Earth, any observers based and around Earth could not see the impact points directly. The impact "point", however, means the crossing point of a fragment, entering to the atmosphere, on the 1 bar pressure level. "Something" on or around the fragment trajectory, positioning high enough above the 1 bar level, were able to be observed from Earth as PC1, if it would sufficiently emit thermal radiation or scatter other light to Earth. The height of Earth-line of sight can be calculated from the predicted or real impact point (Jupiter's latitude and longitude) to be 300-1000 km (see Fig. 26) [29].
Then, what is "something" ? One candidate is the fragment itself which may radiate thermally due to the heating by the atmospheric drag. Another candidate is dust (or comae) around the fragment. The dust may also radiate thermally as meteor during its entry into upper atmosphere or may scatter ther
mal radiation by the entry of the fragment at lower altitude, which is invisible from Earth.
The author, however, rules out the latter by the following reason. During the time difference of 10 seconds between PC1 and the Galileo detection, the fragment fell downward to about 400 km, assuming that the entry speed of 60 km s-1 and the entry angle of 45° did not vary during the entry. If Galileo
detected the entry flash of the fragment around the 1 bar level, at the PC1 the fragment must locate at 400 km height, which is visible from Earth. For the entry at altitude of 400 km, the radiation from the fragment will be brighter than that from the dust around the fragment, because the ablation energy is proportional to the cross section of the entry object [72]. Graham et a/.(1995)
estimated that the PC1 peak flux of the R impact, by their observation with the wavelength of 2.30 J.Lm, was able to account to the entry flash of the fragment with radius of 0.25 km at altitude of 400 km [62]. If ground-based observers saw the scattering light of the entry flash of fragment at 1 bar level from dust at high altitude as the PC1, Gali/eo must see the entry flash of fragment at 400 km below at 10 seconds after the PCl, but such a flash must be obscured by the clouds and may be unobservable.
For the K impact observation of the AAT, as previously described, the flux increased gradually from 10h22m42s, 80 seconds before the PCl at Okayama [42]. In our observation of the K impact, the lightcurve shows similar gradual increase of the flux from the start of the observation, as shown in Fig. 20. Calar Alto observation of the L impact also showed gradual increase of the flux [50].
(In this paper, this phenomenon is called as Oth precursor, PCO, tentatively.) The PCO is distinguishable from the PC1 clearly by the steep start of the PCl.
Probably the PCO is the entry flash of a large amount of dust preceding to the fragment.
6.2 Comparison of Fragments
Because it is clear that the PCl is related to the entry of fragments, one can examine timings of other phenomena, as based on the accepted impact times,
Table 4: Comparisons with Galileo detection times, and PCl times
Galileo PCl
Fragment Detection Instru- Detection
Observa-Time ment Time tory
A No observation 20hllm29s ±05s Calar Alto
B Not seen PPR Not seen
c No return PPR 07hllm39s ±20s AAT
No return NIMS
D No observation llh54m46s? ±Olm AAT
E No observation No observation F No return NIMS Not seen
G 07h33m31s ±05s uvs 07h33m16-46s HST 07h33m32s ±05s PPR 07h32m58s ±0 lm AAT H 19h31m58s PPR 19h31m40s ±05s Pic-du-Midi
19h31m45s ±20s Calar Alto K 10h24m13s ±02s SSI 10h24m02s -09s OAO L 22h16m48s PPR 22h16m32s ±03s Calar Alto
22h16m36s ±20s Pic-du-Midi N 10h29m17s ±02s SSI Not seen AAT,OAO
P2 No return PPR Not seen Calar Alto
Not seen SAAO
Q2 No observation 19h44m47s? ±03s Calar Alto Ql 20h13m52s ±Ols PPR 20h13ml5s ±03s Calar Alto
20h13m19s ±lOs SAAO R 05h35m08s ±lls NIMS 05h34m45s ±04s Keck
05h34m47s ±lOs Palomar s No window PPR 15h16m36s? ±15s SAAO
T No observation Not seen Calar Alto
Not seen SAAO
u No observation Not seen Calar Alto
v No observation Not seen AAT
w 08h06m14s ±02s SSI 08h06m16s-34s HST
For ground-based observations, there were some reports on the B, F, T, U and V impacts, but they are excluded due to no publication on details of these observations. For the D and S impact, the listed PCl might be the PC2. For the E impact, there were no relevant observations.
55
Table 5: Comparisons with predicted impact times and accepted impact times.
Fragment Date Predicted Accepted Observatory
July Time Impact Time /Instrument
A 16 19h59m40s 20hllm29s ±05s Calar Alto B 17 02h54ml3s 02h50m ±03m
c 17 07h02m14s 07hllm39s ±20s AAT D 17 llh47m00s llh53m ±01m
E 17 15h05m31s 15h12m ±02m HST
F 18 00h29m21s 00h37m ±07m
G 18 07h28m32s 07h33m31s ±05s uvs H 18 19h25m53s 19h31m58s ±01s PPR
J 19 01h35m ±60m Calar Alto
K 19 10h18m32s 10h24m13s ±02s SSI L 19 22h08m53s 22h16m48s ±01s PPR
M 20 06h02m ±lOrn
N 20 10h20m02s 10h29m17s ±02s SSI Q2 20 19h31m43s 19h44m47s ±03s Calar Alto Q1 20 19h59m10s 20h13m52s ±01s PPR
R 21 05h25m56s 05h35m08s -22/ + 11s NIMS s 21 15h10m22s 15h16m36s ±30s SAAO
v 22 04h16m53s 04h23m? AAT
w 22 07h57m36s 08h06m14s ±02s SSI
The predicted impact times are by Chodas and Yeomans (JPL/Caltech) (see Table 1) [21]. The accepted impact times are a revised version of the accepted times by Yeomans and Chodas [73]. The accepted impact times are determined using various sources, weighted in the following order of priority : 1) Galileo timings (Fragments G, H, K, L, N, Ql, R, W), 2) the precursors timings (Frag
ments A, C, D, Q2, S, V), and 3) HST observations of impact sites and waves (E). Impacts of fragments P2, T, and U were not detected.
which are the Gali/eo first detection times or the PC1 start times as shown in Table 5 [73]. Similar analysis is carried out by Yamamoto et a/.(1995) [74].
Timings are measured for characteristic points in lightcurve, shown Fig. 23: the start time and the peak time of the PC2, the start time of main event, time of bending point in the increasing phase of main event, the peak and end times of main event, and the end time of plateau. The bending point in the increasing phase is characteristic feature seen in lightcurve with ordinate of log-scale flux for wavelength inK-band. In lightcurve with ordinate of liner-scale, the bending point is seen as temporal stop of rapid increasing, as the lightcurve of the C impact by IRTF (see Fig. 16). In addition, the start times of CO emission are measured from some spectroscopic observations.
Tables 6, 7 and Figs. 24, 25 give the comparisons of timings of the PC2 and the main event among the fragments. Although some data points in Fig. 24 have large uncertainty on time, there is a tendency that the latter impacts have a shorter interval between the impact and the PC2. This suggests relation of the PC2 with the impact geometry, since the impact point of latter fragment became to be closer to the face-side of Jupiter, as viewed from Earth.
On the other hand, at first sight of Fig. 25 there are little differences of the timings of phenomena in the main event among fragments. In particular, for all fragments, whose main events were observed, the start times of the main events agree very well and are 5-6 minutes after the impacts. However, careful inspection of Fig. 25 shows that the peak times and the end times of the main events appear to be related to the sizes of fragments. The impact of larger fragment generally induced a longer lasting and brighter main event, and larger crescent cloud. It is noted that the start times of CO emissions are independent of fragments.
6.3 PC2 and Plume
The PC1 is an entry phenomenon of fragment at high altitude, then the PC2 is thought to be a subsequent phenomenon of the entry, the expanding and rising hot plume.
Early evolution of the plume was observed by Galileo, in particular by the NIMS instrument for the G impact [55]. The NIMS infrared spectroscopic ob
servation of the G impact reveals that a small hot fireball, whose temperature was 2200 K, rose rapidly and that the vertical velocity increased with time, from
1 km s-1 at the initial to 2-3 km s-1 after 20-30 sec.
The subsequent evolution of the plume was observed by HST [53]. By some visible wavelengths, HST mainly caught reflected sunlight from the rising plumes induced by the impacts A, E, G, and W. The top heights of these plumes varied parabolically with time (Fig. 22), which means that each plume traveled along a simple ballistic trajectory. All the maximum heights of these four plume were same, 3000-3300 km above 100 mbar height, although the brightness of the fragments before the impacts were much different each other [16]. The distances
57
of the debris clouds, which were thought to be aftermath of the plume falling down to the atmosphere, from the impact site were about 10,000 km [53]. If the plume was ejected at elevation angle of 45° as predicted by Takata et a/.(1994)
and Crawford et a/.(1994) (27, 29], the initial ejected speed of 17km s-1, whose vertical velocity is 12 km s-1, can explain both the maximum height of 3300 km and the maximum ejected distance of about 10,000 km [71].
Next the relation between the plume and the PC2 is considered. The time variation of the top height of the plume is calculated, based on the above ballistic model. The plume must rise above the height of Earth line of sight if the thermal radiation from plume can be observed for Earth, and above the sunlit height if the reflected sunlight by the plume can be observed. The two geometrical heights at the impact sites of some fragment are calculated, and compared them to the top height of the plume [71]. Fig. 26 shows the time variations of the top height of the plume, the heights of Earth line of sight, and the sunlit heights for some impacts. The start of the PC2 agrees with the crossing time of the ballistic height of the plume top with the height of Earth line of sight. Thus it is concluded that the PC2 is the thermal radiation from the hot plume rising above the limb, as viewed from Earth.
,..., -10
..- 1 0
I
N I
E ::t
E
� 1 0-11
1 0-13
-5 0
PC2 peak
5
Start of marn event
10 15
End of marn event
20
Time [min] after impact
25
Figure 23: Characteristic points in lightcurve. Timings of the characteristic points are measured from
Gali/eo
detection time or PCl time, difference between two times is about 10 seconds. Revised afterYamamoto et a/.(1995) (74].
59
30
Table 6: Timings of the PC2.
fragment 1994 July Accepted PC2
day I hour Impact start time peak time 0 bservatory A 16 1 20 llm29s 13m5ls ±15s 14m5ls ±15s SAAO c 11 1 01 llm39s 12m07s ±09s 13m2ls ±09s OAO
D 11 I 11 53m±Olm 54m46s? ±Olm AAT
E 11 1 15 12m±02m ?
G 18 1 01 33m32s 35ml6s ±Olm HST
H 18 1 19 3lm58s 32m35s ±05s 33ml0s ±05s Pic-du-Midi 32m47s ±12s 33ml5s ±15s Calar Alto 33m00s ±03s 33rn20s ±03s ESOITIMMI K 19 1 10 24rnl3s 25ml7s ±09s 25m26s ±lOs OAO
L 19 1 22 16m48s 17m27s ±03s Calar Alto
17m4ls ±15s Pic- J-band 17m30s ±05s 17m47s ±15s -du- H-band 17m27s ±lOs 17m58s ±15s Midi K-band 17m35s ±15s 18m00-30s NOT
N 20 1 10 29rnl7s Not seen AAT
Ql 20 1 20 13m52s 15m20s ±Olm Calar Alto R 21 1 05 35m08s 35m46s ±07s 35m54s ±07s Keck
35m22s ±lOs 35rn43s ±lOs Palomar s 21 1 15 16m±02m 17m36s? ±15s SAAO
w 22 1 08 06ml4s 06rnl6-34s HST
06m56s ±Olm AAT
Table 7: Timings of the main event
fragment Main Event
Obser-Start time Bending time CO emission vatory
A 16m5ls ±15s 20m22s ±15s SAAO
16m56s ±05s Calar Alto
c 17m00s ±lOs OAO
2lm30s ±lOs IRTF
16m45s ±30s 24m ±lm AAT
D 00ml2s ±lOs OAO
00m30s ±02m 06m ±2m AAT
E 17m56s ±05s Calar Alto
18m ±02m NOT
G 38ml9s±lm? 45m ±2m AAT
39m22s±05s NIMS
H 37m27s ±15s 44m39s Calar Alto
37m00s ±20s Pic-du-Midi
37m40s ±15s 40m25s ±15s SAAO
37m00s ±Olm NOT
K 30m24s ±09s 3lm57s ±lOs OAO
36m ±lm AAT
L 23m00s ±23s Pic-du-Midi
23m00-33s NOT
N 35m ±Olm AAT
Ql 20m03s ±15s 2lm56s ±lOs SAAO
19m47s ±03s Calar Alto
19m30s ±Olm NOT
R 40m57s ±08s 43m00s ±30s Keck
38m30s ±30s Palomar 3.2JJm
39m40s ±lOs Palomar 4.5JJm
40m57s ±17s IRTF 7.85JJm
47m ±2m AAT
s 2lm40s ±30s 24m42s ±30s SAAO
w 12m20s ±Olm 20m ±2m AAT
61
Table.7 (Continue)
Fragment Main Event Plateau
Obser-Maximum time End time End time vatory A 24m52s ±15s 29m22s ±15s 32m53s ±15s SAAO
c 27m00s ±lOs 33m00s ±lOs OAO
22m55s ±lOs IRTF
D 02m44s ±Olm 06m ±Olm OAO
04m ±02m AAT
E 22m42s ±30s 28m43s ±30s (37m ±Olm) SAAO
22m ±02m NOT
G 46m ±Olm 53m ±Olm 61m ±Olm AAT
H 44m04s ±09s 56m06s ±30s Calar Alto
44m20s ±30s Pic-du-Midi
44m56s ±15s 50m27s ±30s 55m58s ±30s SAAO
44m ±Olm NOT
42m40s ±lOs 48m20s ±20s ESO/TIMMI
K 38m37s ±lOs 42m10s ±lOs 46m57s ±lOs OAO
L 30m30s ±30s Pic-du-Midi
31m ±Olm 35m ±02m NOT
N 37m ±Olm AAT/IRIS
Ql 23m50s ±lOs 29m55s ±lOs SAAO
24m ±Olm NOT
R 44m57s ±08s 51m00s ±20s 55m ±Olm Keck
44m41s ±30s 51m00s ±Olm 54m ±Olm Palomar 3.2pm
44m59s ±lOs Palomar 4.5pm
44m30s ±51s IRTF 7.85pm
43m50s-44m50s 10.3 J..lffi
43m20s, 46m 12.2 J..lffi
s 29m16s ±15s 34m50s ±30s SAAO
w 18m ±02m 22m ±Olm AAT
(/)
+oJ c
<D
E
0) co
I...-LL