A PARAMETRIC STUDY OF POSSIBLE SOLUTIONS TO THE HIGH-REDSHIFT OVERPRODUCTION OF STARS IN MODELED DWARF GALAXIES

Both numerical hydrodynamic and semi-analytic cosmological models of galaxy formation struggle to match observed star formation histories of galaxies in low-mass halos (M-H less than or similar to 10(11) M-circle dot ), predicting more star formation at high redshift and less star formation at low redshift than observed. The fundamental problem is that galaxies' gas accretion and star formation rates are too closely coupled in the models: the accretion rate largely drives the star formation rate. Observations point to gas accretion rates that outpace star formation at high redshift, resulting in a buildup of gas and a delay in star formation until lower redshifts. We present three empirical adjustments of standard recipes in a semi-analytic model motivated by three physical scenarios that could cause this decoupling: (1) the mass-loading factors of outflows driven by stellar feedback may have a steeper dependence on halo mass at earlier times, (2) the efficiency of star formation may be lower in low-mass halos at high redshift, and (3) gas may not be able to accrete efficiently onto the disk in low-mass halos at high redshift. These new recipes, once tuned, better reproduce the evolution of f(*) = M-*/M-H as a function of halo mass as derived from abundance matching over redshifts z = 0 to 3, though they have different effects on cold gas fractions, star formation rates, and metallicities. Changes to gas accretion and stellar-driven winds are promising, while direct modification of the star formation timescale requires drastic measures that are not physically well motivated.