A Halo Model of Galaxy Colors and Clustering in the SDSS

TL;DR

This paper develops a halo model linking galaxy colors and clustering in SDSS, showing that color-environment correlations are mainly driven by halo mass, and provides a method to generate realistic mock galaxy catalogs.

Contribution

It introduces a color-dependent galaxy clustering model based on halo mass, with assumptions on color bimodality and satellite red fraction, without extra free parameters.

Findings

Model agrees well with SDSS color distribution and clustering data.

Color-environment correlation primarily driven by halo mass.

Provides a framework for realistic galaxy mock catalogs.

Abstract

Successful halo-model descriptions of the luminosity dependence of clustering distinguish between the central galaxy in a halo and all the others (satellites). To include colors, we provide a prescription for how the color-magnitude relation of centrals and satellites depends on halo mass. This follows from two assumptions: (i) the bimodality of the color distribution at fixed luminosity is independent of halo mass, and (ii) the fraction of satellite galaxies which populate the red sequence increases with luminosity. We show that these two assumptions allow one to build a model of how galaxy clustering depends on color without any additional free parameters than those required to model the luminosity dependence of galaxy clustering. We then show that the resulting model is in good agreement with the distribution and clustering of colors in the SDSS, both by comparing the predicted correlation functions of red and blue galaxies with measurements, and by comparing the predicted color mark correlation function with the measured one. Mark correlation functions are powerful tools for identifying and quantifying correlations between galaxy properties and their environments: our results indicate that the correlation between halo mass and environment is the primary driver for correlations between galaxy colors and the environment; additional correlations associated with halo `assembly bias' are relatively small. Our approach shows explicitly how to construct mock catalogs which include both luminosities {\em and} colors -- thus providing realistic training sets for, e.g., galaxy cluster finding algorithms. Our prescription is the first step towards incorporating the entire spectral energy distribution into the halo model approach.