The F22I catalog

This page describes the F22I wide catalog. It gives:

Description of catalogs:

The relevant files:

Some of these files (for reasons of convenience) are available as links from this page. Others are only available on the VIRMOS disk.
  • The main catalog in the SExtractor ASCII_HEADER format. The columns are the same as previous catalogs. The original image was split into four quadrants. Catalogs were generated for each of the four quadrants and then merged. To make the ID numbers unique, 1,2,3 and 4 million (corresponding to the North-West, North-East, South-East and South-West quadrants respectively). was added to the number. The X_IMAGE and Y_IMAGE columns refer to X and Y in the sub-images, not the original image. The other columns are not affected.
  • The same, gzip'ed
  • F22I.xy: X and Y coordinates with respect to the full original image. The format is 3 columns ID X_IMAGE Y_IMAGE
  • The masks, one per line (RA_1. Dec_1. RA_2. Dec_2 .in decimal degrees) Objects inside one of these boxes  (ie. RA_1<RA<RA_2 and Dec_1<Dec<Dec_2) are corrupted in some way, be it contamination from a bright source, satellite trail, bad fringing or another source.
  • F22Igwyn.flags: Flags for the objects. One line per object in, 6 columns: 
    • ID (matches first column of
    • overall flag = (star flag)+2*(box flag)+4*(mask flag)+8*(bright flag)
    • star flag: 1 if the object is point source
    • box flag: 1 if the object is outside the image boundaries
    • mask flag: 1 if the object is inside a box (i.e. contaminated)
    • bright flag: 1 if the object is brighter than IAB  = 17.5 (MAG_AUTO),
    Following this scheme, usable galaxies have an overall flag of 0, usable stars a flag of 1. Non usuable objects have a flag of 2 or greater. Star-galaxy separation was non-trivial due to seeing variations accross the field. See the full discussion. 
  • F22Ihjmcc.flags: Flags for the objects generated in the same manner as described by Henry McCracken on his webpage with the exception that I have used my adaptive method of star-galaxy separation.
  • F22I.fits (5.1Gb): The whole image 
  • F22I.weight.fits: The whole weight image
  • Sub-images: The image was split into quadrants, for easier viewing and extracting sources (note that East = direction of increasing RA)
    • F22Inw.fits(1.3Gb): North West quadrant
    • F22Inwb4.fits (83Mb): same, binned 4x4
    • F22Inw.weight.fits: weight image
    • F22Inw.mask.fits: mask image (1 in masks, 0 outside)
    • F22Isw.fits(1.4Gb): South West quadrant
    • F22Iswb4.fits (88Mb): the same, binned 4x4
    • F22Isw.weight.fits: weight image
    • F22Isw.mask.fits: mask image (1 in masks, 0 outside)
    • F22Ine.fits(1.2Gb): North East quadrant
    • F22Ineb4.fits (73Mb): the same, binned 4x4
    • F22Ine.weight.fits: weight image
    • F22Ine.mask.fits: mask image (1 in masks, 0 outside)
    • F22Ise.fits(1.2Gb): South East quadrant
    • F22Iseb4.fits (77Mb): the same, binned 4x4
    • F22Ise.weight.fits: weight image
    • F22Ise.mask.fits: mask image (1 in masks, 0 outside)
  • F22I.log: Detailed log of how the images and catalogs were created.

Sky area:

A rectangle between:
RA 22:13:48 22:21:52
Dec -00:31:45 +01:18:00
Or in decimal degrees:
RA 333.450000 335.466666
Dec -0.529166 +1.300000

With allowances for the masked areas, the total area is:  3.345 square degrees.


The following areas were masked:
  • Areas near saturated stars. The brighter the star, the larger the contaminated area. This was done automatically, followed by manual touch ups. Bleeding columns that made it into the final stack were also masked.
  • Satellite trails: Despite the fact that the images were combined by taking the median of the input frames, some satellite trails remain.
  • Areas not observed: One region near the centre of the field was observed under poor conditions. These data were ignored. Subsequent observations cover this area, but these data have not yet been incorporated, due to concerns about proper calibration. Also, the pointings at south-west and south-east corners were centred slightly too far south to overlap completely with the rest of the data
  • The seams at the joins of the four quadrants were masked with 10'' margins.
  • Areas of bad fringing. 

    The figure at right shows  the layout of the masks in the file F22I.boxes. The large box near the centre was not observed. The patch at lower left is an extremely bright star which causes an unacceptable number of false detections. 

    The boundaries of the field are shown in red.The boxes shown in blue may be ignored, as they lie outside the boundaries.

    Notes on reduction procedures:

    This section provides a brief reference for the data reduction procedure. 
  • Exposures used: A complete list of the CFHT CCD frames used in the preparation of this catalog can be found in the file: F22I.explist. The layout of the observations is shown below.
  • Detrending: The bias, dark current, flat fielding and fringe removal of the frames was performed by: Mssrs: Foucaud, Dantel and Mellier.
  • Astrometric calibration: This was done using the AstroGwyn package. Objects in the data frames were matched with stars in the USNO astrometric catalog. Using these matched objects (typically 40-50 per CCD), a second-order distortion correction is computed. This correction is applied to each CCD. Then all the objects in each CCD frame are matched to all the objects in all the other CCD frames. Objects common to both frames are identified and a master list is compiled. This list is then  used to compute a second distortion correction. The procedure is iterated. The final astrometric solution has an external error of about 0.3''. The internal error of the solution between any two raw frames is typically  0.16'' with a worst case of 0.3'' occuring 0.4% of the time.
  • Photometric calibration: This was done using the PhotoGwyn package. Several images were identified has having been taken under photometric conditions. The zero-points were assigned to these images based on the information on the VIRMOS photometric calibration webpage. with the exception of the zero-point for the Oct00-3 run which was found to be 26.260 for the I band, rather than 26.0078 as indicated on that web page. Zero-points were assigned to the images adjacent to the directly calibrated images by comparing the instrumental magnitudes of objects common to both images. The details of this procedure are shown below.  The resulting photometric accuracy in the zero-point is 0.06 magnitudes. The headers of all the images, which contain all the necessary astrometric and photometric calibration information, can be found here. 
  • Image combining: The program SWarp was used to combine the images. The parameters used are in the file F22I.swarp.
  • Image splitting: The resulting image was too large (5.2Gb) to viewed or otherwise processed. It was therefore split into 4 sections using the program toFITS2d.
  • Catalog extraction: Catalogs was extracted from the four sub-images using SExtractor with the following parameter files: 
  • The four catalogs were then merged. This means the some the SExtractor parameters (X_IMAGE, Y_IMAGE) will not be valid. 
    The figure at right shows the layout of the observations. They are labelled as follows:

    ^  ^ ^^^
    |  | |||_ 2-digit pointing
    |  | ||__ the letter p
    |  | |___ date: sub run number
    |  |_____ date: 2-digit year
    |________ date: 3-letter month

    Note that the outlines of the observations are determined directly the astrometrically calibrated data frames.

    The figure at right shows how the photometric calibration was done.

    Lower left: layout of the 22hr pointings, outlined in red and labelled in black.

    Upper left: layout of the 22hr pointings. Photometric exposures are in heavy black and labelled "0". Exposures that "touch" (more than 150 common objects) photometric exposures are in red and labelled "1". Exposures that "touch" level 1 exposures are outlined in blue and
    labelled "2".

    Big right panel: For every exposure, a zero-point was calculated under the (usually erroneous) assumption that it was taken under photometric conditions. This zero-point was calculated using  using the airmass and exposure time from the header and the zero-point for the run. This zero-point is known as the "nominal" or "original" zero-point.

    Some exposures were flagged as being genuinely photometric. These ones are shown in black. For each photometric exposure, zero-points for all the other exposures by comparing magnitudes in the overlap regions. The difference between this "calculated" zero-point and the "nominal" one is plotted along the x-axis.
    The exposures are labelled by number on the left and by the UTC date and HST time are shown on the right. The colour-coding is the same as the upper left panel: black for directly photometric, red for 1 step removed and blue for 2 steps removed.

    The final zero-point for each exposure is the average of the all the "calculated" zero-points for that exposure.

    Data tests:
    The figure at right shows the I band galaxy number counts for the F22I together with several data from the literature.