Technical feedback from Mike Irwin on 27 July.
Mike described a number of techincal issues relevant to the GPS at the Survey Heads meeting in Durham. The Powerpoint presentation is at http://www.ast.cam.ac.uk/vdfs/docs/reports/sv/index.html
Notable points (including responses to specific questions) are:
1) The offset skies included in the GPS SV and the first tranche of GPS proper observing were not really needed, since many sky frames were constructed from median filtered frames and the sky frame closest in time to each data set was used. In addition the non-uniform sky background caused by a substance on the inside of the field lens will be removed before the 2005B WFCAM run. Hence we are unlikely to need specific GPS offset skies in future.
Individual dust spots on the field lens are sufficiently out of focus to appear as doughnuts in raw data.
2) Data taken in thin cirrus conditions have not yet been checked by CASU to see whether they are useful for relative photometry.
3) PSF fitting photometry will be available by end of 2005.
4) Photometric errors for the aperture photometry will be available by early autumn 2005 (response by Nigel).
5) The Zero Points in the headers of frames downloaded from the archive are meaningful and can be used for our own photometry.
6) Photometric system. Archived Data are in the WFCAM system, which is very similar to the MKO system. Zero points are boot strapped from 2MASS at present, but will later be calculated from UKIDSS calibration data. Photometric transformations for sources with colours influenced by extinction should in principle be different from high latitude stars where the colour just comes from the photosphere, but Mike didn't think the difference was large. I will chase this up.
7) Linearity of the arrays is essentially "perfect".
8) Read noise is still 20 electrons (after removing the "curtaining", pattern in each quadrant).
9) Gains are 4.5 to 5.5 e-/DN.
10) Flat fields are stable and wavelength independent but there is significant spatial variation in sensitivity across each array at the +/- 20% level.
11) The bright and dark patterns near each saturated star are caused by cross-talk between adjacent read out channels in each array quadrant. The separation is always 128 pixels along the readout axis. The spurious images appear as the spatial derivative of the saturated star in the read out direction. These cross talk features will be removed from the data as well as possible in future versions of the pipeline.
- 09 Aug 2005