Galaxy assembly bias in cosmological hydrodynamical simulations: a comparison between SIMBA and IllustrisTNG

Abstract

Modelling of large-scale structure is increasingly concerned with galaxy assembly bias (GAB), the dependence of galaxy clustering on quantities other than halo mass. We investigate how baryonic physics affects the strength and redshift evolution of GAB using the largest runs of two state-of-the-art cosmological hydrodynamical simulations: SIMBA and IllustrisTNG. We quantify GAB by comparing the clustering of stellar-mass-selected galaxies to that of shuffled samples, where galaxies are randomly reassigned to haloes of similar mass. We find that GAB in both simulations increases from approximately zero at $z=5$ to a $\sim\,5\%$ change in clustering amplitude at $z=2$. After this epoch, the trends diverge: GAB in TNG continues to increase, reaching $\sim\,10\%$ at $z=0$, while in SIMBA it decreases to nearly zero. By further shuffling galaxies within bins of halo mass and cosmic environment -- characterised by smoothed matter overdensity ($\delta_5$) and tidal anisotropy ($\alpha_5$) -- we show that most of the GAB in both simulations can be attributed to the overdensity, while tidal anisotropy contributes negligibly in both simulations. Exploring this effect from the point of view of the halo occupation distribution (HOD), we find that numbers of central and satellite galaxies vary with overdensity -- but only near the respective turn-on masses for these two constituents: the galaxy contents of high-mass haloes are very nearly independent of environment. We present a simple parameterisation that allows the HOD modelling to be modified to reflect this form of density-dependent GAB.