The standard Halo Occupation Distribution (HOD) models were originally developed based on results from semi-analytic and hydrodynamical galaxy formation models. Those models have since progressed, in particular to include AGN feedback to match the galaxy luminosity function in a universe with the observed baryon fraction. AGN feedback affects the relationship between galaxy stellar mass and luminosity, in particular making the relationship non-monotonic. For matched number density samples, galaxies in luminosity-threshold samples occupy a different range of halo masses from those in stellar-mass-threshold samples. We find that the shapes of the HODs of luminosity-threshold samples are slightly more complicated in semi-analytic galaxy formation models that include AGN feedback than are assumed by standard HOD models. We also find that subhalo abundance matching (SHAM) does not preserve these non-standard shapes. We show that catalogues created using SHAM and the semi-analytic model Galform that have the same large-scale 2-point clustering by construction have different void probability functions (VPFs) in both real and redshift space. We find that these differences arise from the different HOD shapes, as opposed to assembly bias, which indicates that the VPF could be used to test the suitability of an HOD model with real data.
Download