search

LSSTDESC / cosmodc2

Python package creating the cosmoDC2 synthetic galaxy catalog for LSST-DESC
Other
7 stars 1 forks source link

Remodel boost to cluster mass-richness relation #70

Closed aphearin closed 5 years ago

aphearin commented 5 years ago

This commit addresses GitHub Issue #69, which identified a small population of cluster satellites with identical gri restframe magnitudes. The cause of this problem is the way we boosted the mass-richness relation by resampling existing galaxies in the mock; the way this resampling was implemented led to occasional repeated selection of the same objects.

It was further shown that the spatial positions of the galaxies with repeated magnitudes were unexpectedly concentrated. The source of this unphysically large concentration was the same source as the repeated magnitudes. Occasionally, the selected objects were originally located in a smaller-mass halo. So when we distributed the new objects according to an NFW profile, the spatial extent of the profile was being set by the mass of the original halo.

This PR resolves this issue by remodeling this population with the same machinery used to model the original cluster satellites, entirely doing away with the resampling method that caused the problems.

CC @evevkovacs

aphearin commented 5 years ago
  1. Good catch! I just pushed up a commit fixing this error
  2. There's no attempt to recover anything - I just chose something reasonable based on DM-only simulations
  3. That's right, we only boost red sequence membership, mostly because it was easier to implement.