The impacts of the SL9 fragments on Jupiter were observed by a variety of observational methods for the world-wide network of observatories. Therefore we must refer enormous data of these observations to reveal the overview of the impact phenomena. In this section, results of other observatories for each impact are summarized.
5.1 Impact of Fragment A
The impact time of the fragment A was predicted to be at 19h59m UT, July 16, and was observable for Europe and Africa at night, and South America at daytime.
The 3.5-m telescope of Calar Alto, Spain, with MAGIC infrared camera observed the A impact with the time resolution of 2-10 seconds by near-infrared of 2.3,J-Lm [50]. They detected first signal of the A impact at 20hllm29s ±5s, which may be the first precursor of the A impact. The PCl flux was reported to be ... lQ-4 of the peak brightness of the A event and to be 1.6 mJ y (9 x 1Q-16Wm-2J-Lm-1). But a software glitch shortly thereafter suspended the observation. On the other hand, at South Africa Astronomical Observatory (SAAO) Sekiguchi and his colleagues observed the impact by K-band using 75-cm telescope with PANIC (PtSi Astronomical Near Infrared Camera) [51].
This observation could not detect the PCl of the A impact due to the long integration time of 30 seconds, but detected subsequent second precursor from 20h13m51s ±15s. The PC2 reached its maximum flux of 1.2x10-13Wm-2J-Lm-1 (0.19Jy, the K-band corresponds to 2.2J-Lm) at 20h14m51s and then decreased.
By visible wavelength, HST detected the A impact first at 20h13m16-30s [53], which recognized to be the PC2 by comparing with the SAAO observation.
The main event of the A impact started from 20h16m56s ±5s at Calar Alto, also simultaneously at SAAO. At the maximum of the main event, the detector of Calar Alto was saturated, so the peak flux is larger than 16 J y (9.1 x 10-12 Wm-2J-Lm-1). By SAAO observation, the peak flux of the main event was 9.8 x 1Q-12Wm-2J-Lm-1 (15.8Jy) at 20h24m52s ±15s, and the end of the main event was at 29m22s ±15s, and the end of the plateau was at 32m53s
±15s. At mid-infrared, the 2.56-m Nordic Optical Telescope (NOT) with CAMI
RAS camera at La Palma (Canary Island, Spain) observed by 12 J-Lm, but could detect only the main event [52). The peak of main event was about 1200 J y (2.5 x 1Q-11Wm-2J-Lm-1) at 20h23-24m, which was sightly earlier than the peak of near-infrared. HST caught the plume reflecting sunlight [53). Interest
ingly, nothing was seen in a image taking at 15m16s±2s, just before the main event, by HST. But HST detected the plume for the A impact from 18ml6-32s, and observed the plume spreading at 21m16s, the plume falling at 24m16-18s, and the settled disk of the plume at 27ml6-20s. As rotating of the impact site into the view of Earth, a peculiar structure of impact debris appeared: a large
crescent cloud with a central cloud core and a expanding wave.
5. 2 Impact of Fragment E
The impact time of the fragment E was predicted to be at 15h05m UT, July 17, and was observable for South-East and South Asia at night, but there are no large telescope in this area. The E impact was less observed in despite of its medium size.
The initial phase of the E impact is unknown. Calar Alto could not de
tect any precursors, perhaps because poor seeing, daytime observation of the E impact, and therefore a reduced aperture to 50-cm precluded sensitive measure
ments [50]. SAAO observed only after 15h22m with time resolution of 30 sec.
Before that time, they used another observation mode, which has worse time resolution (51]. For the E impact, since Galileo did not observed, eventually we do not have any direct informations on the precursors and the impact time.
The main event of the E impact started at 15h17m56s ±5s for the Calar Alto observation. The peak flux was about lOOJy (5.7 x 10-11Wm-2f1m-1) at 15h21m, but the Calar Alto lightcurve around the peak of the E impact is very irregular and it is not reliable. For SAAO observation, the peak of main event was 3.4 x 10-11Wm-211m-1 (54Jy) at 15h23m ±lm, and the end of main event was at 15h28m30s ±30s. For the NOT mid-infrared observation, the peak of 1211m was about 1500Jy (3.1 x 10-11Wm-2f.lm-1) at 15h22m ±lm (52], again sightly earlier than that of the near-infrared. The peak flux of the E impact was larger than these of the A impact for all those observations, and the E residual debris was also larger than the A and C clouds. Detections of the E impact by HST after 19m16s showed a plume in falling phase (53).
5.3 Impact of Fragment G
The impact time of the fragment G, one of the largest fragments, was predicted to be at 07h28m UT, July 18, and was observable around the Pacific, but the large telescopes at Mauna Kea, i.e., Keck, IRTF, CFHT, and UKIRT, could not observe due to some fog and clouds, and also Okayama in daytime could not due to clouds.
Galileo's first detection of the G impact was at 07h33m31s ±5s by UVS at 292 nm, and 33m32s ±5s by PPR at 945 nm, 33m37s ±5s by NIMS at 4.38 /lm [45, 54, 55]. The detection by UVS lasted only 5 seconds, but PPR and NIMS observed the flash during about one minute. Simultaneous measurements by UVS and PPR shows thermal blackbody radiation at 7800( +500,-600) K (54].
The NIMS spectrum at 33m53s, 22 sec after the UVS detection, shows a thermal source of 2200 K [55].
HST first detection of the G impact was at 33m16-46s by 0.89/lm [53], which is recognized to be PCl comparing with the Galileo observations. On the other
41
hand, AAT /IRIS observed the PCl of the G impact at 32m58s by K-band (1.98-2.4 pm) spectral imaging cubes (three dimensional data set) (42). HST detected thermal radiation of the G plume at 35ml6-32s by 0.89 pm, and observed re
flected sunlight from the plume at 38m16-24s (53), which was just before the main event. The G main event started at 39m22s ±5s by the NIMS (55), but at 38m19s ±1m by AAT. This discrepancy of start time may be due to the difference of sensitivities between NIMS and AAT /IRIS. The peak flux of the main event was 1.02 x 103Jy (5.78 x 10-10Wm-2pm-1) by 2.3pm at 07h46m, and CO emission of 2.3 pm was detected at 07h45m by AAT.
As rotating of the impact site into the view of Earth, a huge Earth-sized crescent dark cloud appeared on the morning terminator and easily observable for ground-based observers by small telescopes with the aperture of only several centimeters. Fig. 21 shows the HST image of the G impact cloud.
5.4 Impact of Fragment H
The impact time of the fragment H was predicted to be at 19h26m UT July 18, and was observable for Africa, Europe, and South America.
Pic-du-Midi Observatory (France) 1-m telescope with DENIS camera (Deep Near-Infrared Survey at ESO observatory) detected the PC1 start at 19h31m40s
±5s and the PCl peak at 31m50s ±5s by K'-band imaging [56], but those results were not calibrated absolutely. The PC1 timings by Pic-du-Midi agree with Calar Alto observation of the PC1 at 31m45s ±20s by 2.3 pm (50]. The PC1 flux by Calar Alto was 0.14Jy (7.9 x 10-14�Vm-2pm-1). Just after those detections, Ga/ileo PPR detected the H impact signal from 31m58s during about 30 seconds for 945 nm [45). The peak flux by PPR was 1.8 x 10-8Wm-2pm-1
at 32m00s.
The PC2 of the H impact started at 32m35s ±5s and reached its peak at 33m10s ±5s for Pic-du-Midi K'-band observation. After the peak, the PC2 maintained its nearly peak flux until the main event. Calar Alto 3.5-m tele
scope observed the PC2 start at 32m47s ±12s and the peak of 2.0 Jy (1.1 x
10-12Wm-2pm-1) at 33m10s ±lOs by 2.3pm, and 2.2-m telescope with a pho
tometer observed same timings of the PC2 with 3.5-m by 3.1 pm, although this photometry was not calibrated absolutely. By 2.3 pm, the PC2 also maintained its flux until 19h35m, then decreased slowly, and by 3.1 pm the flux increased.
The decrease in the 2.3 pm signal between 19h35m and 19h37m and the con
current increase in the 3.1 pm signal suggests a temperature of 900-1300 K at this time. On the other hand, SAAO 0.75-m telescope observed the PC2 peak of 1.3 x 10-12Wm-2pm-1 (2.0Jy) at 33m53s ±15s, later than Pic-du-Midi and Calar Alto, by K-band imaging (51). For mid-infrared, European Southern Ob
servatory (ESO, La Silla, Chile) 3.6-m telescope with TIMMI (Thermal Infrared Multi Mode Instrument) camera observed the PC2 starting at 32m54-57s, and the peak at 33m20s ±3s for 9.1-10.4 pm (57). The PC2 flux was 0.025 of the maximum flux of the H main event by ESO /TIMMI 10 pm.
Figure 21: HST images of the G impact debris cloud. The smaller feature to the left of the G cloud is the D cloud. The left image was taken through a green filter
(555
nm), and the right image shows the same site seen through a filter sensitive to methane (at 889 nm). The dark thick outermost ring's inner edge has a radius of 6000 km and outer edge has a radius of 13000 km. Credit by Hammel (MIT), and NASA.43
According to the Calar Alto observation, the main event of the H impact started at 37m27s ±15s for 3.5-m/2.3 JJ.m, and reached the peak at 44m04s ±9s for 2.2-m/3.1 JJ.m. Also K-band spectra were obtained for near maximum of the main event by the 3.5-m telescope. A spectrum at 41m44s was dominated by continuum emission, but at the peak of the main event, 44m39s, CO band emis
sion (longer to 2.3 JJ.m) and molecular hydrogen emission (shorter to 2.15 JJ.m)
were prominent in the spectrum. By the Pic-du-Midi observation, the start of the main event was about 37m, and the images are saturated from 19h39m to 19h5lm. But the peak was estimated to be at 44m20s by looking to the diame
ter of the saturated images. At SAAO, the start time was 37m40s ±15s, and the peak flux was 4.8 X 10-11 wm-2 jJ.m-1 (78 Jy) at 44m56s ±15s. In the increas
ing phase of the main event, the brightening rate became to decrease at 40m25s
±15s. The end of main event was at 50m27s ±30s, and the end of plateau of the H impact was at 55m58s ±30s by SAAO, and those are consistent with the Pic-du-Midi observation [58]. For Calar Alto, H ami/ton et a/.( 1995) reported the end of main event at 56m06s ±30s [50], but this is the end of plateau of the H impact.
For the mid-infrared observations, ESO /TIMMI observed a gradual increase of the flux from 1 9h36m40s ±20s and the following rapid increase from 19h37m30s
±20s, and NOT detected the main event at about 19h37m ±lm. The peak of the main event was 2900Jy (6.0 x 10-11wm-2JJ.m-1 ) at 19h44m ±lm by NOT
l2jJ.m, but ESO/TIMMI reported the peak at 42m40s ±lOs with better time resolution for 10 JJ.m. Also ESO /TIMMI observed the end of main event at 48m20s ±20s, which was again earlier than that of 2 JJ.m. It is noted that the radiation area expanded parallel to Jovian limb from 51m08s, and that this apparent size was 10,000 km.
5.5 Impact of Fragment L
The impact time of the fragment L, one of the largest fragments, was predicted to be at 22h09m UT, July 19, and was observable for the Americas and the Atlantic.
Calar Alto 3.5-m telescope observed the initial phase of the L impact with excellent resolution for 2.3 JJ.m: the flux increased gradually from 22h16ml8s
±3s, and PCl started at 16m32s ±3s, and reached its maximum of 0.36 J y (2.0 x 10-13Wm-2JJ.m-1 ) at 16m41s ±3s [50]. Since the lightcurve of the L impact at Calar Alto shows gradual increase of flux before sharp increase of PCl, this slow brightening may be PCO of the L impact, similar to that of the K impact. By Pic-du-Midi K-band imaging, PCl of the L impact was first detected at 16m36s ±20s and reached to peak at 16m54s ±20s, but by the J- and H-band images PC1 did not appear [56]. Galileo PPR observed the L impact for 945 nm and detected first signal at 16m48s ±ls, and this flash lasted during 40 seconds [45]. The PPR peak flux was 2.6 x 10-8Wm-2JJ.m-1 at the distance of 1.6 AU between Jupiter and Galileo.
At Calar Alto observation, PC2 of the L event started at 17m27s ±3s, but the PC2 peak, larger than 5Jy (2.8 x 10-13Wm-2J..lm-1), caused saturation of the detector. Pic-du-Midi observed using three filters, J, H, and K bands with time resolution of 16-18 sec for each filter. PC2 started between 17m17s and 17m37s for K-band, and at 17m30m ±5s for H-band, but the start time was not clear for J-band. The maximum time of the PC2 is measured to be for each filter as 17m4ls, 17m47s, and 17m58s in J, H, and K bands, respectively.
Relative peaks of the J, H, and K-bands were 0.18, 0.6, and 28, respectively.
NOT also detected the PC2 from 19hl8m to 19h20m by mid-infrared of 12 J..lm, and showed the peak of the PC2 was between 18m00s ±5s and 18m31s [52]. The peak flux was larger than 300 J y (6.2 x 10-12 W m -2 J..lm-1). The difference of the peak times between the wavelengths used Pic-du-Midi and NOT are discussed by Drossart et a/.(1995) [58].
The infrared detectors of Calar Alto and Pic-du-Midi were saturated with the L main event, which started at 23m00s ±23s (Pic-du-Midi, K-band). By looking to the diameter of the saturated images of Pic-du-Midi, the peak of the main event was estimated to 30m30s. But by 12 J..lffi NOT observed the main event from 21m20s ±30s, 100 seconds earlier than that by the K-band. The peak of the main event was about 13,000Jy (2.7 x 10-10Wm-2J..lm-1) at about 31m, which agrees with the peak of the K-band. Lagage et a/.(1995) combined the peak flux of NOT 10 J..lffi with another 2 J..lm observation, to estimate a color temperature of 700 K at the peak [52]. After the peak of the main event, the radiation spot was spatially resolved, as indicated by its elliptical shape. A maximum spot extension was about 30,000 km. This expansion of radiative area was also shown in the H impact [57].
The L plume could be detected by ground-based observations for shorter, near-visible infrared wavelengths. The 70-cm solar Vacuum Tower Telescope (VTT) of the Kiepenheuer-Institute on Tenerife observed the L impact by near
infrared methane band at 892 nm [59], and the 1-m Jacobus Kapteyn Telescope (JKT) at La Palma observed with a filter at 0.9075 J..lm [60]. For those wave
lengths, the observed flux seems to be reflected sunlight from the impact plume rather than thermal radiation from a hot material [60]. Both lightcurves ob
tained by VTT and JKT are very similar. VTT, having better time resolution than JKT, detected the L plume first at 22h20m52s ±20s, and JKT detected first at 22ml5s ±2m. The flux increased gradually to reach the maximum at 24m52s ±20s for VTT and at 24m47s ±76s for JKT, and then decreased. The maximum flux at JKT was 1.0 x 10-11Wm-2J..lm-1 (2.7 Jy). The maximum flux at VTT was 2.1 x 10-4 of the total flux of Jupiter, which is estimated to be 6.2 x 10-12Wm-2J..lm-1 (1.6Jy) [61]. After the maximum, however, both observatories observed a sudden bright flare, whose peak was at 3lm52s ±20s with a lifetime of 90 seconds for VTT, and at 3lm46s ±76s, only one observing point, for JKT. By JKT, the flare flux was 1.22 x 10-11Wm-2J..lm-1 (3.3Jy).
The flare flux at VTT was 3.4 x 10-4 of the total flux of Jupiter, which is equal to be 1.0 x lo-11wm-2J..lm-1 (2.7 Jy) [61]. Also Schleicher et a/.(1995)
45
reported the Gregory-Coude-Telescope observation at 611 nm, which observed the plume but could not detect this flare event [59). If this flare is thermal ra
diation, the color temperature of this flare can be estimated by the NOT 12 pm and the VTT and JKT optical observations. At 32m, which is one minute after the peak by the NOT observation, the 12 pm flux by NOT is about 11,000 J y (2.3 x 10-10Wm-2pm-1 ). The pure flare fluxes, or the excess fluxes of the flare flux from the decreasing trends of the lightcurves are 7.1 x 10-12Wm-2pm-1 at VTT, and 6.1 x 10-12Wm-2pm-1 at JKT. The estimated color temperatures are 900 K (JKT) and 925 K (VTT).
5.6 Impacts of Fragments Ql and Q2
The impact times of the fragments Q2 and Q1 were predicted to be at 19h31m UT and 19h59m July 20, respectively, and were observable for Africa, Europe and South America.
The Q2 impact could be detected only by Calar Alto 2.3 pm observation [50]. They detected a precursor at 19h44m47s ±3s, and a faint main event from 52m24s ±15s to 56m ±1m. The peaks of the precursor and the main event are about 0.015 J y (8.5 X 10-15 Wm -2 pm-1) and 0.01 Jy (5.7 X 10-15 wm-2 pm-1 ), respectively. It is not clear whether this precursor is PC1 or PC2. Time dura
tion of about 7m30s between the precursor and the main event is longer than durations of about 6 minutes for other impact events.
The Q1 impact was detected at 20h13m52s ±1s by Galileo PPR with two wavelengths of 678 nm and 945 nm [45). The ratio of the 678 nm flux to the 945 nm flux indicates a color temperature of 18,000 K, if blackbody emission is assumed. At 13m15s ±3s Calar Alto by 2.3 pm detected a precursor of the Q1 impact, which is considered to be PC1 of the Q1 impact, comparing to the Galileo PPR time, and the precursor peak was 0.040-0.050 Jy ( (2.2-2.8)x10-14Wm-2pm-1) at 13m30s ±20s [50]. However, no distinct second precursor was detected. The Calar Alto lightcurve of the Q impacts shows a small bump, whose peak was at 16m ±1m, just after the PCl. This bump may be the PC2 of the Q1 impact.
The main event started at 19m47s ±3s at Calar Alto, and at 20m03s ±15s at SAAO K-band [51), and at 19m30s ±1m at NOT 12 pm [52]. The peak of the main event was 5.6 x 1o-uwm-2pm-1 (90Jy) at 23m50s ±lOs at SAAO, and 1700 Jy (3.5 X 10-11 wm-2 pm-1) at 24m ±1m at NOT.
5.7 Impact of Fragment R
The impact time of the fragment R was predicted to be at 05h25m UT, July 21, and was observable around the Pacific. But Okayama could not observe the R impact due to bad weather.
Galileo PPR observed the R impact for 678 nm, but could not detect any
thing for 5.3 seconds sampling [45). On the other hand, NIMS detected the R
signal from 05h35m08s -22/ + lls (55).
The R impact was observed by many ground-based observatories; the 10-m W. M. Keck telescope by 2.3 pm imaging [62] and IRTF by mid-infrared imagings of three wavelengths (43] at Mauna Kea, the 5-m Hale telescope at Palomar Observatory by 3.2 pm and 4.5 pm imagings and 8-13 pm spectroscopy (63], and AAT/IRIS spectral imaging in K-band [42).
Keck observation with good time resolution of 8 seconds showed that PCl of the R impact started at 05h34m44.5s ±4s and reached its peak of 0.43 Jy (2.43 x l0-13vvm-2pm-1) at 34m55.2s ±4s. Sequential PC2 began at 35m46.4s ±4s and reached its peak of 0.52Jy (2.9 x 10-13Wm-2pm-1) at 35m54.ls ±4s, and then decreased slowly. On the other hand, AAT, observing by same wavelength with Keck, detected first signal of the R impact at 34m00s, but with worse time resolution of ±2m.
For longer wavelengths, Palomar detected, but could not temporally resolve PCl of the R impact, because of shorter lasting time of PCl than that by 2.3 pm and long time resolutions of 30 sec at 3.2 pm, and of 10 sec at 4.5 pm.
The detected times of PCl were 34m50s ±30s at 3.2 JJm, and 34m52s ±lOs at 4.5 pm. The flux of PCl by 3.2 pm was 1.5 Jy ( 4.4 x l0-13W m-2 pm -1 ), and that by 4.5 pm was 1.8 J y (2.6 x 10-13Wm-2 pm-1 ). At 4.5 pm, PC2 started at 35m22s ±lOs and reached its maximum of 10.8 Jy (1.60 x 10-12Wm-2 pm-1) at 35m48s ±lOs, then decayed quickly until 36m20s. At 3.2 pm, the rising phase of PC2 was unresolved and the peak flux was 3.2Jy (9.4 x l0-13Wm-2pm-1) at 35m5ls -1m/ +30s, but the decay time of PC2 at 3.2 JJ.m was longer than that at 4.5 JJ.m, but shorter than that at 2.3 pm, and the PC2 end was at 37m30s-38m00s. IRTF could not detect any precursor by all three thermal infrareds of 7.85, 10.3, and 12.2 pm with time interval of 17 sec between observations at different wavelengths.
The main event of the R impact started at 40m57s ±4s by Keck, but at 38m30s-39m00s by 3.2 JJ.m and 39m40s ±lOs by 4.5 pm of Palomar, and at 40m57s by 7.85 pm of IRTF (no signs at 40m40s by 12.2 pm). The peak times of the main event well coincided in those observations ; 44m57s by Keck 2.3pm, broad peak at ""'44m55s by Palomar 3.2/4.5pm, 44m30s by 7.85pm and 43m50s-44m50s by 10.3 pm, and double peaks at 43m20s and 46m by 12.2 pm of IRTF with 51 seconds resolution. The peak brightness of 2.3 pm was 65 Jy (3.7 X lo-ll wm-2 pm-1) by Keck, but 400 J y (2.3 X l0-10 wm-2 pm-1) at 05h46m by AAT. The 3.2 JJ.m peak was 600 Jy (1.7x 10-10Wm-2 pm-1 ), and the 4.5pm peak was 800Jy (1.2x 10-10Wm-2pm-1) by Palomar. The IRTF peaks were 0.84 x 10-10Wm-2pm-1 (1730Jy) by 7.85pm, 0.39 x 10-10Wm-2pm-1 (1380Jy) by 10.3pm, and 0.17 x 10-10Wm-2pm-1 (844Jy) at 43m20s and 0.20 x 10-10Wm-2pm-1 (993Jy) at 46m by 12.2pm.
Some infrared spectrum also were obtained for the R impact. Nicholson et a/.(1995) shows three spectrum in 8-13 pm obtained by Palomar (63). In a spec
trum at 05h37m, between PC2 and the main event, no impact signal appeared, although in this time the bright spot after PC2 was detected in the 2.3, 3.2, and
47
4.5 J-Lm imagings. Next spectrum at 05h43m, just before the peak of the main event, showed remarkably enhanced flux in addition to an emission from silicate dust grains as a 10 J-Lm bump in the spectrum. Nicholson et a/.(1995) also cal
culated color temperatures from Palomar and Keck observations, and estimated the PC1 temperature of 1000(±120) K, and the PC2 peak of 630(±50) K, which is unexpectedly low, but 950(±150) K at 36m30s and 820(±100) K just before the main event. In the main event, the color temperature remains relatively constant, varying from a minimum of 600 K at 05h41m to a maximum of 1150 K at 05h48m. On the other hand, AAT spectral imaging shows CO erruss10ns (2.3-2.4J-Lm) after 05h47m in K-band (2.0-2.4J-Lm) [42].
5.8 Impact of Fragment S
The impact time of the fragment S was predicted to be at 15h10m UT, July 21, and was observable for Africa and Europe, but in daytime.
In K-band imaging of the S impact by SAAO 75-cm telescope, a precursor was detected, but its start time is uncertain [51]. This precursor showed two peaks at 15h16m36s ±15s and 17m36s ±15s. Those peaks might correspond to PC1 and PC2, respectively. But in another SAAO observation by 1.9-m telescope with an infrared photometer, only one precursor was observed at about 05h 18m and lasted for about 45 seconds for K-band photometry [64]. Calar Alto could not detect any precursor due to various effects [50].
The main event began at 21m40s ±30s for SAAO 75-cm, and at 22m00s for SAAO 1.9-m, and reached the peak of 1.12 x 10-10Wm-2J-Lm-1 (180Jy) at 29m16s ±15s by SAAO 75-cm.
Galileo PPR observed the S impact, but its observational window from 15h21m to 15h59m did not cover the entry of the S impact, and thus could not detect any signal [45].
5.9 Impact of Fragment W
The impact time of the fragment W, last impact of SL9, was predicted to be at 07h57m UT, July 22, and was observable around the Pacific. However, clouds obscured skies of Okayama and Hawaii.
Galileo SSI observed the W impact for 0.56 J-Lm and detected the W impact signal, starting at 08h06m14s ±2s and fading over 7 sec [46). HST observed by 540 nm at 06m16s (the integration time of 0.3 sec), which happened to be same time of the SSI detection (53]. This HST image shows a bright emission point in .Jupiter's shadow. AAT/IRIS detected a precursor at 06m56s [42], which is considered to be PC2 by comparing with SSI and HST detections. And AAT observed the main event of the W impact from 12m20s ±2m, the peak of 200.Jy (1.1 x l0-10Wm-2J-Lm-1) at 18m ±2m by 2.3J-Lm, and CO emission from 20m ±24m. HST observed the W impact plume as compact point-like object
at 09m16-26s by 409 nm, but large plume at 16ml6-34s in the main event by 332 nm.
5.10 Impacts of Other Fragments
The impacts of fragments B, F, P2, T, U, and V showed faint brightenings and small debris clouds, or no detectable brightenings and clouds. These impacts are introduced in this subsection, but are not discussed in the following sections of this paper, because of the absent of detail informations on these impacts.
The B impact was detected at July 17.122 UT (2h56m, July 17) by Keck 3.27-3.44 J-Lm [65], but the details of this observation did not published yet. Even after the careful inspections of the data, no signatures of the B impact could be seen by IRTF 2.2 J-lffi [43] and ESO 3.6-m 10 J-Lm observations [57].
ESO 3.6-m/TIMMI detected a minute signature of the F impact at July 18.060 UT (Olh26m, July 18) by mid-infrared [57, 66]. However, IRTF could detecte no obvious effect of the F impact by 10.3 J-Lm, although there were some clouds in the sky [43].
Hamilton et a/.(1995) reported a mystery spot at 22h, July 19 before the L impact by 2.3 J-Lm imagings of Calar Alto 3.5-m telescope [50]. This spot may be a debris cloud of impact of fragment J, which faded from view in December
1993.
Also the impact of disappeared fragment M was detected at July 20, 20.259 UT (06h13m) by 2.1 J-Lm of the San Pedro Martir of the Mexican National As
tronomical Observatory [67], and at 20.256 UT (06h09m) by Keck [68].
Any signatures could not be detected at the P2 impact by Calar Alto [50], SAAO [51], Teide Observatory and Pic-du-Midi [68], at the T impact by Calar Alto and SAAO, and at the U impact by Palomar, Teide, Swedish Solar Tele
scope at La Palma, University of Massachusetts Whately Observatory, and Mc
Donald Observatory [69].
The V impact was observed at 04h23m, 22 July by AAT 2.35 J-Lm imaging, but the flux of the V impact was not reported [42]. Other observatories (Mc
Donald, Palomar, Steward Observatory) could not detected any signatures of the V impact [69].
5.11 HST Observations
HST observed the debris clouds besides the impact plumes and measured the positions and sizes of the clouds [53]. Table 3 is a summary of the impact positions and the cloud sizes.
The crescent ejecta cloud by the G impact is measured to have inner radius of 6000 km and outer radius of 13,000 km [53]. The K and L clouds have almost the same sizes to the G cloud. Sizes of other impact clouds are not measured and unknown.
49
Table 3: Summary of detected impact locations and relative cloud sizes
Fragment HST Detections Class
Latitude Longitude Method A -43.41 ±0.05 187.8 ±.03 wave 2a
-43.54 ±1.0 186.3 ±2.0 site
B -42.79 ±1.0 71.1 ±2.0 site 3
c -43.41 ±1.0 225.0 ±2.0 site 2a
D -43.29 ±1.0 33.5 ±2.0 site 3
E -43.48 ±0.05 153.5 ±0.2 wave 2a -44.54 ±1.0 153.5 ±2.0 site
G -43.65 ±0.04 25.7 ±0.2 wave 1
-43.66 ±1.0 26.8 ±2.0 site H -43.66 ±1.0 101.4 ±2.0 site 2a K -43.29 ±1.0 282.6 ±2.0 site 1 L -42.79 ±1.0 351.6 ±2.0 site 1 N -43.41 ±1.0 73.1 ±2.0 site 3
Q2 -44.67 ±1.0 47.5 ±2.0 site 3
Q1 -44.37 ±0.1 64.0 ±0.5 wave 2b -43.41 ±1.0 66.3 ±2.0 site
R -44.17 ±0.1 46.8 ±0.5 wave 2b
-44.50 ±1.0 43.6 ±2.0 site s -43.91 ±1.0 34.0 ±2.0 site 2c w -44.29 ±1.0 284.8 ±2.0 site 2c
Impact sites for fragments F, P2, T, U, and V were not detected. Latitudes are planetocentric and longitudes are System III. Class is classification of the sizes of the impact debris clouds, based on the first view after impact. Class 1 (G, K, and L) =dark region> 10,000 km, large eject, probably multiple waves;
class 2a (A, C, E, and H) = 4000 km < impact site < 8000 km, medium eject, possibly multiple waves; class 2b (R and Q1) = medium but slightly smaller eject, probably single wave; class 2c (S and W) = < 6000 km, impacts occurred near earlier impact sites and are somewhat confused; class 3 (B, D, N, and Q2)
= < 3000 km, no eject, no wave; class 4 (F, P2, T, U, and V) = not detected, not shown in table. After Hammel et a/.(1995) (53].
Also by HST observations, Hammel et a/.(1995) measured plume heights of the A, E, G, and W impacts above 100 mbar [53]. Fig. 22 shows time variations of the heights after the imapcts. For the A impact, Ham mel et a/.(1995) regared
the first detection time (20h13m16s) by HST as the impact time of the A.
However, since it is better to consider that the first detection at 20hllm29s by Calar Alto is the impact time, the heights of the A plume are plotted with times after the Calar Alto detection in Fig. 22. The impact times will be discussed in detail at Section 6.
51
3500
3000
6A
X
E ,-, 2500
0
G
.::s:.
E
....___.
.E 2000 0 w
en
<J.) I
<J.)
E
:J o_
1500
1000
500
0 -300
•••
•
0 300 600 900 1200 1500
Time [sec] after impact(s)
Figure 22: Plume heights as functions of times after impacts. Revised after
Hammel et a/.(1995) [5
3]
. The observational times are the midpoints of exposures. Open marks show the reflected sunlight by the plume, and closed marks show the thermal radiation in shadow. The connecting lines are drawn to aid the eyesight, but appear to show plume evolutions. A question mark on the A line at 240 sec shows a HST observation showing no impact signature at 20h15ml8s.
Uncertainties in the measurements are of the order of 370 km for the A plume and 170 km for the other plumes.