| Abstract: The Halo Occupation Distribution (HOD) framework provides a key theoretical bridge between the observed clustering of active galactic nuclei (AGN) and the underlying dark matter halo population. Despite its widespread use, standard HOD prescriptions often assume weak or negligible redshift evolution, an assumption that remains poorly tested for luminous quasars. We use state-of-the-art cosmological hydrodynamic simulations, IllustrisTNG and SIMBA, to investigate the redshift, luminosity, and black hole mass dependence of quasar HODs.
Across both simulations, we find that quasar activity is significantly suppressed at a characteristic halo mass scale of 10^13 solar mass, leading to a pronounced departure from commonly adopted HOD models. This quenching becomes increasingly significant at low redshifts for luminosity-selected quasar samples. We demonstrate that neglecting this effect can introduce substantial biases in clustering-based inferences of quasar host halo masses, with errors of up to an order of magnitude in the central occupation and tens of percent in the inferred satellite fraction. While the central occupation is strongly quenched, the satellite occupation retains a power-law–like form, resulting in a satellite fraction that increases monotonically toward low redshift.
We also find that, in contrast to luminosity-selected quasars, the halo occupation of mass-selected supermassive black holes exhibits minimal redshift evolution and is well described by a step function plus power-law form, similar to that found for galaxies and low-luminosity AGN. By comparing different feedback implementations in SIMBA, we find that jet-mode feedback plays a central role in quenching central quasar activity. These results highlight the necessity of incorporating redshift-dependent and feedback-driven effects into HOD modeling, with important implications for the interpretation of host halo masses from AGN clustering in the context of forthcoming large-scale survey data. (arXiv:2508.08851v2) |
