Faint field galaxies: an explanation of the faint blue excess using number evolution models

Abstract

Pure luminosity evolution models for galaxies provide an unacceptable fit to the redshifts and colors of faint galaxies. In this paper we demonstrate, using HST morphological number counts derived both from the $I_{814}$-band of WFPC2 in the Medium Deep Survey (MDS) and the Hubble Deep Field (HDF) and from the $H_{1.6}$-band of NICMOS, and ground-based spectroscopic data of the Hawaii Deep Field and the Canada-France Redshift Survey, that number evolution is necessary for galaxies, regardless of whether the cosmic geometry is flat, open, or $Λ$-dominated. Furthermore, we show that the number evolution is small at redshifts of z<1, but large at z>1, and that this conclusion is valid for all the three cosmological models under consideration. If the universe is open or $Λ$-dominated, the models, which are subject to the constraint of the conservation of the comoving mass density of galaxies, naturally predict a population of star-forming galaxies with the redshift distribution peaking at $z=2\sim3$, which seems to be consistent with the recent findings from Lyman-break photometric selection techniques. If the cosmological model is flat, however, the conservation of the comoving mass density is invalid. Hence, in order to account for the steep slope of B-band number counts at faint magnitudes in the flat universe, such a star-forming galaxy population has to be introduced ad hoc into the modelling alongside the merger assumption.