Astrophysics & Cosmology from Line Intensity Mapping vs Galaxy Surveys

TL;DR

This paper evaluates the potential of line intensity mapping (LIM) to map high-redshift galaxies and matter density, analyzing degeneracies, measurement strategies, and comparing LIM's effectiveness to galaxy surveys for cosmology and astrophysics.

Contribution

It extends previous LIM analysis by incorporating cosmic variance, pixel correlations, luminosity bias, and redshift distortions, providing a comprehensive assessment of LIM's capabilities.

Findings

LIM can probe galaxies too faint for detection in high-noise regimes.

Redshift-space distortions help break degeneracies in LIM measurements.

LIM can outperform galaxy surveys in tracing matter density under certain conditions.

Abstract

Line intensity mapping (LIM) proposes to efficiently observe distant faint galaxies and map the matter density field at high redshift. Building upon the formalism in the companion paper, we first highlight the degeneracies between cosmology and astrophysics in LIM. We discuss what can be constrained from measurements of the mean intensity and redshift-space power spectra. With a sufficient spectral resolution, the large-scale redshift-space distortions of the 2-halo term can be measured, helping to break the degeneracy between bias and mean intensity. With a higher spectral resolution, measuring the small-scale redshift-space distortions disentangles the 1-halo and shot noise terms. Cross-correlations with external galaxy catalogs or lensing surveys further break degeneracies. We derive requirements for experiments similar to SPHEREx, HETDEX, CDIM, COMAP and CONCERTO. We then revisit the question of the optimality of the LIM observables, compared to galaxy detection, for astrophysics and cosmology. We use a matched filter to compute the luminosity detection threshold for individual sources. We show that LIM contains information about galaxies too faint to detect, in the high-noise or high-confusion regimes. We quantify the sparsity and clustering bias of the detected sources and compare them to LIM, showing in which cases LIM is a better tracer of the matter density. We extend previous work by answering these questions as a function of Fourier scale, including for the first time the effect of cosmic variance, pixel-to-pixel correlations, luminosity-dependent clustering bias and redshift-space distortions.