Cosmic variance or galaxy bias? Disentangling finite-volume and galaxy formation effects in cosmological analysis

Abstract

Current and forthcoming galaxy redshift surveys, such as DESI and Euclid, are going to bring cosmological analysis to an unprecedentedly exquisite level of precision in the determination of the cosmological parameters. However, these efforts require a high degree of control over theory and systematics, to obtain unbiased results. In this sense, the cosmic variance associated to finite-volume effects represents a major challenge and should adequately accounted for. In this work, we revisit the definition of cosmic variance and develop a novel framework to describe it using a `galaxy biasing' formalism. In particular, we use halo/galaxy Eulerian perturbation theory to relate the density field from an arbitrary cosmic realization to its counterpart having statistical properties reproducing the ensemble average, introducing a new set of bias parameters. We then apply this idea to the description of the non-linear shift of BAO, disentangling the source of uncertainty from cosmic variance and galaxy biasing associated with the measurement of the BAO scale. We finally check our analytical argument against cosmological variance-suppressed $N$-body simulations, finding an expected reduction in the uncertainty on the BAO peak position. We conclude that extra care should be used when inferring cosmological information from perturbative approaches involving the estimation of bias parameters and propose new practical strategies to optimally leverage the novel formulation of cosmic variance presented herein in cosmological analysis.