Project

General

Profile

Gal-G-1-5

The main goals of this set of tests is to check the accuracy of DESDM galaxy magnitudes and colors, internally and with reference to truth fields of varying source density and seeing conditions.

Prerequisites

  1. Imager is functional and focussed, all filters are available.
  2. Photometric nights (preferably all data should be taken in the same night), dark time.
  3. Overscan/bias/dome flattening is done and catalog with galaxy coordinates, plus magnitudes [mag_auto and mag_model (for total magnitudes) and mag_detmodel (for colors)] and their errors is available via DESDM. In particular, detmodel magnitudes are supposed to be seeing-independent and mag_auto is not strongly seeing dependent.

Procedure

Dataset:
~3 non-dithered observations of the fields specified bellow in all bands, with seeing varying between observations in the range ~ 0.7-1.5 arcsec with steps of ~ 0.5 arcsec.

Proposed observing procedure:
  1. Run a SISPI script to take one set of images near the start night: 1 exposure in each filter, exptime ~100 seconds, visit both fields (total estimated time for this: ~20 minutes)
  2. Note down the seeing at the time.
  3. Proceed with other observations planned for the night
  4. When seeing value changes by more than ~0.4’’ from any set previously done, repeat the script.
  5. Stop when you have completed the full range between ~0.7 and ~1.5 arcsec, e.g.:
    <0.7, 0.7-1.1, 1.1-1.5, >1.5
Fields:
  1. Cluster core TBD. Suggestion: Bullet Cluster (RA=06:59, Dec=-55:57)
  2. VVDS-Deep 02hr field (RA=02:26, Dec=-04:30, galactic latitude b=-58 deg, SN deep field and photoz field),
  3. VVDS-Deep CDFS field (RA=03:32, Dec=-27:48, SN deep field and photoz field),
  4. High stellar density: Open cluster (TBD. Maybe the M46/M47 pair, at b=-4, but need to check if there is truth table for this target.)

Note:
It is important that these exposures get flag appropriately flagged/tagged in the logbook or comments header tag (or whatever we think is appropriate) so that later we can easily find the dataset we need for this test.

Analysis method:
  1. Match the repeated exposures in RA and DEC and measure the relative degradation in completeness, magnitude and color errors as a function of seeing.
  2. Select a subsample of deblended objects only and repeat the step above.
  3. Compare the results for the various fields. (Maybe break in subfields to test the results as a function of number of # of galaxies per sq-arcmin)
  4. Scale the results to full depth data!
  5. Basic plots:
    1. mag_dispersion - mag_err vs. seeing
    2. color_dispersion - color_err vs. seeing
    3. relative completeness vs. mag (for each seeing value)
    4. relative purity vs. mag (for each seeing value)
    5. mag_dispersion vs. mag (for each seeing value)
    6. color_dispersion vs. mag (for each seeing value)

Veredict

We pass the Gal-G tests if:

  1. magnitudes and colors are biased (vs. truth or vs. 0.9’’ seeing) by <0.01 mag at each seeing value. Assuming that the seeing is just one component of the 0.02 mag photometry error budget at each seeing value. (Gal-G-1)
  2. color dispersion < 2% (CR-3)
  3. mag errors for deblended objects brighter than 22nd i-band mag < 0.1 mag @99% CL (CR-9) (Gal-G-5)
  4. completeness variation between the fields at densities between 10 and 100 galaxies per square arcmin < 5% (CR-6) (Gal-G-2) (Gal-G-3)
  5. purity variation between b=20 and b=90 < 5% (CR-7) (Gal-G-4)
  6. rejection rate of stars at magnitudes up to i = 21 > 99% (CR-8)
  7. limiting mags are > 22.8, 23.4, 24.5, 23.3 in g,r,i,z respectively (CR-2)

Consequences

??

Results

(Dec. 12, 2012) DES docdb items 6720 and 6701 have comparisons of DECam catalogs with SDSS catalogs.

Other notes

??