Geophysical Research Letters

Longitude dependences of Saturn's ultraviolet aurora



[1] Based on periodicities in the kilometric radio emissions, the Saturn Longitude System 4 (SLS4) was used to organize the far ultraviolet (120–150 nm) aurora observed by the Ultraviolet Imaging Spectrograph on the Cassini spacecraft. Individual Ultraviolet Imaging Spectrograph pixels were projected onto the ionosphere of Saturn, transformed into the SLS4 north and SLS4 south longitude systems, accumulated over all over auroral observations from 2007 to early 2009, and binned into 1°×1° bins of colatitude. The intensity of the northern aurora showed little variation in its SLS4 north system, but the intensity of the southern aurora exhibited an enhancement of over ~10 kR between ~140–280° SLS4 south longitude. This enhancement may represent the auroral signature of a southern ionospheric vortex proposed in MHD models of Saturn's magnetosphere to explain its periodicities. The loci of the northern intensity peaks and the 3 kR boundaries varied little over 360° of longitude, while the equatorward boundary of the southern aurora varied by ~5° in SLS4 south longitude, reaching its most equatorward location of ~23° colatitude between 100° and 180° longitude. The polygonal centroids of the aurora in both north and south were consistent with offsets of no more than ~1° in both hemispheres.

1 Introduction

[2] Saturn's main auroras are thought to be located at the boundary between open and closed field lines of its magnetosphere [Cowley et al., 2004; Bunce et al., 2008]. The generative mechanism of the aurora are field-aligned currents (FAC) flowing in the vicinity of these boundaries; the FACs have been observed in association with Saturn's aurora [e.g., Talboys et al., 2011]. Recent MHD models of Saturn's magnetosphere have posited the existence of one or two polar ionospheric vortices that cause the FAC and, when mapped upward along field lines, set up the well-known periodicities in the magnetosphere [Jia et al., 2012; Jia and Kivelson, 2012]. A vortex in the northern polar region would rotate with one speed, while a vortex in the southern hemisphere would rotate with a slightly different speed, thus generating the dual periodicities documented in many of Saturn's magnetospheric phenomena [Jia and Kivelson, 2012]. Heretofore, however, no evidence has been found for either of these vortices.

[3] Much of what is known about the morphology and dynamics of Saturn's ultraviolet aurora has come from observations made by instruments on the Hubble Space Telescope. Several campaigns between 1997 and 2009 confirmed the following characteristics of Saturn's aurora: the main aurora lies between 14° and 17° colatitude with a median width of ~2° [e.g., Badman et al., 2006], intensities in the far ultraviolet (FUV) extend from a few kR up to ~75 kR, the brightest aurora are in the dawn sector [e.g., Gérard et al., 2004, 2006], and auroral features that rotate with the planet but at speeds well below corotation [e.g., Clarke et al., 2005]. Equinoctial observations reveal hemispheric asymmetries in the auroral morphology because of the northern offset of the planetary dipole [Nichols et al., 2009]. The aurora responds to solar wind pressure pulses and also has a strong correlation with Saturn kilometric radio (SKR) emissions [e.g., Kurth et al., 2005]. Intensities of both the north and south aurora depend on the phase of the SKR emissions [Nichols et al., 2008, 2010a], and the centers of the aurora oscillate with the well-known SKR periodicity [Nichols et al., 2010b].

[4] The regularity of these SKR emissions led to the construction of several Saturn longitude systems (SLS). SLS systems are complicated because the frequency of Saturn's radio emissions varies about 1% over the course of several years [Galopeau and Lecacheux, 2000]. Based on measurements by the Cassini spacecraft, the SLS3 variable longitude system was designed to organize magnetospheric and ionospheric observations from 2004 to 2007 [Kurth et al., 2007, 2008]. The SKR was then found to have a dual periodicity with a shorter period associated with sources in the northern hemisphere and a longer period associated with sources in the southern hemisphere [Gurnett et al., 2009]; the dual periods themselves varied over many years and actually “merged” shortly after Saturn equinox in 2009 [Gurnett et al., 2010]. These complications prompted the construction of the SLS4 longitude system that varies in time and has a north and south component [Gurnett et al., 2011a]. This longitude system is freely available at the University of Iowa website ( Both the SLS3 and SLS4 systems are ordered so that when the Sun is at 100° longitude the SKR achieves a maximum, and the longitude measures westward (clockwise) in azimuth [Kurth et al., 2007; Gurnett et al., 2011a]. The SLS4 system differs somewhat from the variable phase system derived from magnetic field measurements [Andrews et al., 2010; Provan et al., 2011], although both systems have essentially the same period and variation.

[5] A recent study organized infrared aurora emissions from H3+ at 3.6 µm in rotational phase and local time [Badman et al., 2012]. Both rotational and local time dependences were detected in the infrared intensities, with the infrared intensities peaking between dawn and noon in local time. In rotational phase, the northern aurora differed from the southern by ~180°, with some indication that the southern periodicity appeared in the northern hemisphere. The rotational modulation was interpreted in terms of rotating FAC systems, one in the north and one in the south [Andrews et al., 2010; Southwood, 2011]. Presumably, these would be the same FAC systems appearing in the ionospheric vortex models.

[6] This paper utilizes the SLS4 north and south longitude systems [Gurnett et al., 2011a] to organize observations of Saturn's aurora made by the Cassini Ultraviolet Imaging Spectrograph (UVIS) from 2007 to 2009. The approach is statistical. By binning a large number of images in the SLS4 system, certain statistical features—if they are present—can be identified in the aurora. This statistical map of the aurora might then be compared to the statistical maps of magnetospheric phenomena that are themselves well-organized in SLS longitude [e.g., Carbary et al., 2008]. Similar binning has been used to show the organization of the aurora in local time [Carbary, 2012].

2 Instrument and Method of Analysis

[7] The observations described herein derive exclusively from the UVIS spectrographic imager on Cassini [Esposito et al., 2004]. UVIS consists of spectrographic imagers in two wavelength regimes: the extreme ultraviolet (56–118 nm) and the far ultraviolet (112–191 nm). Because Saturn's UV auroral emissions appear are primarily H Lyman α (122 nm), this paper uses observations only from the FUV channel. Each UVIS image consists of up to 1024 pixels along the spectral dimension and 64 pixels along the spatial dimension. The spatial resolution of one pixel is 1.0 × 1.5 mrad, which represents the low resolution slit used for auroral observations. The UVIS data and SPICE pointing information are available from the Planetary Data System.

[8] The investigation begins by collapsing each spectrographic image along the wavelength dimension. During this collapsing, wavelength filtering was performed to remove dayglow background at wavelengths longer than ~152 nm. UVIS accumulated a two-dimensional image when the spacecraft scans the slit field of view across the aurora. After a slow scan across the auroral region, the spacecraft rapidly “flies back” to begin another scan. The auroral scans occurred episodically when the spacecraft was at high enough latitude to observe the aurora, and long gaps consequently appear in the auroral observations. Each pixel from each “scanned” image was transformed from the UVIS coordinate system to its projection onto an ellipse 1100 km above the Saturn ellipsoid (Req = 60,268 km, Rpo = 54,364 km), which is thought to be the 1 bar level of the atmosphere [e.g., Gérard et al., 2009; Stallard et al., 2012]. The initial projection was performed in Saturn-Spin-Sun coordinates (SZS) in which the Z axis lies along the spin axis of Saturn, the X axis is in the direction of the Sun, and Y completes the system. These projected points were then transformed into SLS4 north and south longitudes using the convention λ = (λSUN - α) MOD 360°, where λSUN is the SLS4 west longitude of the Sun [, 2011] and α is the local time angle of the pixel measured counterclockwise from noon. This local time angle is given by α = ATAN2DEG(YSZS, XSZS) MOD 360°, where ATAN2 designates an arc tangent function with values between −180° and +180°. Pixel intensities were adjusted for slant viewing by using a cos(μ) correction, where μ is the local zenith angle of the pixel. The pixel locations were then transformed to west longitude coordinates: XSLS= θcosλ, YSLS= −θsinλ where θ is colatitude (measured from the north or south pole) and λ is the SLS4 longitude. Finally, statistical maps of the aurora were constructed by accumulating the corrected pixel intensities into colatitude bins of 1°×1° in XSLS and YSLS. Within these bins, representative intensities were derived from the medians of the intensities, while representative uncertainties derive from half the difference between the upper and lower quartiles of the intensities. Use of medians and quartiles tends to remove effects of statistical outliers caused, for instance, by transients that are not representative of the binned intensities. Similar processing has been widely applied in the construction of maps of the terrestrial aurora [e.g., Shue et al., 2001, Liou et al., 2001; Carbary, 2005].

[9] The maps of Saturn's aurora shown here involve spectrographic images of the aurora obtained between 6 April 2007 and 23 January 2009. Only complete scans of the aurora were included in the bin accumulations; “fly backs” of the UVIS slit across the aurora were specifically excluded. A total of 326 scans of the northern aurora were made and 61 scans of the southern. Over 1360 Saturn “revolutions” of 10.7 hours each occurred during the northern scans, while 440 revolutions took place during the southern scans. UVIS generally viewed the aurora from nearly overhead (μ < 45°) so that slant corrections were not severe. Finally, each UVIS scan encompassed the entire auroral oval from day to night and dawn to dusk. Therefore, UVIS observations of the aurora do not suffer from the extreme viewing angles and/or incomplete auroral coverage that marks observations made by the Hubble Space Telescope, for instance.

3 Statistical Maps in SLS4 Longitude

[10] Saturn's northern aurora was mapped to the SLS4 north longitude system and the southern aurora to the SLS4 south longitude system. Figure 1 exhibits these maps, with the northern auroral map in the top panel and the southern aurora in the bottom. Although the longitude systems are left-handed, the maps represent “true” views of an observer looking down from the north pole and rotating with the SLS4 north or south longitude system. The color scales are linear and indicate the southern aurora is brighter than the northern [e.g., Carbary, 2012]. The solid curves show the peaks in the latitudinal profiles of the aurora, while the dashed lines show the 3 kR levels. These 3 kR levels will be taken as the “boundaries” of the aurora, although virtually any other level might have been selected because no criteria have yet been established to define the “boundaries” of Saturn's aurora. In this instance, the peaks and boundaries were computed using two-dimensional interpolations in colatitude along 24 equally-spaced radials of constant longitude [e.g., Carbary, 2005; Carbary, 2012].

Figure 1.

Bin median maps of UV fluxes in the northern polar region in SLS4-north longitude (top) and the southern polar region in SLS4-south longitude (bottom). These maps represent the views of an observer rotating with the SLS4-north or SLS4-south longitude systems and looking down from the north pole. A left-handed polar longitude grid, marked at 45° increments, is overplotted. Dotted lines mark 5° circles of colatitude measured from the respective poles. Solid curves indicate the peak intensities of the aurora as found in colatitude profiles, while the dashed curves show the 3 kR boundaries. Asterisks indicate the centers of the aurora determined using polygonal centroiding of the peak intensities. Note that 100° points toward the Sun at SKR maximum.

[11] The northern auroral map represents a composite of 326 scans, while the southern map consists of 61 scans. Bearing in mind these statistics, one recognizes a general smoothness of the northern aurora compared to the southern. In contrast to the northern aurora, the southern evidences a broad crescent-shaped enhancement occupying nearly 180° in longitude. This southern enhancement spans SLS4 south longitudes from ~140° to ~280°. There is no such enhancement in the north in SLS4 north longitude. The persistence of the southern enhancement was checked by dividing the SLS4-south scans into two halves according to day number, and virtually the same pattern appeared in each half.

[12] The asterisks in each panel show the locations of the auroral centers as determined by polygonal centroiding:

display math(1)

where {xk, yk} are the Cartesian coordinates of the auroral peaks, and k = 0, 1, …24. A is the area of the polygon. One finds ANORTH = 734 ± 14 square degrees, while ASOUTH = 925 ± 11 square degrees (1 square degree ≈ 9×105 km). The xc and yc pair are converted to polar coordinates ρc and λc, where λc is the SLS4 longitude of the center. The northern aurora then has its center at ρNORTH = 0.51° ± 0.02° at λNORTH = 104° ± 0° (SLS4 north longitude), and the southern aurora is centered at ρSOUTH = 1.17 ° ± 0.01° at λSOUTH = 160° ± 1° (SLS4 south longitude). Note that the auroral centers are not in opposition (i.e., not 180° apart), and that the southern center over twice as far from the pole as the northern center. However, these center locations are consistent with offsets measured previously [Nichols et al., 2008, 2010b].

[13] The left panels of Figure 2 show the locations of the auroral peaks (solid lines) and the 3 kR boundaries (dashed lines) in SLS4-north and SLS4-south longitudes in their respective polar regions. The locations of the northern auroral peaks did not vary by more than ~1° from their mean value of 15.4° colatitude. The northern poleward boundary (lower dashed line) displayed more irregularity, up to ~2°, than the equatorward boundary (upper dashed line). The southern auroral loci showed more variation in SLS4 south longitude. The southern peak locations varied by over ~1° in colatitude from their average of 17.2°, with a maximum near 190° SLS4 south longitude, and a minimum near 340°. Like the poleward boundary in the north, the southern poleward boundary was irregular. However, the southern equatorward boundary varied smoothly between colatitudes of 18.5° (at 0° longitude) to 23.5° (150° longitude). This boundary was perhaps the best-organized of all the boundaries in this study. For a suggestive comparison, the orbit of Titan at 20 RS was mapped to the respective ionospheres (dot-dash lines) using the magnetic field model of Achilleos et al. [2010]. Titan serves as a proxy for the dayside magnetopause.

Figure 2.

Longitude dependences of auroral intensity peaks (solid lines) and 3 kR boundaries (dashed lines) in the northern region in SLS4-north longitude (top left, 1), compared to peaks and 3 kR boundaries in the southern region in SLS4-south longitude (bottom left, 2). Note that these are still (left-handed) west longitudes. The horizontal dotted lines show averages of the intensity peak locations. The uncertainties are based on differences between the upper and lower quartiles of the bin medians, not on the standard deviations of averages. The magnetically mapped orbit of Titan is shown as a dot-dashed line; this is a proxy for the magnetopause. The longitude dependences of auroral peak intensities in the northern region in SLS4-north longitude (top right, 3) are compared to peak intensities in the southern region in SLS4-south longitude (bottom right, 4). The mean intensities are indicated as dotted lines.

[14] Finally, right panels of Figure 2 display the auroral peak intensities as functions of SLS4 longitude. As indicated by Figure 1, the northern auroral intensities appeared essentially featureless with a mean value of 5.7 kR. On the other hand, southern intensities showed a pronounced enhancement of 20 kR at ~215° SLS4 south longitude. The region of enhancement extended from 140° to 280° longitude, and rose ~15 kR from a “background” level of ~5 kR, and were ~10 kR above a mean of 9.5 kR. Note that this enhanced region extends ~140° in longitude, thus occupying essentially one hemisphere.

4 Discussion

[15] This paper represents an attempt to organize Saturn's ultraviolet aurora in a longitude system defined by the radio emissions of Saturn, which are often used to arrange magnetospheric phenomena. This organization is complicated by the fact that Saturn has two longitude systems, one associated with northern SKR emissions and one associated with southern SKR emissions. The northern longitude system should best order the northern aurora, while the southern system should best order the southern aurora. Generally, the northern aurora showed little structure in either intensity or shape in the SLS4-north longitude system, while the southern aurora exhibited notable shape and intensity structure in the SLS4-south longitude system.

[16] Some evidence already exists for an “anomaly” locked to the rotating ionosphere of Saturn. Magnetic field perturbations at high latitudes tend to oscillate at the SKR periods of the respective hemispheres in which the oscillations are observed [Andrews et al., 2010; Southwood, 2011], suggesting a polar or ionospheric origin for the FAC that cause these oscillations [Andrews et al., 2012]. Langmuir probe measurements of electron densities at high latitudes indicate that plasmapause-like boundaries exist at high latitudes, and that these boundaries rotate with the periods of the respective northern and southern SKR periods [Gurnett et al., 2011b]. Fluctuations of radio periods are investigated at time scales of years to a few months, and the southern SKR rotational modulation is consistent with an intrinsically rotating phenomenon, in contrast with the early Voyager picture [Lamy, 2011].

[17] The magnetospheric effects of such an anomaly have been deduced in a series of MHD models [Jia et al., 2012; Jia and Kivelson, 2012]. In these models, the anomaly is an ionospheric vortex of circulating plasma that generates FACs threading the magnetosphere. One such vortex would span ~180° of longitude and ~10° of latitude, and would lie at high latitudes between ~10°–20° colatitude. A southern vortex would rotate at the southern SKR period and a northern vortex at the northern period. This arrangement appears to produce essentially all the periodicities evident in Saturn's magnetosphere such as those observed in the magnetic fields, charged particles, and radio emissions [e.g., Jia and Kivelson, 2012].

[18] The organization of the southern aurora in SLS4 south longitude strongly suggests the existence of such a vortex, at least in the south. The auroral signature of this vortex would be the FUV enhancement between ~140° to ~280° SLS4 south longitude and 10°–20° colatitude, in very good agreement with that predicted by the vortex model.

[19] However, the northern auroral map in SLS4 north longitude does not show evidence for a corresponding vortex in that hemisphere. This result may indicate that the vortex was not operating at the times of the UVIS observations, or that the vortex effects are simply below the threshold for instrumental detection, or even that the northern are not organized in SLS4 northern longitude. A number of investigations have suggested differences in the amplitudes of the southern and northern periodicities, and that the southern SKR period has usually been the most prominent during the Cassini mission [Gurnett et al., 2009, 2010; Lamy, 2011; Andrews et al., 2012]. On the other hand, frequency-time spectrograms of the SKR emissions during 2007–2009 do show both north and south components of the periodicities [Gurnett et al., 2010; Lamy, 2011].

5 Conclusions

[20] Saturn's FUV auroras were statistically mapped for the first time into SLS4 longitude system, the northern aurora being placed into SLS4-north longitude and the southern aurora being placed into the SLS4-south longitude. The northern aurora showed little variation in its SLS4-north system, either in intensity or shape. In contrast, the southern aurora exhibited a very broad enhancement of over ~15 kR between ~140–280° and a variation of ~5° in SLS4-south longitude, reaching its most equatorward location of ~23° colatitude between 100° and 180° longitude. The centers of the aurora are consistent with offsets of ~1° or less from the spin poles of the planet. The appearance of the southern aurora may be interpreted in terms of an ionospheric vortex coincident with the auroral enhancement, while no northern vortex apparently operated during the FUV observations of this investigation.


[21] This research was supported in part by the NASA Office of Space Science under Task Order 003 of contract NAS5-97271 between NASA Goddard Space flight Center and the Johns Hopkins University. The authors thank Don Gurnett and the RPWS team for providing the SLS4 longitudes used in this study, and Wayne Pryor and the UVIS team for helping with the analysis of the FUV scans.

[22] The Editor thanks Margaret G. Kivelson and an anonymous reviewers for their assistance in evaluating this paper.