Predictions of quasar clustering: Redshift, luminosity, and selection dependence

We show that current clustering observations of quasars and luminous active galactic nuclei (AGNs) can be explained by a merger model augmented by feedback from outflows. Using numerical simulations large enough to study clustering out to 25 comoving h ^-1 Mpc, we calculate correlation functions, biases, and correlation lengths as a function of AGN redshift and optical and X-ray luminosity. At optical wavelengths, our results match a wide range of current observations and generate predictions for future data sets. We reproduce the weak luminosity dependence of clustering over the currently well-measured range and predict a much stronger dependence at higher luminosities. The increase in the amplitude of binary quasar clustering observed in the Sloan Digital Sky Survey (SDSS) is also reproduced and is predicted to occur at higher redshift, an effect that is due to the one-halo term in the correlation function. On the other hand, our results do not match the rapid evolution of the correlation length observed in the SDSS at z ≃ 3, a discrepancy that is at least partially due to differences in the scales probed by our simulation versus this survey. In fact, we show that changing the distances sampled from our simulations can produce changes as large as 40% in the fitted correlation lengths. Finally, in the X-ray, our simulations produce correlation lengths similar to that observed in the Chandra Deep Field (CDF) North, but not the significantly larger correlation length observed in the CDF-South.