A. INTRODUCTION The Near Infrared Mapping Spectrometer (NIMS) on the Galileo spacecraft took unique data of Comet Shoemaker-Levy/9's impact with Jupiter. A preliminary analysis of this data is presented in this submission to the Planetary Data System (PDS). It consists of nine small tables (with detached labels) and this document. B. BACKGROUND The Galileo spacecraft, enroute to Jupiter, was situated 240 million kilo meters from Jupiter with a spacecraft-Jupiter-Sun phase angle of 51 degrees during the collision of Comet Shoemaker-Levy/9 with the planet. This geometry allowed a direct view of the impacts, which occurred on the nightside of Jupiter, not viewable from the Earth, and provided an opportunity to investigate the early temporal evolution of the impact events. Much of the radiation occurs in the infrared region, and time-resolved infrared spectral observations, obtained over a broad wavelength range, are ideal for studying these phenomena. The Galileo Near Infrared Mapping Spectrometer (NIMS) instrument [1] observed the C, F, G, and R events, simultaneously with the Photopolarimeter (PPR) and Ultraviolet Spectrometer (UVS) instruments. Only data for the G and R events were telemetered to Earth. In order to ensure successful observations of the impacts, given uncertainties in the absolute spacecraft pointing, a "checkerboard" scan pattern was used, covering Jupiter and the immediate vicinity. One dimension of scanning was provided by the NIMS mapping capability, giving a 10 mrad column of 20 pixels. Each pixel is acquired in 1/63 sec and is 0.5 mrad by 0.5 mrad in size. (Jupiter's diameter as seen from Galileo was 0.6 mrad). The spacecraft scan platform provided the second dimension, scanning back and forth by 3 mrad at 0.92 mrad/sec and a period of 10 2/3 sec. Jupiter was in the field-of-view for only a fraction of each scan, giving a net time resolution of 5 1/3 seconds. The instrument was operated in the "Fixed Map" mode in which, for each spatial pixel, 17 spectral bands are simultaneously monitored. The wavelength setting was chosen to include continuum bands, where the atmospheric gases are transparent, and bands with differing absorption strengths so as to perform vertical sounding of the fireball in the atmosphere. It also included a band for possible H3+ emissions. For short wavelengths, the intense reflected sunlight signal precludes ready identification of fireball emission while Jovian thermal emission obscures the fireball signature in the 5 micron region Between these limits, in the 1.8 to 4.4 micron region, the reflected sunlight signal is weak and little atmospheric thermal emission occurs. Consequently, we employ this region for our preliminary analysis. The corresponding wavelengths and atmospheric absorption properties are listed in Table 1. The spectral resolution for each wavelength channel is 0.025 microns. ----------------------------------------------------------------------------- Table 1. Wavelengths and Jovian Atmospheric Absorption Properties Det. No. Wavelength Wavenumber Absorber, Emitter (microns) (cm-1) ======== ========== ========== ========================== 6 1.84 5430 Continuum 7 2.12 4710 Molecular hydrogen, pressure induced 8 2.40 4160 Methane (stratosphere) 9 2.69 3720 Continuum 10 2.97 3370 Ammonia (troposphere) 11 3.25 3075 Methane (stratosphere) 12 3.53 2830 " " , H3+ 13 3.82 2620 " " 14 4.10 2440 Continuum 15 4.38 2280 Phosphine (troposphere) ----------------------------------------------------------------------------- The data provided here for the G and R events are of three types: calibration data, raw and averaged data numbers, and processed data, giving source intensities in physical units. The raw data are extracted from NIMS Experiment Data Records (EDRs), archived separately on the CD-ROM volume GO_1004. Two references of use are given below [2,3]; some of the above has been abstracted from [2]. In addition, a comprehensive paper on the G fireball is being published in Icarus [4], and an analysis of the G and R splash spectra is currently underway. C. DATA PROCESSING NOTE It should be emphasized that Jupiter, during the SL-9 impact, was only slightly more than a single NIMS pixel (.5 mrad) across. So in the EDRs, non- dark NIMS DNs are pretty much restricted to a 2x2 array of pixels in each of the 17 detectors, each time the secondary mirror scans cross Jupiter. (The spacecraft scan platform was moving back and forth, from side to side, across Jupiter's expected position.) Therefore the usual method of organizing NIMS data, as spectral image cubes generated from the time-ordered EDRs, is not applicable. Fragments G and R were observed for some time, but it was possible to return only about 20 of the 2x2 arrays, in each detector, per fragment. These data are buried in the EDRs for this period, which consist mostly of dark sky DNs. But they have been extracted from the EDRs and organized into two tables, one per fragment, of sets of 17 detector DNs -- containing at least one non-dark DN -- as functions of time. The portion of Jupiter in the field of view corresponding to each (partially dark sky) pixel had to be carefully modelled, and "effective" DNs for Jupiter computed for each detector for each crossing. These were then converted to radiance according to the best understanding of the angular sensitivity of the instrument. The results also appear in tables, which are essentially 17 by ~20 value tables of radiances, equivalent to ~20 17-value spectra in time. Additional tables contain calibration and reference and derived quantities, comprising the "science" derived from the NIMS data for the impact. The complete set of tables is probably the most useful NIMS product for this final SL9 archive. It's unlikely anyone but the NIMS team will wish to work from the original data on the EDRs, but they are available on the previously mentioned CD-ROM volume GO_1004, along with documentation on their structure and some software. D. DATA FILES This submission consists of nine (9) data tables, each accompanied by a detached PDS label. The tables are in PDS ASCII format, with extension .TAB. The corresponding labels have the save name, with extension .LBL. The tables are: CAL_DATA.TAB: A file of calibration data and related information useful for interpretation. JREF_DNS.TAB: Reference spectra for the undisturbed full disc Jupiter, in data numbers (DNs). JREF_GAM.TAB: Reference spectra for Jupiter's morning hemisphere, just prior to the G fireball event, in DNs. JREF_RAM.TAB: Reference spectra for Jupiter's morning hemisphere, just prior to the R fireball event, in DNs. G_DATA.TAB: Raw DNs versus time for pixels containing the G impact site. R_DATA.TAB: Raw DNs versus time for pixels containing the R impact site. SI_G_1.TAB: Source intensities versus time for the G event using chi squared minimization for the fireball period, which finds the fraction (eta) of reflected light to subtract to obtain the best fit of a blackbody spectrum. A regression fit of eta to detector 1 was developed from the fireball period and applied to the pre- and post-fireball periods. SI_G_2.TAB: Same as above, but the regression algorithm found in the chi squared minimization was used for all data, including the fireball period. This is to test the sensitivity of the results to the analysis procedure. SI_R_2.TAB: Same as above, but for the R event. E. DATA FILE STRUCTURES (duplicated in PDS form in the detached labels) For each type of table, a detailed description of the structure of each entry (row) is given, including the starting and ending byte positions of each field. These byte positions specify the actual fields and do not include the comma separators or end-of-entry markers. 1. CAL_DATA.TAB: Calibration and related data 1 - 3 NIMS detector (of 17). 5 - 10 Wavelength of NIMS detector, in microns. 12 - 17 Mean data number (0-1023) of detector while looking at dark sky. 19 - 23 Standard deviation of dark value, the intrinsic instrumental noise 25 - 29 Standard deviation of the mean of the dark values. 31 - 37 Sensitivity of the detector for a point source, at the Jupiter distance from Galileo during the SL-9 impact, in units of terawatts per steradian-micrometer-DN. It will be used to compute source intensity from measured DNs. 39 - 48 Solar irradiance at Jupiter's distance from the sun, in microwatts per cm**2-micron. 50 - 55 Geometric albedo of Jupiter, computed from the solar irradiance and the weighted mean of the pre-impact, pre-G-impact and pre-R-impact full disc signals. The absolute value was found by normalizing to Karkoschka's (Icarus 111, 174-192, 1994) values at the two shortest wavelengths, convolved using the NIMS slit function. The geometric albedo is used to correct for reflection of the fireball intensities from the underlying cloud deck. 2. JREF_DNS.TAB: Reference spectra for undisturbed full disc Jupiter. These are mean data numbers, not dark corrected, and are useful only in a relative sense, since pointing differences change the amount of Jupiter in the field-of-view. 1 - 3 NIMS detector (of 17). 5 - 10 Wavelength of NIMS detector, in microns. 12 - 19 Mean data number, not dark corrected, for entire disk of Jupiter, before impact of the G fragment of the SL-9 comet -- 26 pixels contribute to each detector average. 21 - 28 Mean data number, not dark corrected, for entire disk of Jupiter, before impact of the R fragment of the SL-9 comet -- 8 pixels contribute to each detector average. 30 - 37 Mean data number, not dark corrected, for entire disk of Jupiter, before impact of any fragments of the SL-9 comet -- 8 pixels contribute to each detector average. 39 - 46 Weighted mean of the previous three columns. 3. JREF_GAM.TAB [JREF_RAM.TAB]: Reference spectra (in DNs) for Jupiter's morning hemisphere, just prior to the G [R] fireball event. 1 - 3 NIMS detector (of 17). 5 - 12 Wavelength of NIMS detector, in microns. 14 - 22 Mean data number, not dark corrected, for the morning hemisphere of Jupiter, just prior to the G [R] event. This hemisphere contains the impact sites. These are relative values, since pointing differences affect the amount of Jupiter in the field-of-view. 4. G_DATA.TAB [R_DATA.TAB]: Raw DNs versus time for G [R] impact site. These data are extracted from NIMS Experiment Data Records (EDRs) for the SL-9 Impact, archived on CD-ROM volume GO_1004. 1 - 3 Crossing number: As the spacecraft scan platform scanned across Jupiter's position, the impact site (the morning hemisphere) was generally contained within one, and sometimes two, NIMS mirror scans. For each scan, two pixels are illuminated, due to overlap in the spatial response. The 'crossing number' identifies each scan: even crossing numbers correspond to night to dayside scans odd numbers the reverse. Some crossings are missing due to transmission problems or instrument cycling or limited playback time. Occasionally the same crossing will appear twice, due to overlap in the scan platform scan pattern. In some cases, the signal is clearly weaker due to decreased responsivity at the edge of the field, but is retained for relative spectral information. 5 - 14 Spacecraft Clock Realtime IMage count (RIM): The RIM count is incremented every 60 2/3 seconds. It contains 91 minor frames or MOD91 counts. 16 - 21 Spacecraft Clock MOD91 count (minor frame): The MOD91 count is incremented every 2/3 second, and assumes 91 values, 0 to 90, within each RIM. Since there are two mirror scans per minor frame, and Jupiter appears in the center of the scan, .25 or .75 has been added to the MOD91 count as appropriate. 23 - 26 NIMS mirror position: The secondary mirror scans through 20 positions, designated 0-19. During the SL-9 impact, Jupiter appears mainly in the central mirror positions 9 and 10, which are the only positions for which data is included in this table. 30 - 84 NIMS raw data numbers: Sensor data from the 17 NIMS detectors, in order 1-17. Ten-bit raw data numbers (0-1023) from the current mirror position are expanded to 16 bits. (The DNs are right-justified in 17 4-byte fields, with commas between the fields.) 5. SI_G_1.TAB: Source intensities versus time for G impact site, chi squared minimization for fireball, derived regression fit before and after. SI_G_2.TAB [SI_R_2.TAB]: Source intensities versus time for G [R] impact site, regression algorithm from chi squared minimization used for all the data. 1 - 3 Consecutive item number, beginning at 1, of NIMS mirror scan crossings of Jupiter included in this table. 5 - 8 Crossing number: see 1st field of structure 4 (G_DATA.TAB) above. 10 - 17 Spacecraft Clock Realtime IMage count (RIM): see 2nd field of structure 4 (G_DATA.TAB) above. 19 - 24 Spacecraft Clock MOD91 count (minor frame): see 3rd field of structure 4 (G_DATA.TAB) above. 26 - 33 Earth received time: the time, in seconds, at which the event would have been observed on Earth. For the G impact, it is relative to 1994 day 199, 07:33:32 UT, which corresponds to spacecraft clock RIM 2486478, MOD91 count 67.99. For the R impact, it is relative to 1994 day 202, 05:35:08 UT, which corresponds to spacecraft clock RIM 2490633, MOD91 count 67.75. 35 - 40 Jupiter reference fraction (eta): the fraction of the Jupiter reference subtracted from the dark corrected data numbers. It is the best fit of a blackbody for the continuum wavelengths (1.84, 2.67, 2.99 and 4.38 microns), minimizing chi-squared over approximately 60 seconds of fireball phenomena. A regression fit of eta to detector 1 signal levels was used to determine eta for the pre- and post-fireball period in SI_G_1.TAB. However it was used for *all* the data in SI_G_2.TAB and SI_R_2.TAB. We use the weighted mean of the full disc Jupiter reference values here; very little difference in derived fireball temperatures was found using the morning hemisphere G or R reference spectra. 42 - 44 NIMS detector (of 17): Only data from those detectors found use- ful so far (6-15) have been included in this table. 46 - 52 Wavelength of NIMS detector, in microns. 54 - 60 Lower limit of source intensity, the power radiated per unit solid angle and per micron, expressed as terawatts per steradian-micron (See next field for algorithm.) The error estimate includes the intrinsic noise of the instrument and pointing-induced variations within a spectrum. Variations in overall intensity level are not included. 62 - 68 Best estimate of source intensity, the power radiated per unit solid angle and per micron, expressed as terawatts per steradian-micron. It is computed by (1) dark-correcting and summing sensor values from the two successive mirror positions containing the signal and multiplying by the instrumental sensitivity, (2) similarly dark-correcting the weighted mean Jupiter reference spectrum (from JREF_DNS.TAB) and normalizing it to detector 1, (3) subtracting the fraction eta of the Jupiter reference from the sensed results, (4) correcting the results for cloud reflection and (5) putting them into source intensity units. 70 - 76 Upper limit of source intensity, the power radiated per unit solid angle and per micron, expressed as terawatts per steradian-micron (See previous field for algorithm.) The error estimate includes the intrinsic noise of the instrument and pointing-induced variations within a spectrum. Variations in overall intensity level are not included. F. REFERENCES [1] Carlson, et al., "Near-Infrared Mapping Spectrometer Experiment on Galileo", Space Science Reviews 60: 457-502, 1992. [2] Carlson, et al., "Galileo Infrared Observations of the Shoemaker-Levy 9 G Impact Fireball: A Preliminary Report", Geophysical Research Letters, 22, 12, 1557-1560, 1995 [3] Carlson, et al., "Some Timing and Spectral Aspects of the G and R Collision Events as Observed by the Galileo Near Infrared Mapping Spectrometer", ESO SL 9 Conference Proceedings, p. 69, Garching, 1995. [4] Carlson, et al., "Temperature, Size and Energy of the Shoemaker-Levy 9 G-impact Fireball", Icarus, 1997 (in press). G. CONTACTS Robert W. Carlson, NIMS Principal Investigator Jet Propulsion Laboratory, M/S 183-601 4800 Oak Grove Drive Pasadena, CA 91109 (818) 354-2648, fax (818) 393-4605 rcarlson@issac.jpl.nasa.gov Bob Mehlman, NIMS Data Processing UCLA Institute of Geophysics and Planetary Physics Box 951567 Los Angeles, CA 90095-1567 (310) 825-2434, fax (310) 206-3051 rmehlman@igpp.ucla.edu