Home | Overview | Browser | Access | Login | Cookbook | nonSurvey

Problem with astrometry for GPS in DR8

The issue lies with the WCS header of the catalogue that was used to calculate the RA and Decs of the objects in gpsDetection and gpsSource. The GPS re-processing for DR8 saw the updating of the PV2_3 value depending on the waveband. Previously all wavebands had PV2_3 set to -50. The revised values being: J band, PV2_3 = -46 and K band, PV2_3 = -60. Although the headers of the affected files contain the revised values the rest of the WCS was not updated (i.e. corresponds to PV2_3 = -50). Consequently positions in the affected frames are calculated incorrectly and get worse the further the position is away from the centre of the pawprint. At the far corners of a pawprint the positions are out by 0.4-0.5 arcsec in J and 1.0-1.1 arcsec in K

This problem occurs in data taken in 06A and earlier and affects some 456 multiframes (1809 out of 75,296 detectors) in DR8 which is 2.5%

Below is link to a file listing the affcetd files. List of affected frames (multiframeIDs, framesetIDs and file names)

The problems affects RA/Decs store in both gpsDetection and gpsSource.

This issue only affects DR8, the positions from the equivalent catalogues in DR7 are OK.

The files will be reprocessed or excluded from the next release.