Halo assembly bias from Separate Universe simulations

TL;DR

This paper calibrates halo assembly bias using Separate Universe simulations, providing high-precision measurements of bias parameters and revealing non-universality in their relations, which aids in detecting and mitigating assembly bias in cosmology.

Contribution

It introduces a new analytical framework combined with simulations to precisely calibrate halo assembly bias and extends this to higher-order bias parameters, including the first calibration of $b_2$.

Findings

High-precision calibration of linear assembly bias $b_1$.

First calibration of second-order bias $b_2$ related to assembly bias.

Detection of non-universality in the $b_1 - b_2$ relation.

Abstract

We present a calibration of halo assembly bias using the Separate Universe technique. Specifically, we measure the response of halo abundances at fixed mass and concentration to the presence of an infinite-wavelength initial perturbation. We develop an analytical framework for describing the concentration dependence of this peak-background split halo bias -- a measure of assembly bias -- relying on the near-Lognormal distribution of halo concentration at fixed halo mass. The combination of this analytical framework and the Separate Universe technique allows us to achieve very high precision in the calibration of the linear assembly bias $b_1$, and qualitatively reproduces known trends such as the monotonic decrease (increase) of $b_1$ with halo concentration at large (small) masses. The same framework extends to the concentration dependence of higher order bias parameters $b_n$, and we present the first calibration of assembly bias in $b_2$. Our calibrations are directly applicable in analytical Halo Model calculations that seek to robustly detect galaxy assembly bias in observational samples. We detect a non-universality in the $b_1 - b_2$ relation arising from assembly bias, and suggest that simultaneous measurements of these bias parameters could be used to both detect the signature of assembly bias as well as mitigate its effects in cosmological analyses.