Spatial variation of zero-point systematics
Issue raised Jun 08 2017, 06:04 UTC; last modified Jun 19 2017, 02:42 UTC
We have identified spatially varying systematics in the SkyMapper DR1 photometric zero-points (ZPs), which primarily affect the i-band and z-band magnitudes.
The reliance of DR1 photometry on ZPs determined from sources in the APASS catalogue means that presence and pattern of systematics in that catalogue are reflected in the properties of DR1. For example, the APASS team has indicated that the quality of their photometry and astrometry degrades towards the corners of their fields, and we have found that this can result in ZP residuals across a given SkyMapper image. The DR1 data processing fits a bi-linear ZP gradient across each image, but the APASS variations arise on different spatial scales, which the fitted gradient can in some instances exaggerate rather than mitigate. The effect is particularly pronounced for the DR1 photometry in i-band and z-band (both of which are primarily derived from the APASS i-band)
The plot below shows the spatial distribution of differences between the i-band magnitude in DR1 and that predicted for stars in APASS that were considered in the ZP fit (about 6 million sources).
Among the APASS stars that are well-measured in DR1, 65% have i-band offsets < 0.05mag, 88% < 0.1mag, and 98% < 0.2mag. However, some spatial regions can reach +/- 0.4mag offsets. The corresponding percentages for g-band and r-band are (89%, 98%, 99%) and (85%, 97%, 99%), respectively.
We are currently investigating options to address this issue.