New empirical constraints on the cosmological evolution of gas and stars in galaxies

Abstract: We combine the latest observationally motivated constraints on stellar properties in dark matter haloes, along with data-driven predictions for the atomic (HI) and molecular (H$2$) gas evolution in galaxies, to derive empirical relationships between the build-up of galactic components and their evolution over cosmic time. At high redshift ($z \gtrsim 4$), the frameworks imply that galaxies acquire their cold gas (both atomic and molecular) mostly by accretion, with the fraction of cold gas reaching about 20% of the cosmic baryon fraction. We infer a strong dependence of the star formation rate on the H$_2$ mass, suggesting a near-universal depletion timescale of 0.1-1 Gyr in Milky Way sized haloes (of masses $10^{12} \ M{\odot}$ at $z = 0$). There is also evidence for a near-universality of the Kennicutt-Schmidt relation across redshifts, with very little dependence on stellar mass, if a constant conversion factor ($\alpha_{\rm CO}$) of CO luminosity to molecular gas mass is assumed. Combining the atomic and molecular gas observations with the stellar build-up illustrates that galactic mass assembly in Milky-Way sized haloes proceeds from smooth accretion at high redshifts, towards becoming merger-dominated at late times ($z \lesssim 0.6$). Our results can be used to constrain numerical simulations of the dominant growth and accretion processes of galaxies over cosmic history.