HST Far-Ultraviolet Imaging of Jupiter During the Impacts of Comet Shoemaker-Levy 9 John T. Clarke, Renee Prange, Gilda E. Ballester, John Trauger, Robin Evans, Daniel Rego, Karl Stapelfeldt, Wing Ip, Jean-Claude Gerard, Heidi Hammel, Manish Ballav, Lotfi Ben Jaffel, Jean-Loup Bertaux, David Crisp, Claude Emerich, Walter Harris, Mihaly Horanyi, Steven Miller, Alex Storrs, Harold Weaver Hubble Space Telescope far-ultraviolet images of Jupiter during the Shoemaker-Levy 9 impacts show the impact regions darkening over the 2 to 3 hours after the impact, becoming darker and more extended than at longer wavelengths, which indicates that ultraviolet-absorbing gases or aerosols are more extended, more absorbing, and at higher altitudes than the absorbers of visible light. Transient auroral emissions were observed near the magnetic conjugate point of the K impact site just after that impact. The global auroral activity was fainter than average during the impacts, and a variable auroral emission feature was observed inside the southern auroral oval preceding the impacts of fragments Q1 and Q2. The -20 separate bodies from the disrupt- ed comet Shoemaker-Levy 9 impacted Ju- piter's predawn southern hemisphere (1) between 16 and 22 July 1994. Entering at a relative speed of 60 km s` and a zenith angle of 450, the fragments deposited large amounts of energy into Jupiter's atmo- sphere. Many of the impact plumes rose 2000 to 3000 km above the visible cloud tops (2). The cooling material then spread laterally and fell, with the particles and molecules diffusing through the atmosphere with vertical diffusion rates determined by the state of the atmosphere. The high ini- tial temperature of the plumes (3) probably dissociated most of the molecules, with con- J. T. Clarke, G. E. Ballester, M. Ballav, and W. Harris are at the Space Physics Research Laboratory, University of Michigan, Ann Arbor, Ml 48109, USA. R. Prang6 is at the Institut d'Astrophysique Spatiale, CNRS-Universit6 Paris XI, 91405 Orsay Cedex, and the Institut d'Astrophysique de Paris, CNRS, 75014 Paris, France. J. Trauger, R. Evans, K. Stapelfeldt, and D. Crisp are at the Jet Propul- sion Laboratory, Pasadena, CA 91109, USA. D. Rego is at the Space Physics Research Laboratory, University of Michigan, Ann Arbor, Ml 48109, USA, and the Institut d'Astrophysique Spatiale, CNRS-Universit6 Paris Xl, 91405 Orsay Cedex, France. W. lp is at the Max-Planck- Institut fur Aeronomie, D-37191 Katlenburg-Lindau, Ger- many. J.-C. Gerard is at the Institut d'Astrophysique, Universit6 de Liege, 4000 Li6ge, Belgium. H. Hammel is at Massachusetts Institute of Technology, Cambridge, MA 02139, USA. L. B. Jaffel is at the Institut d'Astrophysique de Paris, CNRS, 75014 Paris, France. J.-L. Bertaux is with the Service d'Aeronomie du CNRS, Verrieres le Buisson, 91371 France. C. Emerich is at the Institut d'Astrophysique Spatiale, CNRS-Universit6 Paris Xl, 91405 Orsay Cedex, France. M. Horanyi is with LASP, University of Colorado, Boulder, CO 80309, USA. S. Mill- er is with the Department of Physics and Astronomy, University College, London, UK. A. Storrs and H. Weaver are at the Space Telescope Science Institute, Baltimore, MD 21218, USA. 1302 sequent recombination and photochemistry occurring as the plumes rose and cooled. Many of the cometary and atmospheric constituents that were anticipated in the impact plumes are strong absorbers at far- ultraviolet (FUV) wavelengths. Preimpact models of the passage of the fragment bodies through Jupiter's magnetosphere (4) also suggested that several mechanisms could produce observable effects in Jupiter's aurora or ionosphere, including the effects of the extended dust cloud on Jupiter's magneto- sphere and electrodynamic effects associated with the rapid cross-field motion of the frag- ments. We therefore acquired FUV images of Jupiter to (i) study the response of Jupi- ter's upper atmosphere to the impacts, (ii) measure upper atmospheric winds through observed motions of impact-related absorb- ers, and (iii) search for auroral emissions associated with the comet material passing through Jupiter's magnetosphere or from the aftereffects of the impacts. Before these ob- servations, there had been no direct mea- surements of the winds in Jupiter's upper atmosphere (at pressures p < 1 to 10 mbar) (5), and Jupiter's FUV aurorae have only been imaged recently by the Hubble Space Telescope (HST) (6). A new capability for FUV imaging of Jupiter's aurora with the HST Wide Field Planetary Camera 2 (WFPC2) and the post-COSTAR (Correc- tive Optics Space Telescope Axial Replace- ment) Faint Object Camera (FOC) is pre- sented in this paper. This is necessarily a progress report, after just 3 months of anal- ysis, and more detailed reports on these im- ages will be presented at a later time. SCIENCE * VOL. 267 * 3 MARCH 1995 Observations and Data Reduction The FUV images of Jupiter were obtained with both the WFPC2 and the FOC be- tween 13 July and 9 August 1994. The FUV images were scheduled 3 days before the impacts, within 2 to 4 hours after two of the first five impacts, one and three jovian rotations after the first images, then at roughly daily intervals spaced through- out the remaining week, and finally 1 and 2 weeks after the impacts ended. The tim- ing intervals were planned to determine the time scales for auroral and impact site changes. Because of the longer integration times required for the FUV images and the expectation of short-lived plume phenom- ena, all images during impacts used visible or near-UV filters by agreement with the HST visible imaging team. The limited time for FUV imaging dictated that we concentrate on the side of Jupiter near a central meridian longitude (CML) = 1800 (System III). This side afforded the best view of the northern auroral zone, some- times showing the extensions of magnetic field lines from the impact sites to the northern hemisphere (the magnetic con- jugate points) even when the impacts oc- curred beyond the limb, and offered a time series of the development of the C, A, and E impact sites (7). We later obtained im- ages of opposite sides of Jupiter spaced half of a jovian rotation apart to record nearly the full range of impact sites on 21 July (after the R impact) and on 9 August (after all impacts). We also obtained im- ages shortly after the K impact (expected to be one of the larger fragments) and a series of four images just before the impact of fragment P2. The latter two image series were intended to search for fragment-in- duced auroral emissions. Each WFPC2 imaging set consisted of a pair of exposures with the Wide Field Camera (WFC) with the filters F160BW (1150 to 2100 A) and F160BW + F13OLP (1300 to 2100 A). This approach facilitat- ed cosmic ray identification and removal and the isolation of the shortest wave- length emissions (including H Lyman a.) as a difference of images. Comprehensive discussions of the properties of the WFPC2 and image reduction are present- ed elsewhere (8). Cosmic ray events typi- cally deposit 10 times more charge per pixel than Jupiter and are readily identi- fied by the deviation from the median of neighboring pixel values. The final test of the reality of any small-scale feature in WFPC2 images is the appearance of the feature in each of a pair of images, because the cosmic ray coincidence rate is low. No spatial filtering has been performed on WFPC2 images. Locations on Jupiter in the FUV images have been determined by on March 29, 2016 Downloaded from on March 29, 2016 Downloaded from on March 29, 2016 Downloaded from on March 29, 2016 Downloaded from on March 29, 2016 Downloaded from on March 29, 2016 Downloaded from