UniverseMachine: The Correlation between Galaxy Growth and Dark Matter Halo Assembly from z=0-10

TL;DR

<3-5 sentence high-level summary> UniverseMachine offers a flexible empirical framework that directly connects galaxy growth to dark matter halo assembly by parameterizing star formation rates as functions of halo potential, assembly history, and redshift, and by calibrating against a comprehensive set of multi-wavelength observations from z=0 to z=10. The method employs merger-tree inputs from Bolshoi-Planck, a two-component SFR distribution, and a Bayesian MCMC sampler to produce self-consistent SFHs, SMHM relations, and quenching statistics, including satellites and environmental effects. Key findings reveal a strong link between halo assembly and galaxy growth, a redshift-evolving halo-mass quenching threshold, distinct SMHM relations for centrals versus satellites and star-forming versus quenched populations, and substantial rejuvenation in some systems; Planck cosmology further modulates the inferred SMHM and high-z growth. The public UniverseMachine Data Release (DR1) provides halo/galaxy catalogs, lightcones, and code, enabling broad community use and future empirical-modeling explorations across cosmic time.

Abstract

We present a method to flexibly and self-consistently determine individual galaxies' star formation rates (SFRs) from their host haloes' potential well depths, assembly histories, and redshifts. The method is constrained by galaxies' observed stellar mass functions, SFRs (specific and cosmic), quenched fractions, UV luminosity functions, UV-SM relations, IRX-UV relations, auto- and cross-correlation functions (including quenched and star-forming subsamples), and quenching dependence on environment; each observable is reproduced over the full redshift range available, up to 0<z<10. Key findings include: galaxy assembly correlates strongly with halo assembly; quenching at z>1 correlates strongly with halo mass; quenched fractions at fixed halo mass decrease with increasing redshift; massive quenched galaxies reside in higher-mass haloes than star-forming galaxies at fixed galaxy mass; star-forming and quenched galaxies' star formation histories at fixed mass differ most at z<0.5; satellites have large scatter in quenching timescales after infall, and have modestly higher quenched fractions than central galaxies; Planck cosmologies result in up to 0.3 dex lower stellar mass-halo mass ratios at early times; and, nonetheless, stellar mass-halo mass ratios rise at z>5. Also presented are revised stellar mass-halo mass relations for all, quenched, star-forming, central, and satellite galaxies; the dependence of star formation histories on halo mass, stellar mass, and galaxy SSFR; quenched fractions and quenching timescale distributions for satellites; and predictions for higher-redshift galaxy correlation functions and weak lensing surface densities. The public data release (DR1) includes the massively parallel (>10^5 cores) implementation (the UniverseMachine), the newly compiled and remeasured observational data, derived galaxy formation constraints, and mock catalogues including lightcones.

Figures (39)

• Figure 1: Visual summary of the method for linking galaxy growth to halo growth (§ \ref{['s:methodology']}).
• Figure 2: Left panel: Comparison between observed stellar mass functions (Appendix \ref{['a:smf']}) and the best-fitting model. References for observations are in Table \ref{['t:smf']}. Right panel: Comparison between observed quenched fractions (Appendix \ref{['a:qf']}) and the best-fitting model. Observed quenched fractions are adapted from Bauer13, Moustakas13, and Muzzin13. Notes: almost all data from both panels were used to constrain the best-fitting model. The exceptions are the $z=4-8$ SMFs from Song15, which are shown for comparison only; these were not used in the fitting as the same underlying data is already represented in the $z=4-8$ UVLFs and the UV--SM relations (Fig. \ref{['f:uv_comp']}). The Bolshoi-Planck simulation used is incomplete for low-mass haloes, contributing to an underestimation of the SMF below $10^{7}\;\mathrm{M}_{\odot}$ at $z=0$, a limit which rises smoothly to $10^{8}\;\mathrm{M}_{\odot}$ by $z\sim 8$.
• Figure 5: Comparison between galaxy autocorrelation and cross correlation functions at $z\sim 0$ for the best-fitting model and the observed results rederived from the SDSS in Appendix \ref{['a:cf']}. Notes: all data from all panels were used to constrain the best-fitting model. Observational errors shown are jackknife estimates from the observational sample. Actual fitting used covariance matrices as detailed in Appendix \ref{['a:cf']}.
• Figure 6: Comparison between observed galaxy autocorrelation functions at $z\sim 0.5$Coil17 and the best-fitting model. Notes: all data from this panel were used to constrain the best-fitting model. Redshift errors for PRIMUS were rederived according to Appendix \ref{['a:cf']}.
• Figure 9: Left panel: best-fitting median ratio of stellar mass to peak halo mass ($M_\mathrm{peak}$) as a function of $M_\mathrm{peak}$ and $z$. Right panel: best-fitting median stellar mass as a function of $M_\mathrm{peak}$ and $z$. Error bars in both panels show the 68% confidence interval for the model posterior distribution. Notes: see Figs. \ref{['f:comp_z0']}-\ref{['f:comp_z4']} for a comparison with past results. See Appendix \ref{['a:smhm_fits']} for fitting formulae.