Photometry of upsilon Geminorum during the graze of UT 2005 Apr 16

Michael Richmond
May 14, 2005

Data files can be found at the end of this page

On UT 2005 April 16 (the evening of April 15, 2005 in the US), the Moon slipped across the sky very close to the star upsilon Geminorum. From some locations, the star appeared to graze the dark limb of the moon. With careful measurements of the times at which the star disappears and re-appears, one can figure out properties of the Moon's limb.

A number of people used the IOTAoccultations group to coordinate their observations and share the results.

I did not observe the event, but I watched and listened as people described their results. Several astronomers recorded the event on videotape (my sincere apologies if I err on names or details -- I'll gladly make corrections if necessary):

The observers digitized their videotapes. Walt Morgan kindly collected the data from the Morana and Breit, placed it all onto a CD, and mailed it to me.

What I have done is to measure the brightness of the star as a function of time during "interesting" periods: when the star disappears and re-appears behind the Moon's limb. Let me explain very briefly what I did:

Because each tape contained thousands of frames, and because the position of the star drifted slowly across each frame with time, I did not measure every frame of each tape. Instead, I picked out sections of 200 to 700 frames around all the events I could see visually by watching the movie on my computer's screen. I ended up selecting the following events, common to all (or almost all) three datasets. In chronological order, they are

                                           is it present in 
  Event    meaning                  Breit         Morgan       Morana
   D1    first disappearance         yes          yes           yes

   F     "flash" -- brief            yes          yes           no
           re-appearance and

   R1    first full reappearance     yes          yes           yes(brief)
   D2    second disappeance          yes          yes           yes

   R2    second reappearance         yes          yes           yes(brief)

   P     small dip in brightness     yes          no            no

   D3    third full disappearance    yes          yes           yes

   R3    final re-appearance         yes          yes           yes

The Breit and Morgan tapes show very similar sequences of events; not surprising, since the telescopes were so close together. The Morana tape features quite a different set of features: the gap between R1 and D2 is much shorter than in the other two tapes, as is the gap between R2 and D3.

There could very well be other events in the tapes. As I stated earlier, I concentrated only on the frames around these obvious features.

Some of the observers recorded the time directly on their videotape; others recorded an audio signal which provides regular timestamps. I did not attempt to convert frame number to proper Universal Times; please contact the individual observers for that information. I did verify that all tapes provide roughly 30 frames per second. In the graphs below, I will usually show sections of 200 frames, which corresponds to about 6 seconds of time.

I will now run through the sequence of events, one tape at a time, and conclude with a short summary of important features.

Events on Breit's tape

Below is event D1 on Breit's tape. I show measurements made through four different apertures. Based on these results, I concluded that the smallest apertures (with radii of 4 pixels and 6 pixels) were best for this dataset.

Note that any value of instrumental magnitude below about 20 means "no light was detected." The star disappears completely in about 8 frames.

Next is the brief "flash" event I call F. The peak appears around frame 1177. Remember that instrumental magnitude values greater than about 20 mean "no signal", and anything less than 19.5 is dubious. The short, small bump between frames 1170 and 1180 is real: one can see the star on the individual frames. At its peak, the star reaches instrumental magnitude 18.5 in the 4-pixel aperture, which is FAR below its magnitude (about 15) at full brightness.

The star re-appears fully at event R1:

Next is the second full disappearance D2:

The star re-appears about five seconds later, event R2:

There is then a brief, partial dip in brightness around frame 3130, which corresponds roughtly to UT 04:40:03.

The star makes a third full disappearance around 04:40:15:

The final event is R3, as the star reappears for good. Note the extennded "precursor" which appears some 50 frames before the main brightening, and the gradual nature of the main brightening itself.

Morgan's tape

The first event on Morgan's tape is D1.

I used larger apertures to measure the stellar brightness on Morgan's tape, as the PSF was larger and somewhat "broken" in appearance. In this case, instrumental magnitude values below about 22 mean "no light was detected."

The "flash" event F does appear very cleanly on Morgan's tape. Note that the star appears to gain its full brightness briefly before disappearing; in Breit's tape, the flash was both much briefer and much fainter. The rise to maximum light is clearly more extended than the fall afterwards.

The star then gradually re-appears fully at R1:

The next two events, D2 and R2, occur relatively close together in time on Morgan's tape: they are separated by only about 130 frames.

I checked carefully for any trace of the "dip" event, P, which appears in Breit's tape between R2 and D3. The event ought to be in the region of the tape shown below, near frame 5413, but I see nothing.

The next event I did see is D3, followed very closely by the final reappearance R3. Note that the following graphs overlap in time by 100 frames.

In between D3 and R3, you can clearly see a real, partial brightening. It lasts for roughly 30 frames.

Morana's tape

I found that small apertures gave the best results on Ed Morana's tape. He was located relatively far from the other observers, so the sequence of events, and their appearance, is quite different on his tape. The instrumental magnitude corresponding to "no light detected" is about 19.0 in the 4-pixel aperture.

We start with D1:

I checked the tape for any evidence of the "flash" event F between D1 and R1, but did not find it.

The next event is therefore R1; but the star re-appears only briefly before it disappears fully again in D2. I guess it's possible that this brief appearance corresponds to event F on the other tapes, but I don't think that the timing is right. Recall that full brightness corresponds to instrumental magnitude approx 16.3 in the 4-pixel aperture, so this brief brightening doesn't reach that level.

The next appearance is again a very brief one, and a faint one. Most of the data shown in the graph below indicates "no light detected". However, the red points (magnitude through aperture of radius 4 pixels) rise above the critical level of 19.0 for an extended period from about frame 4560 to 4590, a duration of about 1 second.

The final event is the complete re-appearance of the star, back to its normal instrumental magnitude of 16.3 or so.

Photometry data files

You can grab copies of the photometric measurements below. Each file has a format something like this:

 1  231.00  445.00     12   2.46   15.097  0.013  14.722  0.011  14.685  0.01 2  14.704  0.014  14.718  0.015    0

where the columns are:

  1. meaningless integer
  2. row position on frame
  3. col position on frame
  4. background sky value (counts)
  5. uncertainty in background sky value (counts)
  6. magnitude through 4-pixel aperture
  7. uncertainty in magnitude through 4-pixel aperture
  8. magnitude through 6-pixel aperture
  9. uncertainty in magnitude through 6-pixel aperture
  10. magnitude through 8-pixel aperture
  11. uncertainty in magnitude through 8-pixel aperture
  12. magnitude through 10-pixel aperture
  13. uncertainty in magnitude through 10-pixel aperture
  14. magnitude through 12-pixel aperture
  15. uncertainty in magnitude through 12-pixel aperture
  16. meaningless flag value

A value of "99.00" for a magnitude or its uncertainty means "no meaningful measurement".

The data files are simple ASCII text. There is one line per frame of videotape. I do not have any internal markers indicating the frame from which each measurement is taken; you will have to add the starting frame number to the line number. Sorry :-(

Perhaps someone else can use these datafiles as a starting point and improve them by adding explicit frame numbers or UT values to each line ....


All three videotapes show evidence for an asymmetry in the light curves of upsilon Geminorum during the graze: in general,

One might explain this repeated pattern by the presence of a faint companion to the star, which appears on the "leading" side of the star (i.e. the side first encountered by the Moon's limb as it sweeps past the star).

The faint, extended "precursor" to event R3, the faint F event on the Breit tape, the faint bump between D3 and R3 on the Morgan tape, and the faint R2/D3 event on the Morana tape, may be additional evidence for a faint companion. Note that the flash F was too bright on the Morgan tape to be ascribed to the companion -- that must have been the primary star itself.

Now, Richard Nolthenius observed this event at 118:12:04 West, +35:07:04 North, near California City, California, several hundred miles to the south of the three observers whose videotapes are described here. His record shows that the extended periods of faint light occurred after the DIS-appearances of the star, rather than before the RE-appearances. That pattern is the opposite to the one recorded on videotape by the Bay Area observers.

Does that cast doubt on the entire idea that ups Gem has a faint companion? Not necessarily. Richard notes that it may simply mean that the companion is nearly due north of the primary, and, from his location, the limb cut across the pair of stars in a slightly different direction.

More comments may be added later ....