Implement environment-independent red sequence SEDs

[aphearin]

In the current cosmoDC2 model, for galaxies of the same stellar mass the distributions of restframe (g-r) and (r-i) colors of the two populations are identical:

• red sequence cluster galaxies
• red sequence field galaxies

These two populations also have identical spectral normalization <Mr|M> in the restframe. This conditioning on restframe color/flux enforces that these two populations will have similar SEDs, since fixing `M, Mr, Mg, Mi` strongly restricts galaxy SED space. However, in the current model the SED distributions of cluster vs. field red sequence galaxies are not identical; in observer-frame bands, red sequence cluster galaxies will be redder than red sequence galaxies in the field.

Note that this issue is distinct from #6, which will be resolved prior to releasing cosmoDC2 simply by smoothing out the hard-edge Mhost-dependence of red sequence colors. Here we are not just requiring that the distributions be smooth; here the cluster WG has requested that the following equality hold:

That is, for galaxies of the same stellar mass, the statistical distribution of any SED characteristic should be independent of host halo mass, not merely smooth and gradual with halo mass. At the present time, this issue is marked post-cosmoDC2 as it is not critical to implement this prior to the image simulation.

[erykoff]

I chatted with Eve (I think) last week about this, and this is not precisely true. The mass-kink is visible in absolute color as well as observed color. See here for v4.14_test for centrals of a slice at z=0.3: The following shows the absolute g-r color for centrals as a function of redshift. Density plot is 9e12<M<3e13, points are M>3e13: So this might be a selection effect, though I'm not exactly sure how, but something is happening so that the absolute colors also have a sharp transition.