We investigate the dependence of dark matter halo clustering on halo formation time, density profile concentration, and subhalo occupation number, using high-resolution numerical simulations of a LCDM cosmology. We confirm results that halo clustering is a function of halo formation time, and that this trend depends on halo mass. For the first time, we show unequivocally that halo clustering is a function of halo concentration and show that the dependence of halo bias on concentration, mass, and redshift can be accurately parameterized in a simple way: b(c,M|z) = b(M|z) b(c|M/M*). The scaling between bias and concentration changes sign with the value of M/M*: high concentration (early forming) objects cluster more strongly for M <~ M* while low concentration (late forming) objects cluster more strongly for rare high-mass halos, M >~ M*. We show the first explicit demonstration that host dark halo clustering depends on the halo occupation number (of dark matter subhalos) and discuss implications for halo model calculations of dark matter power spectra and galaxy clustering statistics. The effect of these halo properties on clustering is strongest for early-forming dwarf-mass halos, which are significantly more clustered than typical halos of their mass. Our results suggest that isolated low-mass galaxies (e.g. low surface-brightness dwarfs) should have more slowly-rising rotation curves than their clustered counterparts, and may have consequences for the dearth of dwarf galaxies in voids. They also imply that self calibrating richness-selected cluster samples with their clustering properties might overestimate cluster masses and bias cosmological parameter estimation.
Download