Rasicam v QC¶
Rasicam puts out various numbers that attempt to measure the extent of cloud cover. Several of these
values are now written to the DM Oracle tables in tables gruendl.rasicam (basic rasicam data) and
gruendl.rasicam_decam (rasicam data correlated with exposure number). Independently, I have been
measuring cloud extinction by matching stars against the Nomad and APASS catalogs; let us call that
the QC Cloud parameter (in magnitudes).
I have compared the two sets of data for one night - 20131030, which is a night of highly variable
cloudiness, and thus a good test. The five parameters in the table are "gskyvar", "gskyhot", "lskyvar",
"lskyhot", and "lskypow".
gskyhot, lskyhot, and lskypow all show correlations; however, lskypow does the best by far:
1. The horizontal axis is exposure number, not UT. Just easier to plot.
2. I have shifted the lskypow curve left by 1.5 exposures, since there seems to be an offset in the
matchup of nearest rasicam image to an exposure. Presumably one could fine-tune the matchup using
UT timestamps rather than exposure number.
Here is a night (20131001) where rasicam's scaling is off at the start of the night:
Also, around exposure 492, the Blanco's pointing hits the rasicam spider, where lskypow is not useful.
A chunk of cloud passes through and is seen by QC but not rasicam.
Conclusion: Rasicam's lskypow parameter is useful for detecting clouds, but one still would want a)
an independent measure from QC, along with a way to look at the actual image.