not bright-side spokes shining through optically thin parts of the B ring. The shapes seen on the dark side indicate a morphological behavior similar to that observed on the illuminated side, with spokes tilting away from the radial direction in the sense of the Keplerian motion of ring particles. Furthermore, dark-side spokes are not tilted enough to be created on the illuminated side and then passed through the ring plane a quarter of an orbit later. The conclusions that these features may be created on the dark side of the B ring while others form in Saturn's shadow have implications for the mechanism of spoke formation, casting some doubt on the idea that charging of small ring particles by photoionization alone is responsible for levitating them out of the ring plane.
Embedded moonlets. Following the discovery by Voyager 1 of radial structure in the rings, several hypotheses emerged to explain the different types of structure. Several variants of instability or "clumping" mechanisms were suggested (86), as well as the effects of small, embedded moonlets (77, 78). The moonlet hypotheses offered the possibility of understanding both the many narrow features in optically thick regions and the broad, clear gaps in optically thin regions. Because of the importance of the existence or nonexistence of objects with radii as large as 10 to 15 km for theories of ring origin, Voyager 2 observations of one optimal ring region, the Cassini division, were carefully analyzed. The region contains two empty gaps ~ 250 and ~ 450 km wide which, if cleared by moonlets, would indicate moonlet diameters of 20 to 30 km respectively (77). The region was observed nearly continuously for about 16 hours, longer than a local orbital period, at a resolution of about 80 km/lp. The sequential images of the Cassini gap were reprojected as linear swaths and stacked vertically, with each successive swath shifted horizontally to compensate for Keplerian motion, so that any moonlet present in succeeding photos would line up vertically. The data were filtered to enhance the appearance of unresolved point sources with respect to background noise. By using simulations, it was ascertained that a point source with a signal as low as two to four times the mean noise level would be visible in such a display. Almost 340 degrees coverage of both gaps in the Cassini division was obtained, and no moonlets were detected.
This negative result is a significant constraint on the size of any moonlet embedded within either gap. The limiting size is a function of geometric albedo. Satellites 1980S1, 1980S3, and 1980S28 have albedos of ~ 0.3 at this phase angle. If the albedo of a ring moonlet were as high as 0.3, our observations would rule out moonlets as large as 6 km in diameter in the Cassini gaps; if it were as low as 0.1, only moonlets as large as 10 km in diameter would be ruled out.
Whether a moonlet smaller than the limiting sizes above could clear the Cassini gaps depends on the density of the moonlet and the ring thickness, which is characterized by particle dispersion velocity (77). For values of this velocity as large as 0.5 cm sec^-1, obtained by associating observed viscosity with particle collisions (84), the gaps would require a moonlet of density 1.3 g cm^-3 with a 20 km diameter. Even with particle dispersion velocities at their absolute lower limit, given by the condition of minimal stability to gravitational disturbances, such a moonlet would have to be at least 5 km in diameter. Therefore, a combination of unusually (but not impossibly) low-surface albedo and the lowest possible particle random velocities would be required for the Cassini gaps to be cleared by solitary moonlets. The 350 km wide A ring gap remains a possible location for several moonlets 10 to 20 km in diameter; the kinky ringlets in this gap may be a clue to the presence of moonlets there. However, neither the A ring gap nor the 250 km wide "eccentric ring" gap in the outer C ring at 1.44 R_S was observed sufficiently to establish limits on the sizes of possible moonlets there.
Last updated Feb-27-1997