Updated/revised 96072B magnitude analysis

Ted Molczan (molczan@fox.nstn.ca)
Sat, 4 Jan 1997 22:41:45 -0500

I have updated and revised my 96072B magnitude analysis, and
comparison with other Titan IV 2nd stages:
Object   State    Obs  Std Mag  Payload
------  --------  ---  -------  -------
90050E  flashing   19   4.67     NOSS

91017B  flashing  124   5.14     Lacrosse
91017B  steady    494   4.82

91076B  flashing   52   5.33     NOSS

96029B  both       87   5.16     NOSS

96038B  flashing   28   4.46     SDS
weighted ave      804   4.92     All

96072B  both       20   3.03     KeyHole

The std mag is defined for a range of 1000 km, and a phase angle
(phi) of 90 deg. At phi=90, a sphere would be 50 percent illuminated.
[100 * (1 + COS(phi))/2]

First, the revisions. Rainer Kracht has informed me of an error 
in his recent analysis of 96029B and 96038B, which resulted in 
including only the maxima of flashing observations, instead the 
average of minima and maxima. As a result, the standard magnitudes
I reported on 31 Dec, were too bright. Rainer was unable to revise
the analysis on short notice, so he kindly provided me the data,
and I completed it for this report.

The table shows that all five of the Titan 4 2nd stages observed
prior to 96072B had std magnitudes around 5; the weighted average
of all 804 observations (every single one by Russell Eberst!)
was about 4.9.

As for 96072B, there are now 20 reliable reports from 4 Australian
observers: Anthony Beresford, Gordon Garradd, Robert McNaught
and Ian Porter.

There are very many person-years of observing experience among them, 
so I am confident in their reports. Their observations, in many cases, 
were more general than Russell's; however, typically there was 
sufficient information to enable me to confidently reconstruct the 
circumstances. I discarded a few obs that were completely ambiguous.

I have found that using a std mag of almost exactly 3.0 enables me 
to account for the observations - that is the average of the 
difference between prediction and observation is nearly zero.

As a result, I find that 96072B is 1.9 magnitudes brighter than all 
five of its predecessors. I do not know the cause.

I have not done a formal analysis of statistical significance; however, 
a casual comparison of the data shows that there is little overlap 
between 96072B and 96029B or 96038B - that is 96072B rarely is as 
faint as the others, and they are rarely as bright as 96072B. Six of 
the 20 96072B observed std mags were brighter than the others' single 
brightest std mag.

Also, this 1.9 std mag variation is much larger than that of any of
the other rocket body types analyzed by Rainer. Typically, they fall
within a range of about 0.5 mag, rarely to 1 mag, for large samples.

Clear skies!
Ted Molczan