LETTERdoi:10.1038/nature09928

The auroral footprint of Enceladus on SaturnWayne R. Pryor1,2*, Abigail M. Rymer3*, Donald G. Mitchell3, Thomas W. Hill4, David T. Young5, Joachim Saur6,Geraint H. Jones7,8, Sven Jacobsen6, Stan W. H. Cowley9, Barry H. Mauk3, Andrew J. Coates7, Jacques Gustin10, Denis Grodent10,Jean-Claude Gerard10, Laurent Lamy11, Jonathan D. Nichols9, Stamatios M. Krimigis3,12, Larry W. Esposito13,Michele K. Dougherty14, Alain J. Jouchoux13, A. Ian F. Stewart13, William E. McClintock13, Gregory M. Holsclaw13,Joseph M. Ajello15, Joshua E. Colwell16, Amanda R. Hendrix15, Frank J. Crary5, John T. Clarke17 & Xiaoyan Zhou15

Although there are substantial differences between the magneto-spheres of Jupiter and Saturn, it has been suggested that cryovolcanicactivity at Enceladus1–9 could lead to electrodynamic couplingbetween Enceladus and Saturn like that which links Jupiter withIo, Europa and Ganymede. Powerful field-aligned electron beamsassociated with the Io–Jupiter coupling, for example, create anauroral footprint in Jupiter’s ionosphere10,11. Auroral ultravioletemission associated with Enceladus–Saturn coupling is anticipatedto be just a few tenths of a kilorayleigh (ref. 12), about an order ofmagnitude dimmer than Io’s footprint and below the observablethreshold, consistent with its non-detection13. Here we report thedetection of magnetic-field-aligned ion and electron beams (offsetseveral moon radii downstream from Enceladus) with sufficientpower to stimulate detectable aurora, and the subsequent discoveryof Enceladus-associated aurora in a few per cent of the scans of themoon’s footprint. The footprint varies in emission magnitudemore than can plausibly be explained by changes in magneto-spheric parameters—and as such is probably indicative of variableplume activity.

There have been 12 close Cassini encounters with Enceladus in the sixyears since the spacecraft arrived at Saturn. During a fly-by on 11 August2008, the spacecraft passed within 55 km of the moon at 21:06 UT.Cassini approached Enceladus from downstream (with respect to thebackground plasma flow) while moving north–south (Supplemen-tary Fig. 1). Just before closest approach, a spacecraft roll brought twoplasma sensors into the optimum orientation for measuring alongSaturn’s (approximately dipolar) magnetic field lines. At this time,powerful ion and electron beams were observed propagating fromSaturn’s northern hemisphere (Fig. 1). Neither sensor was accessibleto particles originating from Saturn’s southern ionosphere. Beams wereobserved from 3.6 to at least 23.3 RE (radius of Enceladus RE 5 252 km)downstream (positive X in Fig. 1) from Enceladus. At 21:05 UT, ,1 minbefore closest approach, with Cassini still 3.6 RE downstream of themoon, the flow of magnetic-field-aligned ions and electrons abruptlyceased. (The final burst of low energy electron flux observed after closestapproach at ,21:07 in Fig. 1b is actually the tail of a non-field-aligneddistribution and is produced by a different process to that which pro-duces the beams.)

At approximately 20:59 and 21:02 UT, the magnetic-field-alignedelectrons flicker in energy between peaks near 10 eV and 1 keV; bi-modal electron populations are observed for about 1 min either side ofthese transitions (Fig. 1b). These changes in the characteristic energy ofthe field-aligned electron flux, while not currently well understood, are

associated with changes in the magnetic field perturbation (Fig. 1c),suggesting an actual change in the total field-aligned current density. AtJupiter, variations in auroral radio emission14 and a ‘string-of-pearls’ultraviolet aurora associated with the Io footprint15 have been inter-preted as being due to multiple reflections of a standing Alfven wavecurrent system driven by Io. It is possible that the flickering in energy ofthe beams observed downstream of Enceladus is the equatorial sig-nature of a standing wave pattern like that observed at the Io footprint.It has been suggested that the locations of beams observed near Io arecontrolled by the product of the Alfven wave travel time towards Jupiterand the plasma convection speed past the moon16. If this value, in unitsof the moon’s radius, is larger at Saturn, then the beams are expected tobe further downstream than at Io.

We estimate the wave travel time to be of the order of 150 s (using theelectron density derived from Cassini data17 and assuming a dipolefield). Assuming an average plasma velocity of ,20% of full co-rotationbetween Enceladus and the onset of the beams, we find a downstreamshift of the beams of 3.6 RE. This is consistent with the observed distancedownstream from the moon where the beams begin. However, giventhe spatial1 and (likely) temporal18,19 variability of the Enceladus vents,filamentary current structures associated with local variable mass load-ing might contribute to the variability of the observations. Assuming thefield-aligned electrons are incident on Saturn’s ionosphere, theobserved flux excites hydrogen molecules at Enceladus’ footpoint, pro-ducing ultraviolet emission between 3 6 0.2 and 12 6 3.0 kR. That isabove the measurement threshold of the Cassini UltraViolet ImagingSpectrograph (UVIS)20.

Two weeks later, on 26 August 2008, the UVIS recorded threesuccessive polar views (two of which are shown here as Fig. 2) thatshow an unambiguous auroral footprint (boxed area at top left ofFig. 2a and b). UVIS spectra of the footprint look similar to the simul-taneously measured emissions from the brighter main auroral ovalseen near 75u latitude in Fig. 2. Both compare well with an H2 elec-tron-impact laboratory spectrum and are thus consistent withemissions due to electrons precipitating on atomic and molecularhydrogen at an emission altitude ,1,100 km above the 1 bar level inthe atmosphere21. Using this altitude and a quantitative Saturn mag-netic field model22, we calculate that the northern Enceladus footprintshould occur at a latitude of 64.5uN for nominal magnetosphericconditions. (The southern footprint would occur at 61.7u S becauseSaturn’s magnetic dipole, although spin-aligned within observationaluncertainties, is displaced about 0.04 RS (Saturn radii) northward fromSaturn’s geometric centre22.) The modelled footprint latitude is not

*These authors contributed equally to this work.

1Science Department, Central Arizona College, Coolidge, Arizona 85128 USA. 2Space Environment Technologies, Pacific Palisades, California 90272, USA. 3Applied Physics Laboratory, Johns HopkinsUniversity, Laurel, Maryland 20723, USA. 4Department of Physics and Astronomy, Rice University, Houston, Texas 77005, USA. 5Space Science and Engineering Division, Southwest Research Institute, SanAntonio, Texas 78238, USA. 6Institut fur Geophysik und Meteorologie, Universitat zu Koln, Cologne, D-50923, Germany. 7Mullard Space Science Laboratory, Department of Space and Climate Physics,University College London, Holmbury St Mary, Dorking, Surrey RH5 6NT, UK. 8The Centre for Planetary Sciences at University College London/Birkbeck, London WC1E 6BT, UK. 9Department of Physics andAstronomy, University of Leicester, Leicester LE1 7RH, UK. 10Laboratoire de Physique Atmospherique et Planetaire, Departement d’Astrophysique, Geophysique et Oceanographie, Universite de Liege,Liege, B-4000, Belgium. 11Laboratoire d’Etudes Spatiales et d’Instrumentation en Astrophysique, Observatoire de Paris, Centre National de la Recherche Scientifique, Universite Pierre et Marie Curie,Universite Paris Diderot, 92195 Meudon, France. 12Academy of Athens, Soranou Efesiou 4, 115 27, Athens, Greece. 13Laboratory for Atmospheric and Space Physics, University of Colorado, Boulder,Colorado 80303, USA. 14Space and Atmospheric Physics, The Blackett Laboratory, Imperial College, London SW7 2AZ, UK. 15Jet Propulsion Laboratory, Pasadena, California 91109, USA. 16Department ofPhysics, University of Central Florida, Orlando, Florida 32816, USA. 17Astronomy Department, Boston University, Boston 02215, USA.

2 1 A P R I L 2 0 1 1 | V O L 4 7 2 | N A T U R E | 3 3 1

Macmillan Publishers Limited. All rights reserved©2011

very sensitive to auroral altitude; it would shift only 0.04u equatorwardif the assumed auroral altitude were increased to 1,200 km. It is also notvery sensitive to the size of the magnetospheric cavity, varying by only,0.1u over the whole range of sizes observed during the 6-year Cassinimission.

The location of the observed northern footprint is consistent withthe expected location. The brightness centroid of the first spot (Fig. 2a)was at latitude 64.1u6 0.4u and longitude 286.0u6 0.5u, thus about1.7u downstream of the sub-Enceladus longitude of 287.7u. Here wehave set errors equal to the pixel size. The brightness centroid of the

a

1 2 5 10 20 50 100 200 500 1,000EUV counts per pixel

b

1 2 5 10 20 50 100 200 500 1,000EUV counts per pixel

Figure 2 | Cassini images of Saturn’s northern aurora, including theEnceladus auroral footprint. a, b, Successive UVIS EUV polar-projectedimages of Saturn’s north polar region from 26 August 2008 (day of year 239);02:16–03:28 UT (a) and 03:38–04:50 UT (b). Images were formed by slowlyslewing the spacecraft and its long-slit ultraviolet spectrometer. During thisinterval, Cassini moved from sub-spacecraft latitudes of 74uN to 65uN, andfrom 8.1 to 6.0 RS (Saturn radius RS < 60,300 km) from Saturn’s centre. Each

image represents two spacecraft slews across the planet. The colour bar showsEUV emission per pixel. The white boxes are centred on 64.5uN and the sub-Enceladus longitude, cover 4u in latitude and 10u in longitude, and enclose thepredicted magnetic mapping of the satellite Enceladus to Saturn’s daysideatmosphere. Satellite footprint emission is visible in both boxes. The north poleis at the centre; the latitude circles are 5u apart, and the hashed white lineindicates the day/night terminator. The Sun is to the left.

EL

S e

lectr

on

s

log

(en

erg

y (keV

)) 1

0

–1

–2

20:52:30

55

–9

0 k

eV

INC

A

ion

flu

x

20

10

0

–10

Azim

uth

al B

-field

pert

urb

atio

n (n

T)

20:55

23.3

–2.0

41.7

21:00

13.4

–1.4

23.0

Time (UT)

X (REnc)

Y (REnc)

Z (REnc)

21:05

3.6

–0.4

4.3

21:10

–6.2

0.8

–14.3

30

20

10

0

20

10

2.0

1.0

0.0

–1.0

Up

ward

ele

ctro

n

flu

x (m

W m

–2)

Pitc

h a

ng

le (°)

1012

1011

1010

109

DE

F (s

–1 m

–2 s

r –1)

Pro

ton d

iff. flux

(cm

–2 sr –1

s–1 k

eV

–1)

Closestapproach

Data

gap

a

b

c

+Z North

+XAlong

co-rotation

+YToward

Saturn

Anti-draped ΔBDraped ΔBUpward FAE flux

21:07:3021:0521:02:3021:0020:57:3020:55

Figure 1 | Cassini particle and field observations on 11 August 2008.a, Protons (55–90 keV) observed by the Cassini Ion and Neutral Camera(INCA)26. Colours indicate proton differential flux. Contours 30u and 60u fromthe magnetic field direction are overplotted in white. b, Electrons (1 eV to22 keV) measured by the most field-aligned detectors of the CAPS ElectronSpectrometer (ELS)27. Electron differential energy flux (DEF) is indicated by thecolour bar. Field-aligned electrons are observed when the instrument measuredwithin ,20u of the magnetic field line, as indicated by the white line (pitchangle). The blacked out regions are data gaps. c, Azimuthal perturbation (DB)in the magnetic field during this interval, calculated by subtracting a modelbackground field from the total field measured by the Cassini magnetometer

(MAG)28,29. Positive DB (red) is in the direction of co-rotation; the signatureexpected from simple field line draping around an obstacle is characterized bynegative DB (blue) above the equator. Positive DB (red) indicates an anti-draped perturbation in the super-co-rotational sense. Overplotted in green isthe total upward field-aligned electron (FAE) energy flux derived by numericalintegration of the electron data in b (see Supplementary Information foradditional discussion and error analysis). Calculations of electron energy loss inH2 atmospheres indicate that 1 mW m22 of particle energy input produces,10 kR of auroral ultraviolet emission30. The observed electron energy flux istherefore expected to produce a ultraviolet emission brightness between2.8 6 0.2 and 11.9 6 3.0 kR.

RESEARCH LETTER

3 3 2 | N A T U R E | V O L 4 7 2 | 2 1 A P R I L 2 0 1 1

Macmillan Publishers Limited. All rights reserved©2011

second spot was at latitude 64.3u6 0.3u and longitude 316.8u6 0.4u,about 0.8u downstream of the sub-Enceladus longitude of 317.6u.These values are very close to the anticipated Enceladus footprintlatitude of 64.5u. This agreement confirms both the identification withEnceladus and the accuracy of the magnetic field model, a ‘ground-truth’ observation which proved vital in confirming the detailed mag-netic field configuration of Jupiter23 and is equally valuable at Saturn.The predicted southern footprint has not yet been detected (28 of the310 non-detections were of the southern ionosphere); the southernfootprint may be dimmer than its northern counterpart, as is the casefor the main aurora24.

The 504 km diameter of Enceladus maps along Saturn’s magneticfield lines to a quasi-elliptical spot ,52 3 29 km in Saturn’s atmo-sphere, for average magnetospheric conditions. This spot would bespatially unresolved by UVIS. For a steady fixed source, the UVISobservations suggest emission connected to an Enceladus interactionregion at the equator extending as far as 20 RE downstream with aradial extent between 0 and 20 RE, consistent with the location of thebeams observed at the equator.

The slewing UVIS slit passed over and recorded the Enceladus-related spot on 26 August 2008 at 3:00, at 4:20 and at 8:02 UT. Thespot dims as it moves from near dawn towards noon (8:20 to 9:12 to11:54 local time). The total combined (extreme ultraviolet (EUV) plusfar ultraviolet (FUV)) spot brightnesses in the three UVIS images were1,550 6 340 R, 1,130 6 200 R and 450 6 290 R. These should be con-sidered lower limits, assuming the spatial pixel is uniformly filled bysignal, because the true emission region size is not known (see Sup-plementary Information for more details). Thus (even when visible)the Enceladus auroral footprint varies in brightness by a factor of about3. At Jupiter, footprint emission variability is principally caused by therocking of the magnetospheric plasma sheet, as the magnetic dipolemoment is inclined with respect to the spin axis. At Saturn, there is nosubstantial rocking of the plasma sheet at the location of Enceladus,but we still see these large brightness variations. This variation could,in principle, reflect variations of plume activity18,19, of ionization rates(owing to varying background plasma conditions), or of magneto-spheric size (when the magnetosphere is compressed, auroral emis-sions are generally enhanced). The last two factors do not typicallyexhibit order-of-magnitude variations17,25. The most likely cause forthe observed large-scale variability, therefore, is time-variable cryo-volcanism from Enceladus’ south polar vents, suggesting that plumeactivity was particularly high during August 2008. Thus, systematicmonitoring of Enceladus’ ultraviolet auroral footprint might provideevidence of plume variability, which is an important open issue.

Received 13 July 2010; accepted 10 February 2011.

1. Porco, C. C. et al. Cassini observes the active south pole of Enceladus. Science 311,1393–1401 (2006).

2. Spencer, J. R. et al. Cassini encounters Enceladus: background and the discoveryof a south polar hot spot. Science 311, 1401–1405 (2006).

3. Dougherty, M. K. et al. Identification of a dynamic atmosphere at Enceladus withthe Cassini magnetometer. Science 311, 1406–1409 (2006).

4. Tokar, R. L. et al. The interaction of the atmosphere of Enceladus with Saturn’splasma. Science 311, 1409–1412 (2006).

5. Jones, G. H. et al. Enceladus’ varying imprint on the magnetosphere of Saturn.Science 311, 1412–1415 (2006).

6. Spahn, F. et al. Cassini dust measurements at Enceladus and implications for theorigin of the E ring. Science 311, 1416–1418 (2006).

7. Waite, J. H. et al. Cassini ion and neutral mass spectrometer: Enceladus plumecomposition and structure. Science 311, 1419–1422 (2006).

8. Hansen, C. J. et al. Enceladus’ water vapor plume. Science 311, 1422–1425(2006).

9. Brown, R. H. et al. Composition and physical properties of Enceladus’ surface.Science 311, 1425–1428 (2006).

10. Connerney, J. E. P., Baron, R., Satoh, T. & Owen, T. Images of excited H31 at the foot

of the Io flux tube in Jupiter’s atmosphere. Science 262, 1035–1038 (1993).11. Clarke, J. T. et al. Ultraviolet emissions from the magnetic footprints of Io,

Ganymede, and Europa on Jupiter. Nature 415, 997–1000 (2002).12. Pontius, D. H. Jr & Hill, T. W. Enceladus: a significant plasma source for Saturn’s

magnetosphere. J. Geophys. Res. 111, A09214, doi:10.1029/2006JA011674(2006).

13. Wannawichian, S., Clarke, J. T. & Pontius, D. H. Jr. Interaction evidence betweenEnceladus’ atmosphere and Saturn’s magnetosphere. J. Geophys. Res. 113,A07217, doi:10.1029/2007JA012899 (2008).

14. Gurnett, D. A. & Goertz, C. K. Multiple Alfven wave reflections excited by Io, origin ofthe Jovian decametric arcs. J. Geophys. Res. 86 (A2), 717–722 (1981).

15. Bonfond, B. et al. UV Io footprint leading spot: a key feature for understanding theUV Io footprint multiplicity? Geophys. Res. Lett. 35, L05107, doi:10.1029/2007GL032418 (2008).

16. Jacobsen, S. J. et al. Location and spatial shape of electron beams in Io’s wake. J.Geophys. Res. 115, A04205, doi:10.1029/2009JA014753 (2010).

17. Persoon, A. M., Gurnett, D. A., Kurth, W. S. & Groene, J. B. A simple scale heightmodel of the electron density in Saturn’s plasma disk. Geophys. Res. Lett. 33,L18106, doi:10.1029/2006GL027090 (2006).

18. Saur, J. et al. Evidence for temporal variability of Enceladus’ gas jets: modeling ofCassini observations. Geophys. Res. Lett. 35, L20105, doi:10.1029/2008GL035811 (2008).

19. Smith, H. T. et al. Enceladus plume variability and the neutral gas densities inSaturn’s magnetosphere. J. Geophys. Res. 115, A10252, doi:10.1029/2009JA015184 (2010).

20. Esposito, L. W. et al. the Cassini ultraviolet imaging spectrograph investigation.Space Sci. Rev. 115, 299–361 (2004).

21. Gerard, J.-C. et al. Altitude of Saturn’s aurora and its implications for thecharacteristic energy of precipitated electrons. Geophys. Res. Lett. 36, L02202,doi:10.1029/2008GL036554 (2009).

22. Burton, M. E., Dougherty, M. K. & Russell, C. T. Model of Saturn’s internal planetarymagnetic field based on Cassini observations. Planet. Space Sci. 57, 1706–1713(2009).

23. Connerney, J. E. P., Acunna, M. H., Ness, N. F. & Satoh, T. New models of Jupiter’smagnetic field constrained by the Io flux tube footprint. J. Geophys. Res. 103,11929–11939 (1998).

24. Nichols, J. D. et al. Saturn’s equinoctial auroras. Geophys. Res. Lett. 36, L24102,doi:10.1029/2009GL041491 (2009).

25. Kanani, S. J. et al. A new form of Saturn’s magnetopause using a dynamic pressurebalance model, based on in situ, multi-instrument Cassini measurements. J.Geophys. Res. 115, A06207, doi:10.1029/2009JA014262 (2010).

26. Krimigis, S. M. et al. Magnetospheric imaging instrument (MIMI) on the Cassinimission to Saturn/Titan. Space Sci. Rev. 114, 233–329 (2004).

27. Young, D. T. et al. Cassini plasma spectrometer investigation. Space Sci. Rev. 114,1–112 (2004).

28. Dougherty,M. K. et al.TheCassini magnetic field investigation. Space Sci. Rev. 114,331–383 (2004).

29. Jia, Y.-D. et al. Time varying magnetospheric environment near Enceladus as seenby the Cassini magnetometer. Geophys. Res. Lett. 37, L09203, doi:10.1029/2010GL042948 (2010).

30. Waite, J. Jr et al. Electron precipitation and related aeronomy of the Jovianthermosphere and ionosphere. J. Geophys. Res. 88, 6143–6163 (1983).

Supplementary Information is linked to the online version of the paper atwww.nature.com/nature.

Acknowledgements We acknowledge support from the NASA/ESACassini Project andNASA’s Cassini Data Analysis Program.

Author Contributions A.M.R. andW.R.P. discovered the electronbeamsand the auroralfootprint, respectively, and wrote most of the paper. D.G.M. discovered the ion beamsand contributed to the text and interpretation. T.W.H. contributed extensively to the textand interpretation. D.T.Y. is CAPS PI and contributed extensively to the text andinterpretation. J.S., G.H.J., S.J., B.H.M. and A.J.C. advised on the interpretation of the insitu data. S.W.H.C. performed the field line mapping and provided advice on the paper.J.G., D.G., J.-C.G., L.L. and J.D.N. advised on the interpretation of the UVIS data. S.M.K. isthe MIMI PI and oversaw the ion data. M.K.D. is the MAG PI and oversaw themagnetometer data. L.W.E. is the UVIS PI and oversaw the UVIS data. A.J.J. and F.J.C.designed the auroral observation campaign. A.I.F.S., W.E.M., J.M.A., J.E.C. and A.R.H.helped to process the UVIS data. J.T.C. provided advice on the HST observations. X.Z.contributed to auroral discussions related to comparisons with terrestrial auroralprocesses.

Author Information Reprints and permissions information is available atwww.nature.com/reprints. The authors declare no competing financial interests.Readers are welcome to comment on the online version of this article atwww.nature.com/nature. Correspondence and requests for materials should beaddressed to A.M.R. (abigail.rymer@jhuapl.edu).

LETTER RESEARCH

2 1 A P R I L 2 0 1 1 | V O L 4 7 2 | N A T U R E | 3 3 3

Macmillan Publishers Limited. All rights reserved©2011