Bad Photometry Maps¶
Working from Regions of Bad Photometry, I have made healpix maps.
Bad Fraction Maps¶
I have pixelized SVA1 with healpix, nside=1024, nest ordering scheme (to match the mangle masks). In each pixel I calculate the number of objects brighter than i<23 (MAG_AUTO), that have "crazy colors". A "non crazy" color is defined as -1 < (DETMODEL_G - DETMODEL_R) < 4 and -1 < (DETMODEL_I - DETMODEL_Z) < 4. Detmodel colors were used because they have much better behavior in regions where one bad is merely very shallow rather than corrupted by an image artifact (streak, scattered light, etc.) (A brief note on this aspect of DETMODEL colors vs AUTO colors is coming soon.)
Visual inspection shows that this is a reliable indicator of concentrations of bad objects.
Only regions with at least 30 galaxies in the pixel are considered. This number was calculated based on it being a 3sigma deviation to the low side of the number of galaxies with i<23 in all pixels in SVA1. Thus, pixels with fewer galaxies are clearly very shallow and should be avoided.
Overall, about 0.7% of the area has a bad fraction of >0.2, and 2.3% of the area has a bad fraction of >0.1.
Although many satellite streaks are narrower than the size of the pixels here, for the sake of expediency I think we can be conservative and mask out extra area.
Note the large bad area near R Doradus in the SW. This Mira variable is very red, and is J=-2.3, the second brightest star in the sky after Betelgeuse. (Thanks to Allistair for this tidbit!). I propose we mask out a region of area ~ 10 deg^2 around this area.
(to be uploaded)