Galaxy properties from the ultraviolet to the far-infrared: Λ cold dark matter models confront observations

Abstract

We combine a semi-analytic model of galaxy formation with simple analytic recipes describing the absorption and re-emission of starlight by dust in the interstellar medium of galaxies. We use the resulting models to predict galaxy counts and luminosity functions from the far-ultraviolet (FUV) to the submillimetre, from redshift five to the present, and compare with an extensive compilation of observations. We find that in order to reproduce the rest-UV and optical luminosity functions at high redshift, we must assume an evolving normalization in the dust-to-metal ratio, implying that galaxies of a given bolometric luminosity (or metal column density) must be less extinguished than their local counterparts. In our best-fitting model, we find remarkably good agreement with observations from rest ~1500 Å to m. At longer wavelengths, most dramatically in the submillimetre, our models underpredict the number of bright galaxies by a large factor. The models reproduce the observed total IR luminosity function fairly well. We show the results of varying several ingredients of the models, including various aspects of the dust attenuation recipe, the dust emission templates and the cosmology. We use our models to predict the integrated extragalactic background light, and compare with an observationally motivated extragalactic background light EBL model and with other available observational constraints

Dusty star-forming galaxies at high redshift

“…In other words, galaxies at a comparable luminosity at highredshift compared to a local analog needed less dust obscuration. This made reproducing the observed properties of dusty galaxies at high-z more difficult, however, and as a consequence the Somerville et al (2012) model was unable to match the observed 250 µm SPIRE number counts, as well as the 850 µm Scuba counts. As shown by Niemi et al (2012), the Somerville et al (2012) model works rather well at lower redshifts (i.e.…”

“…It is noted, however, that this model over produces stars in massive galaxies, and therefore is incompatible with observations of local galaxies. Somerville et al (2012) presented an updated form of the Somerville & Primack (1999) SAM (i.e. the "Santa Cruz SAM") that was combined with estimates for dust attenuation by both diffuse dust and stellar birth clouds surrounding young clusters.…”

“…The uncertainty, of course, is whether locally-calibrated observational templates are applicable to galaxies at the same luminosity at high-z. A recent example of this sort of application of observed SED templates to SAMs is presented in Somerville et al (2012); Kim et al (2014).…”

“…Lacey et al, 2008Lacey et al, , 2010Narayanan et al, 2010a;Niemi et al, 2012), and have at the least provided a strong constraint on general galaxy formation models (e.g. Somerville et al, 2012).…”

Redshift evolution of stellar mass versus gas fraction relation in 0 <z< 2 regime: observational constraint for galaxy formation models

“…• Popping et al (2014, SAM): Popping et al implemented pressure-based (Blitz & Rosolowsky 2006) and the metallicity-based (Gnedin et al 2009) H2 formation models in the Santa Cruz semi-analytical model (Somerville & Primack 1999;Somerville et al 2008aSomerville et al , 2012, in which the dark matter halo merger tree was constructed using the Extended Press-Schechter formalism ( star formation from molecular gas , and feedbacks from SNe and AGN. We obtain their MH 2 /M⋆ values (metallicity-based H2 model and star formation model based on Bigiel et al 2008) from Fig.…”