Gal-G11: Composite Cluster Red Sequence Tests

Composite red sequences of galaxies in clusters of known redshift have a measurable intrinsic scatter of <0.05 mag RMS in g-r and <0.03 mag RMS in r-i and i-z.

This test makes use of the existing composite red sequence calibration code for the redMaPPer cluster finder. Given a set of clusters with spectroscopic redshifts, the calibration measures the color, slope, and intrinsic scatter of the red sequence over a wide area. These tests were used to show that the DC6B color measurements were introducing an additional color scatter of ~6% (g-r) which would make it not possible to detect clusters with any reasonable purity or completeness.

Due to the availability of pre-survey data, this test will be the most interesting at 0.2<z<0.7 in the Stripe 82 overlap region. The number of cluster spectra (for clusters with >~10 red galaxies) at higher redshifts is very limited in the SV region.


  • Full-depth coadded images over >20 deg^2 of SDSS Stripe 82 with DETMODEL colors AND/OR
  • Individual images over >20 deg^2 of SDSS Stripe 82 with DETMODEL colors.
  • Mangle masks useful but not required.
  • Set of cluster spectra, including XMM-LSS clusters in overlap region and Stripe 82 redMaPPer clusters (0.2<z<0.7).
  • To be useful, we require >~ 5 clusters per 0.05 redshift bin (and more is better). This requirement is met with >20 deg^2 of overlap with Stripe 82.


  • Run redMaPPer red sequence calibration to measure the color, slope, and scatter of the composite red sequence for the clusters with spectra.
  • The intrinsic scatter of the red sequence in g-r should be ~0.05 at z<0.4, and the scatter in r-i and i-z should be 0.02-0.03 at z<0.7.


If the red sequence scatter is larger than cluster finding is going to be impacted, perhaps significantly. The source of the problem may be the color measurement (DETMODEL magnitudes) or field-to-field variations, as the composite red sequence is implicitly a wide-area test of the photometry.