CALMON version 3.0 - Proposed Change to Remove Aberrant Standards
CALMON uses aperture magnitudes in computing the nightly zero point. Due to the nature of aperture photometry, some standards are affected by spurious artifacts such as radiation hits or dead pixels. This can cause an entire calibration scan to
fail the dispersion test thus affecting the final zero point solution. The PSF based calibration for the same night will likely derive a different and usually better calibration solution. This has caused QA analysts to override the default selection of the aperture based solution and adopt the PSF based zero point solution for a given night. Having non uniformity in the derivation of the final accepted nightly zero point solution is not ideal. This document proposes a way to identify and eliminate the aberrant standards prior to deriving the zero point solution thereby improving the quality of the aperture based nightly zero point solutions and consistency of calibration throughout the mission.
One method to identify anomalous standards within scans is to compare the PSF magnitude to the aperture magnitude and reject standards when the aperture magnitude exceeds the PSF magnitude in any band by a preset threshold.
This may be the best and quickest method of identifying aberrant standards.
Another method to identify anomalous standards within scans is to examine
the error distributions. CALMON computes the following dispersion
values:
1. Dispersion of all standards within a single scan.
2. Dispersion of each standard within a calibration set of scans (6).
3. Dispersion of all standards in all 6 calibration scans.
I propose that the dispersion of all standards within a single scan be
used in identifying aberrant standards. Not only is this a good logical
point within CALMON to make such determination but also calibration fields
have several standards providing a reasonable population sample for
this statistical measure. Most fields
have now over 20 secondary standards with a handful having between 11 - 20.
The dispersion of each standard within a calibration set would only
provide 6 data points and will not identify secondaries that are
aberrant for the entire calibration set.
The dispersion of all standards in all
6 calibration scans is computed further into CALMON and can provide
another verification as to the quality of the standards.
The dispersion analysis would be a more complicated logic change to CALMON. I suspect the
psf vs aperture comparison would reject the majority of the aberrant standards
for the aperture solution. It needs to be determined if the additional dispersion
check would result in any significant change in the final solution.
Comments and feedback: Sherry Wheelock
email: slw@ipac.caltech.edu
Last update: 10 September 2000
slw@ipac.caltech.edu