Large scale bias and stochasticity of halos and dark matter

Abstract

On large scales galaxies and their halos are usually assumed to trace the dark matter with a constant bias and dark matter is assumed to trace the linear density field. We test these assumption using several large N-body simulations with 384^3-1024^3 particles and box sizes between 100-1000h/Mpc, which can both resolve the small galactic size halos and sample the large scale fluctuations. We explore the average halo bias relation as a function of halo mass and show that existing fitting formulae overestimate the halo bias by up to 20% in the regime just below the nonlinear mass. We propose a new expression that fits our simulations well. We find that the halo bias is nearly constant, b~0.65-0.7, for masses below one tenth of the nonlinear mass. We explore next the relation between the initial and final dark matter in individual Fourier modes and show that there are significant fluctuations in their ratio, ranging from 10% rms at k~0.03h/Mpc to 50% rms at k~0.1h/Mpc. We argue that these large fluctuations are caused by perturbative effects beyond the linear theory, which are dominated by long wavelength modes with large random fluctuations. Similar or larger fluctuations exist between halos and dark matter and between halos of different mass. While these fluctuations are small compared to the sampling variance, they are significant for attempts to determine the bias by relating directly the maps of galaxies and dark matter or the maps of different galaxy populations, which would otherwise be immune to sampling variance.