An Atmosphere on Ganymede from Its Occultation of SAO 186800 on 7 June 1972 Author(s): R. W. Carlson, J. C. Bhattacharyya, B. A. Smith, T. V. Johnson, B. Hidayat, S. A. Smith, G. E. Taylor, B. O'Leary and R. T. Brinkmann Source: Science, New Series, Vol. 182, No. 4107 (Oct. 5, 1973), pp. 53-55 Published by: American Association for the Advancement of Science Stable URL: http://www.jstor.org/stable/1736235 . Accessed: 14/11/2014 12:17

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

.

American Association for the Advancement of Science is collaborating with JSTOR to digitize, preserve and extend access to Science.

http://www.jstor.org

This content downloaded from 131.215.225.130 on Fri, 14 Nov 2014 12:17:50 PM All use subject to JSTOR Terms and Conditions asteroid belt at 8 X 108 This 3. J. R. Arnold, Astrophys. J. 141, 1536, 1548 relative position of the two bodies was g/year. (1965). result is similar ,to the observational 4. E. J. Opik, Advan. Astron. Astrophys. 4, then estimated as 0.3". The adopted of 109 obtained 302 (1966). diameter for was estimate g/year by 5. G. W. Wetherill, in Physical Studies of the prediction purposes Hawkins (16) as well as that obtained Minor Planets, T. Gehrels, Ed. (NASA SP- 5550 ? 130 km (2). 267, National Aeronautics and Space Admin- Successful by Latham et al. (17) from lunar seis- istration, Washington, D.C., 1971), p. 447. observations of the oc- mic data, although the flux obtained by 6. R. E. McCrosky, Smithson. Astrophys. Obs. cultation were made by Hidayat, Carl- Spec. Rep. 252 (1967); Eos Trans. Amer. the Prairie Network (5 X 1010 g/year) Geophys. Union 53, 724 (1972). son, and Johnson at the Bosscha Ob- from all sources, including fireballs 7. D. E. Gault, Meteoritics 4, 177 (1969); J. S. servatory near Lembang, Java, and by Dohnanyi, J. Geophys. Res. 75, 3468 (1970); (18), is higher. Including uncertainties G. W. Wetherill, ibid. 72, 2429 (1967). Bhattacharyya at the Kavalur field in there may be an uncer- S. G. W. Wetherill, Science 159, 79 (1968). station of the Kodaikanal Observatory (dMe/dt), 9. I. S. Astopovich, Astron. Zh. 16, 15 (1939); tainty of an order of magnitude in our F. L. Whipple and R. F. Hughes, in Meteors, in India. B. A. Smith and S. A. Smith, T. R. Kaiser, Ed. (Pergamon, London, 1955), with a calculations; however, the resonant ex- p. 149. observing portable photometer traction mechanism we propose should 10. Y. Kozai, Astron. J. 67, 591 (1962); A. J. and 20-cm telescope from Darwin, Smith, Jr., NASA Rep. TR R-194 (1964). result in a significant meteorite flux at 11. J. G. Williams, thesis, University of Cali- Australia, were just slightly (50 km; Earth. fornia, Los Angeles (1969). see Fig. 1) south of the ,actualocculta- 12. C. J. van Houten, I. van Houten-Groeneveld, Spectrophotometric study (19) of P. Herget, T. Gehrels, Astron. Astrophys. tion zone and obtained negative results. asteroids such as (511) Davida, (814) Suppl. 2, 339 (1970). The sky was of excellent photometric 13. B. G. Marsden, Astron. J. 75, 206 (1970). Tauris, (31) Euphrosyne, (175) 14. G. W. Wetherill and J. G. Williams, J. quality with scintillation noise below Andromache, and Hecuba Geophys. Res. 73, 635 (1968). average and all functioned (108) lying 15. D. E. Gault, E. M. Shoemaker, H. J. Moore, equipment near the boundary of the Kirkwood NASA Tech. Note D-1767 (1963). properly. An attempt was also made by 16. G. S. Hawkins, Astron. J. 65, 318 (1960). Gap could permit their identification 17. G. V. Latham, M. Ewing, F. Press, G. S. D. Sinvhal at the Uttar Pradesh Ob- with known classes of meteorites. Sutton, J. Dorman, Y. Nakamura, N. Toksoz, servatory in Naini Tal, India, with D. Lammlein, F. Duennebier, in Apollo 16 Another asteroidalresonance extraction Preliminary Science Report (NASA SP-315, photoelectric equipment but the fluctu- mechanism, utilizing "secular" reson- National Aeronautics and Space Administra- ations in sky transparency were too tion, Washington, D.C., 1972), pp. 9-19. ances rather than commensurabilities, 18. R. E. McCrosky, Smithson. Astrophys, Obs. large for any events to be detected. has recently been proposed by Williams Spec. Rep. 280 (1968). The photoelectric observations of 19. C. R. Chapman, thesis, Massachusetts Institute (20) and may be of comparable im- of Technology (1972); T. B. McCord, J. B. the event from Lembang (107.6?E, Adams, T. V. Johnson, Science 168, 1445 6.8?S, 1300 m above sea were portance. (1970). level) PETER D. ZIMMERMAN* 20. J. G. Williams, Eos Trans. Amer. Geophys. made with the Bosscha twin 60-cm Union 54, 233 (1973), Paper No. G46. G. W. WETHERILL 21. We thank Dr. J. G. Williams for his con- refracting telescopes mounted in ithe Department of Planetary and Space tinuing assistance and especially for providing same tube. The sky was clear and of several computer subroutines and his un- Science, University of California, published data. We also thank J. G. Higdon excellent photometric quality. Measure- Los Angeles 90024 for writing some computer programs used in ments were obtainedwith a two-channel this work. Research supported by NASA grant NGL-05-007-005. photometer, one channel for the red References and Notes * Present address: National Accelerator Lab- [wavelength (X) ; 6000 A] and the Post Office Box 1. .E. Advan. Astron. oratory, 500, Batavia, Illinois pik, Astrophys. 2, 219 60510. other for the blue (Xk 4500 A), with (1963). 2. E. Anders, Space Sci. Rev. 3, 583 (1964). 18 June 1973 m cooled photomultipliersand pulse-count- ing electronics. Only the data from the red channel, which were of higher quality, are reportedhere. The accumu- An Atmosphereon Ganymedefrom Its Occultation lated counts (at a samplingrate of 22.25 sec- ) were displayed on a visual read- of SAO 186800 on 7 June 1972 out and recorded with a 16-mm cine camera operated at a framing rate of Abstract. On 7 June 1972 the third Jovian satellite Ganymede occulted the about 24 sec-1. Also photographedwas eighth-magnitude star SAO 186800. -Successful photoelectric observations ob- the display from an accurate quartz tained at Lembang, Java (Indonesia), and Kavalur, India, show nonabrupt im- oscillator referenced before and after mersions and emersions, indicating the presence of an atmosphere whose surface the event to radio station WWV by pressure is greater than about 10-3 millibar. By fitting the two occultation dura- using an observatorychronometer. The tions as chords to a model disk, the diameter is found to be 5270 (+ 30, - 200) absolute accuracy of the timing is esti- kilometers, the major error contribution arising from the uncertain atmospheric mated to be ? 1 second; the relative thickness below the occultation layer. The derived mean density is 2.0 (- 0.03, + accuracy during the event was + 0.1 0.2) grams per cubic centimeter. second. Dead-time corrections of ap- proximately 25 percent have been ap- A search for occultation'sof stars by predictions, photographs taken in plied to the data. Direct photographsof planets carried out by Taylor at the March 1972 at the observatoriesat the the event were also taken and have been Royal Greenwich Observatory indi- Cape of Good Hope, South Africa, published elsewhere (3). The photoelec- cated that the star SAO 186800 (mag- and Perth, Australia, were analyzed to tric tracing obtained at Lembang is nitude 8.0, type KO) would be occulted yield more accurate relative positions shown in Fig. 2. The immersion and by Ganymede (JIII) on 7 June 1972 of Ganymede and the star. The pre- emersion midpoint times are 18h47m- (1). The predicted intensity drop was dictions were accordingly refined. The 40.9+ ? 1.2s and 18h50m22.4s 1.2s in about 5 percent at visible wavelengths. predicted area of visibility included corrected universal time (U.T.C.). Since the stellar position was the southern Asia, northern Australia, and The observationsat Kavalur (78.7?E, largest single source of error in the eastern Africa. The uncertainty in the 12.6?N, 800 m above sea level) were 5 OCTOBER 1973 53

This content downloaded from 131.215.225.130 on Fri, 14 Nov 2014 12:17:50 PM All use subject to JSTOR Terms and Conditions made at the Cassegrainfocus of a 1-m N Scale to guiding errors occurred primarily in telescope with an 8-arc-second aper- - the postemersion portion of the trace E-- North Pole?, ture. The photometer signal, recorded and have been suppressed. The Lem- through a Wratten 89B filter by a bang data also show an irregularmodu- cooled RCA 7102 photomultiplier,was lation of several seconds which is due amplified by a GR 1230A electrometer ' / to scattered light from nearby Jupiter. and displayedon one of the four traces f - Kavalur, India This did not interfere with identifica- of a Tektronix 533A oscilloscope with P + tion of the event. In addition there is a a four trace plug-in module in the gradual decrease in overall intensity chopped display mode and a sweep rate during the period of observationdue to Lembang,Java of 2 cm/sec. The other three traces Ganymede's motion away from Jupiter displayed 0.1-second pips from a quartz during the course of the observations. clock and full-second pips from the We emphasize that an instantaneous same source. The scope was photo- occultation would have been identified graphed by a modifiedcine camera with Fig. 1. Apparentpath of the star SAO by an abrupt intensity change super- a precision solid-state timer unit. An 186800 behind Ganymedeas seen from imposed on the more gradual sources India and Java on 7 Accutron clockface was also photo- June 1972. of noise. Such an abrupt change was graphedat the beginning of each frame, not observed. providing time calibration. The clock The Kavalur data (Fig. 3) have been was checked against British Broadcast- more rapid than the ~ 0.05-second in- plotted on a greatly expanded hori- ing Corporation time pips for a few tegrationtime of both observations.The zontal scale for two reasons: (i) the days preceding and following the event, presence of an atmosphere would pro- total amount of reduced data from and found to be very consistent. The duce a more graduallight curve through Kavalur is much less than from Lem- sky was moderately good with a very refraction. bang (approximately 8 seconds on thin wisp of cloud present during the Reduced Lembang data are available either side of immersion and emersion) occultation,but both the immersionand from several minutes before the onset and (ii) the resolutionof individualdata emersion were detected on both the of the event until several minutes after points in the Kavalur data is more im- photographic and potentiometric strip its termination.The tracing of the data portant to support our interpretation chart records. Figure 3 shows the im- in Fig. 2 shows clearly that the occulta- than is true of the Lembang data -on mersion and emersion light curves, tion was in fact observed. The time of that scale. There are breaks in the with midpoint times of 18h49m21.8s?- midoccultation was approximately 1 Kavalur data due to lack of synchroni- 0.1s U.T.C. for immersion and 18152m- minute from that predicted, well within zation between the camera and the 41.8s ?+0.1s U.T.C. for emersion. the accuracy of the prediction. More- oscilloscope, and unfortunately, such The outstandingcharacteristic of the over, photographstaken from Lembang breaks occurred near the times of im- data obtained (Figs. 2 and 3) is that the confirm that the events occurred at the mersion and emersion. The loss of data falloff and subsequent rise in intensity times indicated by the photoelectric during emersion is the more unfortu- appearto be gradualrather than abrupt. records. Immersion appears quite grad- nate because of some uncertainty over By contrast, the occultation of Beta ual, lasting perhaps several seconds. the validity of the data point designated Scorpii C by lo was observed to be Emersion, while less clear, is also non- with a question mark in Fig. 3. The instantaneousto within the time resolu- abrupt; this interpretation is made a immersion curve provides additional tion of the instruments(4). In the ab- little more difficultby the presence of a evidence that the falloff is not instan- sence of an atmosphere on Ganymede, noise spike near emersion (Fig. 2). A taneous and suggests a falloff time of the intensity change should have been few additional fluctuations attributable approximately 0.5 second, although it

27,0001

o CY 0C04 " 25,000

ua

3 24,000

10 sec I t- 300.0 0.0 Time (sec) Time (sec)-- Fig, 2 (left). Photoelectriclight curve of the occultationof SAO 186800by Ganymedefrom Lembang,Java. Note that the time scale is compressed relative to the Kavalur data (Fig. 3) and that each 10-second interval contains 220 data points. Fig. 3 (right). Photoelectriclight curve of the occultationof SAO 186800 by Ganymedefrom Kavalur,India. 54 SCIENCE,VOL. 182

This content downloaded from 131.215.225.130 on Fri, 14 Nov 2014 12:17:50 PM All use subject to JSTOR Terms and Conditions is consistent also with considerably ephemerides of Jupiter and Ganymede. mede should receive first priority for longer falloff times. The emersion curve The equations of condition contained radio occultation or other atmospheric is suggestive of an intensity rise taking three unknowns: corrections to the experiments on the Pioneers and other several seconds, but this may be due in right ascension,declination, and adopted Jupiter spacecraft. part to fluctuationsof Jovian scattered semidiameter of Ganymede. Assuming R. W. CARLSON light (Jupiter's limb was only 20 arc a circular cross section, the diameter of Department of Physics, seconds away) resulting from telescope Ganymede was found to be 5271 km. University of Southern California, guiding oscillations which occurred The formal standarderror of this value Los Angeles 90007 with time scales of the order of several was only 1 km, but this is of no signifi- J. C. BHATTACHARYYA seconds. cance: whereas the standard error of Indian Institute of Astrophysics, The data of Figs. 2 and 3 are not as the times at Lembang is 1.2 seconds, Kodaikanal, India clean as we might like. They are compli- the largest time residual is less than 0.1 B. A. SMITH cated by scattered light fluctuations,oc- second. In view of the standarderror of Department of Astronomy, casional noise spikes, and some data the observed times at Lembang it is New Mexico State University, "dropouts." It must be remembered more realistic to consider that the University Park 88001 that the intensity drop was only of the standard error of the diameter is of T. V. JOHNSON order of 5 percent, that the star was the order of 20 to 30 km. Jet Propulsion Laboratory, itself only eighth magnitude, and that On the other hand, the presence of Pasadena, California 91103 there was some scattered light from an atmosphereon Ganymede makes un- B. HIDAYAT Jupiter. Nevertheless, we feel that the certain the lower limit to the diameter. Bosscha Observatory, data are of sufficient quality to deter- If the surface pressurewere as high as Lembang, Java, Indonesia mine the occultation radius and to sup- about 1 mbar the occultation layer at S. A. SMITH* port the inference that the intensity a pressure of about 10-3 mbar would California Institute of Technology, changes are nonabrupt.Thus it appears be located seven scale heights above Pasadena 91109 that Ganymede does possess at least a the solid surface. Assuming a mean G. E. TAYLOR modest atmosphere. molceular weight of 28 (molecular.ni- Royal Greenwich Observatory, If we fit the two sets of observations trogen) and a temperature of 100?K, Hailsham, Sussex, England as chords to a model disk (Fig. 1) we the scale height would be about 20 km B. O'LEARYt find a discrepancy of about 5 seconds and the solid surface would lie about San Francisco State College, between the absolute times of the two 140 km below the occultation layer. San Francisco, California 94132 observatories. This discrepancy is too The figure would be even greater for R. T. BRINKMANN large to be explained entirely by the constituents of lower molecular weight Lunar Science Institute, gradual nature of the events, yet too and for a warm atmosphere. We con- Houston, Texas 77058 small to erode our confidence that the clude that the diameter of Ganymede References and Notes occultation was in fact observed at is 5270 (+ 30, - - 200) km, and the both locations. There is the mean 2.03 ~ 1. G. E. Taylor, Int. Astron. Union Circ. No. possibility density (- 0.03, + 0.2) 2401 (1972). of an error in the setting of the clock g/cm3. 2. A. Dollfus, Surfaces and Interiors of Planets at Because of its slow and Satellites (Academic Press, New York, one observatory or the other. Also (synchronous) 1970). difficult to interpret completely is the rotation rate (the period of rotation is 3. Sky Telescope 45, 125 (1973). 4. G. E. Taylor et al., Nature 234, 406 (1971); suggestionthat the immersionand emer- 7.155 days) the equatorial flattening of P. Bartholdi and F. Owen, Astron. J. 77, 60 sion recordedfrom Lembangwere more Ganymede should be very small. If we (1972); F. W. Fallon and E. J. Devinney, Icarus 17, 216 (1972); B. A. Smith and S. A. gradual than those from Kavalur. It is make the reasonable assumption that Smith, ibid., p. 218. clear that deriving a meaningful scale Ganymede is a homogeneousfluid body 5. 0. Hansen, thesis, California Institute of Technology (1972). height and composition of Ganymede's in hydrostatic equilibrium, the equa- 6. B. O'Leary and T. C. Van Flandern, Icarus atmospherefrom the shape of the light torial radius would exceed the polar 17, 209 (1972). 7. B. O'Leary, Science 175, 1108 (1972). curves is impossible. radius by less than 1 km and the 8. We acknowledge with thanks the special ef- it is to set a lower tidal directed forts of W. Brunk and others at the National However, possible permanent bulge toward Aeronautics and Space Administration in limit to Ganymede's surface pressure Jupiter would be only - 2 km greater granting funds for the American expeditions on short notice (under NASA grant NGR and to determine the radius at that than in an orthogonal direction (6). 05-036-005 with B.O. as principal investi- level. The data suggest a surface pres- Therefore, the assumption that Gany- gator and coordinator of results). T.V.J. is a National Research Council resident research sure greater than~ -10-3 mbar (at- mede is spherical is well within the ac- associate. We also thank B. J. Harris of Perth tributableto the lack of any noticeable curacy of these observations. Observatory, Australia, and J. Churms and T. W. Russo of Capetown Observatory, South abrupt event in the photoelectric rec- The recognition that observable oc- Africa, for providing the photographic plates ord). Infrared observations an cultations of stars by the Galilean satel- used in refining the predictions; D. T. Vu for suggest supplying photoelectric data; U.S. Informa- upper limit of less than ~ 1 mbar (5). lites are relatively frequent and the im- tion Service personnel at the U.S. Embassy in Further of the nar- of the Jakarta and G. Edwards of the Australian analysis data may provement mechanismsof predic- News and Information Bureau for their assist- row these limits. Spacecraftradio occul- tion could well result in our having ance in the American expeditions; and C. tations Durrington, Qantas Airlines, and British Over- may shed additional light in the accurate information about satellite seas Airline Company for their special help near future on the nature of Gany- atmospheres,diameters, and mean den- in getting the necessary equipment to Darwin, Australia. mede's atmosphere. sities by the end of the decade (7). * Present address: New Mexico State University, An analysis was made by using the The discovery of an atmosphere on Las Cruces 88001. t Present address: Hampshire College, Amherst, four observed times (which were all Ganymede together with previous nega- Massachusetts 01002. given equal weight) in conjunctionwith tive results for Io (4) suggest that Gany- 30 April 1973; revised 12 July 1973 5 OCTOBER1973 55

This content downloaded from 131.215.225.130 on Fri, 14 Nov 2014 12:17:50 PM All use subject to JSTOR Terms and Conditions