One Galaxy Sample to Rule Them All: Halo Occupation Distribution Modeling of DES Year 3 Source Galaxies

TL;DR

The paper develops a halo occupation distribution (HOD) model for DES Year 3 source galaxies to enable lens-equal-source 3x2 analyses on small scales. By matching COSMOS/UltraVISTA photometry to UniverseMachine galaxies and applying DES-Y3 source selection, the authors reveal significant central incompleteness and redshift evolution in the HOD, plus evidence for galaxy assembly bias. They introduce a modified analytic HOD with f_cen(M) to describe central incompleteness and a standard satellite term, and provide priors derived from 1000 realizations that account for photometry and size uncertainties as well as UniverseMachine model variations. An emulator based on Gaussian-process regression delivers 1–3% accurate HOD predictions as a function of redshift and observational errors, enabling forward-modeling of the source population in cosmological analyses. Together, these tools allow robust inclusion of small-scale information in lens-equal-source studies while mitigating biases from the galaxy–halo connection and photometric systematics, advancing DES-like cosmology with improved self-calibration of photo-z and related uncertainties.

Abstract

For the joint analysis of second-order weak lensing and galaxy clustering statistics, so-called $3{\times}2$ analyses, the selection and characterization of optimal galaxy samples is a major area of research. One promising choice is to use the same galaxy sample as lenses and sources, which reduces the systematics parameter space that describes uncertainties related to galaxy samples. Such a ``lens-equal-source'' analysis significantly improves self-calibration of photo-z systematics leading to improved cosmological constraints. With the aim to enable a lens-equal-source analysis on small scales we investigate the halo-galaxy connection of DES-Y3 source galaxies. We develop a technique to construct mock source galaxy populations by matching COSMOS/UltraVISTA photometry onto {\sc{UniverseMachine}} galaxies. These mocks predict a source halo occupation distribution (HOD) that exhibits significant redshift evolution, non-trivial central incompleteness and galaxy assembly bias. We produce multiple realizations of mock source galaxies drawn from the {\sc{UniverseMachine}} posterior with added uncertainties in measured DES photometry and galaxy shapes. We fit a modified HOD formalism to these realizations to produce priors on the galaxy-halo connection for cosmological analyses. We additionally train an emulator that predicts this HOD to $\sim2\%$ accuracy from redshift $z = 0.1 - 1.3$ that models the dependence of this HOD on 1) observational uncertainties in galaxy size and photometry, and 2) uncertainties in the {\sc{UniverseMachine}} predictions.

Figures (7)

• Figure 1: Distribution of specific-SFR (sSFR) in the SMDPL UniverseMachine mock galaxy catalog at $z=0.25$ (blue), $z = 0.50$ (green), $z = 0.75$ (yellow), and $z=1.0$ (red). The dotted vertical line at $\mathrm{sSFR} = 10^{-11} \, \mathrm{yr}^{-1}$ located at a local minimum in the distribution divides galaxies into star-forming and quiescent. We observe that the location of this minimum is insensitive to redshift.
• Figure 2: Color-color diagrams for COSMOS/UltraVISTA galaxies binned by redshift. Black lines divide the the galaxies into quiescent (upper region) and star-forming (lower region) population. A subsample of the galaxies are plotted as points that are colored red (quiescent, $\mathrm{sSFR} < 10^{-11} \, \mathrm{yr}^{-1}$) or blue (star-forming, $\mathrm{sSFR} > 10^{-11} \, \mathrm{yr}^{-1}$) based on the COSMOS/UltraVista estimated sSFR. We observe acceptable agreement between the quiescent/star-forming separation using color information and estimated sSFR.
• Figure 3: Redshift dependent size-mass relations from Mowla_et_al_2019a. There are separate relations for quiescent (solid-line) and star-forming (dashed-line) galaxies that evolve significantly with redshift, and cross over each other at high stellar mass.
• Figure 4: The HOD of UniverseMachine-COSMOS/UltraVISTA matched source galaxies for central (red), satellite (blue) and all (black) galaxies. The left-hand panel shows how the source HOD varies with redshift. We observe significant redshift evolution in the range $z = 0.25 - 1.00$, particularly in the characterstic mass scales to host centrals and satellites. The right-hand panel shows the effect of relaxing selection criteria at $z = 1.00$. The solid line shows the HOD obtained with all selection criteria applied, the dot-dashed curves show the impact of relaxing the magnitude cuts, the dashed shows the impact of relaxing color cuts, and the dotted shows the impact of relaxing size cuts. We see that the significant central-incompleteness we observe is driven by cuts on galaxy size (see text).
• Figure 5: Best fitting HOD models for our mock source galaxies at $z = 0.25$ (top-left), $z = 0.50$ (top-right), $z = 0.75$ (bottom-left), and $z = 1.00$ (bottom right). We see that the satellite term is very well fit by a standard power law, while the modified central term captures overall changes in incompleteness as a function of mass.