Phenomenological power spectrum models for H$\alpha$ emission line galaxies from the Nancy Grace Roman Space Telescope

TL;DR

This paper develops and tests phenomenological models for the power spectrum of Hα emission line galaxies observed by the Nancy Grace Roman Space Telescope, aiming to improve cosmological parameter estimation.

Contribution

It introduces optimized phenomenological models for redshift-space distortions and nonlinear effects, validated with realistic galaxy mocks for Roman's HLSS survey.

Findings

Best model combination with halo model emulation yields unbiased parameters.

Models are comparable to EFT across relevant scales.

Tools are useful for dark energy and gravity tests with Roman.

Abstract

The High Latitude Spectroscopic Survey (HLSS) is the reference baseline spectroscopic survey for NASA's Nancy Grace Roman space telescope, measuring redshifts of $\sim 10$M H$\alpha$ emission line galaxies over a $2000$ deg$^2$ footprint at $z=1-2$. In this work, we use a realistic Roman galaxy mock catalogue to explore optimal phenomenological modeling of the measured power spectrum. We consider two methods for modeling the redshift-space distortions (Kaiser squashing and another with a window function on $\beta$ that selects out the coherent radial infall pairwise velocities, $M_A$ and $M_B$, respectively), two models for the nonlinear impact of baryons that smears the BAO signal (a fixed ratio between the smearing scales in the perpendicular and parallel dimensions and another where these smearing scales are kept as a free parameters, P$_{dw}(k|k_*)$ and P$_{dw}(k|\Sigma_\perp,\Sigma_\parallel)$, respectively), and two analytical emulations of nonlinear growth (one employing the halo model and another formulated from simulated galaxy clustering of a semi-analytical model, $F_{HM}$ and $F_{SAM}$, respectively). We find that the best model combination employing $F_{HM}$ is $P_{dw}(k|k_*)*F_{HM}*M_B$, while the best combination employing $F_{SAM}$ is $P_{dw}(k|k_*)*F_{SAM}*M_B$, which leads to unbiased measurements of cosmological parameters. We compare these to the Effective Field Theory of Large-Scale Structure perturbation theory model $P_{EFT}(k|\Theta)$, and find that our simple phenomenological models are comparable across the entire redshift range for $k_{max}=0.25$ and $0.3$ $h$/Mpc. We expect the tools that we have developed to be useful in probing dark energy and testing gravity using Roman in an accurate and robust manner.