GCM Zeropoints in SVA1

Do the zeropoints vary across the exposures more than we expect? (And what will this do to our photometry?)


The Test

On photometric nights at low airmass, we do not expect the zeropoints to vary significantly even over the large DECam field-of-fiew. (I think...) Dave Burke's (still in development) "Forward Global Calibration Module" (FGCM) has very small variations for photometric nights. But there are some significant differences in the patters from FGCM (Beta) and GCM as run on SVA1.

I took exposures over SPTE, with T_EFF > 0.5. I focused on exposures that were within 0.2 magnitudes of the median zeropoint for each band in SVA1. This selects relatively "good" exposures.

Then, for each exposure, I compute the range in MAG_ZERO from the deepest to shallowest in each exposure. This is always > 1%, and typically 2-3%. Thus, GCM is saying there's at least a 2% variation across the field...does this seem right?