I have looked through all suitable images close to full moon (i.e. at high Sun-Earth-Moon angle) and only came up with a few relevant nights, all within 10 degrees of full moon. We were keen to make a colour map across the face for figuring out what colour range is there, and what colours are reflected under BS sunlight, rather than DS Earthlight.
JD2455814: is the best full moon data available — it’s shown on the left. We ran it through our colour pipeline and get a B-V ~ 0.9. The B and V images turn out to be slightly different sizes, which made the colour map hard to produce. Later it has turned out that this may be due to distortions introduced in the shift to align the images at all. Work is proceeding on this.
JD2455905: data taken when we were following an eclipse — Dec 09, 2011, total lunar eclipse. No non-eclipsed images, so nothing useful on the Brightside colour night (we should check the Moon is red though).
JD2456082 is also eclipsed — partially. No useful colour data as we have only eclipsed images. June 3, 2012
JD2455847 looked like a promising night — full moon at an airmass of 1.54. However, when I make a colour map (right panel above), it has a B-V of ~0.4 across the face. Can only assume that the exposure times are wrong in the headers (as has happened occasionally). One corner of the moon is slightlty eclipsed for
some odd reason in one filter only (V) — opposite Tycho. Perhaps we hit
the dome in this filter? These data should clearly not be trusted further.
Right you are — the variances came out *roughly* in accord with expectations, but not at a level where we could use them for image discrimination — at least that’s my recollection of where we got to…
Great job on sifting through all near-Full exposures! Yes, we did tend to start observing late into an eclipse and not go on till it was over. SO we only have one good night, with B and V? Shoot.
Yes, the shutter precision is a problem. I think we did hit a few roofs on buildings now and then.
In a perfect world the counting statistics would be a ‘proxy’ for the exposure time:
t_exp = sigma^2/(mean flux).
I have seen people from Aarhus Astro use that trick. Our own looks at how the variance of pixels depend on signal strength were, however, not clear-cut I seem to recall?