Each FITS image and catalogue contains a keyword in the primary header that uniquely identifies the version of the product. This keyword is named CASUVERS and takes a value like the following example:
CASUVERS= 'vircam version 1.3' / CASU Release Version Number
Version 1.3 (27/03/2013)
All data have been tagged as version 1.3. Note that nights before 20101101 were still version 1.0 and those before 20110501 were still version 1.1, both sets now include the cumulative changes up to version 1.3. The version 1.3 changes affect all tile catalogues up to 20111231, catalogues after that date are unaffected apart from requiring updating with the latest ESO grades. All processed images are unchanged apart from updated magnitude zero-points. The main changes are:
- A bug involving how the aperture 2 correction was calculated is now fixed and tile catalogues have now been regrouted to include this.
- Prior to regrouting all the stacked pawprint photometric zero-points were recomputed using the latest version of the photometry software.
- Post regrouting all the tile photometric zero-points have also been updated.
- ESO grades have been updated and they should now agree with those supplied by ESO to the PIs directly. This affects the keywords ESOGRADE and OBSTATUS for all data products. Follow the ESO Grades page for more information on their values.
- All tile catalogues have been re grouted taking into account both detector level magnitude zero points variations (tiles before 20101101 did not have those applied) and atmospheric seeing variations.
- Note that WCS coefficients for PV2_3 and PV2_5 were changed from 42.0, -10000.0 pre-20101130 to 44.0, -10300.0 post-20101201. The pre-20101130 astrometry was not updated.
- The internal ZPN->TAN definition bug that affected tile products was fixed August 2012. All products post-2012801 use the corrected ZPN ->TAN transformation. Earlier tile products remain affected at the ~100mas level by this bug. All pawprint products are unaffected.
Version 1.2 (01/05/2011)
Data obtained after the mirror coating intervention has been tagged as version 1.2, i.e., data obtained after 20110501. The changes are:
- Added keyword NUMSTDS which stores number of 2MASS objects in the field of view of each chip or tile potentially usable for the astrometric solution.
- Tiles now inherit the nightly photometric information from pawprints (keywords affected NIGHTNUM, NIGHTZPT, NIGHTZRR).
- Provenance keyword values in tiles do not now have "tmp_" in the provenance information for the component filenames.
- Added a GROUTED keyword to catalogues generated from tiles that have been grouted.
- RADECSYS keyword changes. Primary extension remains FK5 while secondary extensions have the value ICRS added.
- Added new keywords to the tile primary header: PAWMAGZP, PAWELLPT and PAWSEENG which summarize the median values of the zero point, ellipticity and seeing from originating pawprint images to aid in QC procedures.
- The magnitude zero point error estimate for tiles is now calculated from the zero-point variation in the component pawprint images
Version 1.1 (01/03/2011)
All data from 20091015 to 20101030 have been updated (see below) and all data since 20101101 are processed as v1.1
- re-running the classifier on all pawprint catalogues to give improved estimates of aperture corrections and detector saturation levels
- re-running the photometry on all pawprint catalogues with updated colour equations, and extinction coeffs and more careful selection of 2MASS stars in high extinction regions
- generating monthly detector zero-point updates (and illumination corrections); detector level differences will be now routinely included for all VISTA data
- FITS header updates on all pawprint catalogues and images
- running the grouting code on the tile catalogues; this fixes both PSF changes via the various apercors and magzpt changes at either detector level or whole pawprint [note however that grouted tile catalogues pre-20101101 only have the effect of PSF changes fixed and NOT the detector-level magzpt changes - regrouting of these is scheduled for v1.3]
- computation of differential MJD column for all detected objects and output of optional map of differential MJD across tile
- tile catalogue re-classification and re-photometry
- FITS header updates on all tile images
- Updated ESO grades (see here for an up-to-date explanation)
To summarise: no change to images, just header updates; some changes to catalogue table contents and headers
Version 1.0 (21/07/2010)
- First version to contain tile images and catalogues. Note that tile images are Rice compressed. In order to uncompress them to their original state use funpack. For more information about tiles please read the following page.
- Object detection threshold is now 1.25 sigma above background (instead of 1.5 in previous versions) for all catalogues.
- Sky subtraction changes:
- For VVV an attempt is made to pair each tile with another tile of the same filter which were observed within 30 minutes of each other. If it can be done then the sky image is made from the 24 exposures of the two VVV tiles. If it can't be done, then it's just done with the 12 exposures of the single VVV tile.
- For UltraVISTA, the skies are formed using a static object mask which was supplied by the UltraVISTA team.
- All other projects were sky corrected in the way they were previously.
- Two new keywords have been added to image products to reflect the OB status and QC grade as reported by ESO: OBSTATUS and ESOGRADE. The former can be 'Completed', 'Executed' (which means it has to be repeated) or 'Aborted'. The latter is 'A' if fully under constrains, 'B' if mostly (90%) under constrains, 'C' if out of constraints and 'R' rejected. Note that the requested constraints are in the header of the images in the 'ESO OBS' keywords (on data obtained after 20100208). Also note that standard stars observations and non-public survey observations are not graded. A more detailed description is available here.
- Note also that some OBs are aborted and then restarted part way through, and some assigned ESO grades are incorrect.
- The pipeline processing attempts to identify aborted and restarted OBs and uses the "good" observations to produce a "complete" processed OB. This processed OB inherits the ESO grade from the first component and does not necessarily reflect the correct status of the processed OB. We therefore recommend use of the ESO grades only as a first pass indicator of status followed by use of the derived QC information in the files to check if the science requirements are likely to be met.
- Fixed stacking rejection bug which affects pawprints with njitter=2.
Version 0.9 (12/05/2010)
- On the night of 20091119 the chips 5 and 6 were swapped in order to investigate the problem with chip 6 channel 14. We have swapped back these two extensions to make them consistent with the rest of the data.
- Several pawprints on November 2009 were not sky subtracted. This affects VIDEO and VMC and it is corrected now. The list of affected frames is available here.
Version 0.8 (19/03/2010)
- Several internal consistency checks performed on version 0.7
- Fixed Half light radius calculation
- Fixed rejection algorithm on stacking software
- Fixed propagation of astrometry to single pawprint images
Version 0.7 (10/02/2010 - internal release)
- Use covariance matrix in order to calculate equivalent rms sky noise. The magnitude limits and magnitude errors make use of this and are now correct.
- Fixed an astrometry bug on pawprint stacked images.
- Added average confidence in the default aperture for each object in the catalogue (Av_conf table parameter #58).
- The Hall flux and radius has been replaced by the half-light flux and radius (parameters #48 and #53).
- Added CASUVERS keyword to identify the version number of the product.