Simulating realistic Lyman-$α$ emitting galaxies including the effect of radiative transfer

TL;DR

This work develops a high-fidelity LAE mock catalog for $z\sim 2$–$3$ by coupling an empirical UniverseMachine galaxy–halo model to a physically grounded Ly$\alpha$ radiative transfer treatment based on an expanding spherical shell. The model uses three dimensionless parameters $(\alpha,\beta,\gamma)$ to map halo/galaxy properties to the ISM/CGM, dust, and HI content, then applies a redshift-dependent mean IGM transmission to obtain observed Ly$\alpha$ luminosities and spectra. Calibrated to reproduce both the Ly$\alpha$ luminosity function and the LAE angular clustering, the framework also matches EW distributions, Ly$\alpha$ escape fractions, velocity offsets, and dust reddening trends, yielding robust predictions for the LAE–halo connection and satellite populations. The resulting catalog provides a practical tool for upcoming LAE surveys, enabling forward modeling of selection effects and cosmological analyses while highlighting limitations such as the expanding-shell RT simplification and fixed IGM treatment. Overall, the paper demonstrates that a compact, RT-informed, empirically anchored approach can capture the dominant physics controlling Ly$\alpha$ visibility and clustering across $z\sim 2$–$3$, with explicit pathways to improve fidelity in future work.

Abstract

We present an empirical yet physically motivated simulation of realistic Lyman-$α$ emitters (LAEs) at $z\sim2-3$, crucial for ongoing and forthcoming cosmological LAE surveys. We combine an empirical $\mathtt{UniverseMachine}$ galaxy-halo model with a simple spherical expanding shell model for the Lyman-$α$ radiative transfer, calibrating only three free parameters to simultaneously reproduce the observed Lyman-$α$ luminosity function and the angular clustering. Our LAE model is further supported by its consistency with other observables such as the Lyman-$α$ equivalent width distribution, the Lyman-$α$ escape fraction as a function of stellar mass and dust reddening, and the systemic velocity offsets. Our LAE model provides predictions for the halo occupation distributions for LAEs and relationship between Ly$α$ luminosity and halo mass, including the distribution of satellite LAEs. Our work provides a crucial first step towards creating a high-fidelity LAE synthetic catalog for the LAE cosmology surveys. We make our LAE catalog and spectra publicly available upon publication.

Figures (30)

• Figure 1: UV luminosity function at $z\sim2$ (blue) and $z\sim3$ (red) compared to observational data from 10.1093/mnras/stv2857 at $z\sim2$ and Reddy_2008Reddy_2009 at $z\sim3$. The model (thick solid curves) calibrates the SFR to reproduce the UV luminosity function incorporating the empirical dust model from 2020MNRAS.492.5167V. The dust model is not included in dashed lines. The thin dashed lines assume the $\rm SFR_{\rm int}$ from the UniverseMachine model assuming Kennicutt relation, $L_{\rm cont,\lambda} = 1.4 \times 10^{40}\,{\rm erg\,s^{-1}\,\text{\AA}^{-1}}\times(\rm SFR/M_{\odot}yr^{-1})$Kennicutt:1998ARAA.
• Figure 2: Left: ISM escape fraction, $f_{\rm esc}^{\rm ISM}$, as a function of expansion velocity, $V_{\rm exp}$, neutral hydrogen column density, $N_{\rm HI}$, and dust absorption optical depth, $\tau_a$. Top panel shows how $f_{\rm esc}^{\rm ISM}$ changes for various $N_{\rm HI}$ values at a fixed $\tau_a=0.01$. The bottom panel presents the dependence on $\tau_a$ for a fixed $\log_{10}({\rm N}_{\rm HI}/{\rm cm}^{-2})=20$. Right: Example line profile of the galaxies with various shell model parameters. The three parameters in each legend corresponds to $\log_{10}({\rm N}_{\rm HI}/{\rm cm}^{-2})$, $V_{\rm exp}/({\rm km\,s^{-1}})$, and $f_{\rm esc}^{\rm ISM}$. The vertical dashed line shows the rest frame Ly$\alpha$ wavelength, $\lambda_{\rm Ly\alpha}=1215.67\,\textup{\AA}$.
• Figure 3: The impact of Ly$\alpha$ RT on an example Ly$\alpha$ profile. This galaxy resides at $z=2.42$, with $M^{*}=4.59\times 10^{8}\, \rm M_{\odot}$, $\rm SFR=1.01\,\,M_{\odot}\, yr^{-1}$ and the shell model's key parameters of $V_{\rm exp} = 21.41 \,\rm \,\,km/s$, $\rm log_{10} \, N_{\rm HI}= 19.48\,\rm cm^{-1}$ and $\tau_a =0.03$. The light red solid curve represents the emitted Ly$\alpha$ luminosity density from the ISM/CGM. The dark dashed curve corresponds to the observed Ly$\alpha$ spectrum after incorporating both ISM/CGM and IGM effects. The blue dot-dashed curve (right y-axis) represents the IGM transmission. The vertical black dashed line marks the rest-frame Ly$\alpha$ line center wavelength, ($\lambda = 1215.67 \,\text{\AA}$).
• Figure 4: Scaling relations we adopt in this work. We show our galaxy sample at $z\sim 3$ where LAEs ($\mathrm{log}_{10}(L_{Lya}/\,\rm erg\,s^{-1}) \geq 41.7$ and $\mathrm{EW}_0 > 40 \,\, \text{\AA}$) are shown in pink while others are represented in blue. Top left: Stellar half-mass radius, $r_{1/2}$, versus the halo virial radius, $R_{\rm vir}$. The gray solid line represents the empirical relation $r = 0.015 \, \rm R_{\rm vir}$ from Kravtsov_2013 with the shaded region indicating the $2\sigma$ scatter. Top right: Neutral atomic hydrogen gas mass fraction, $\rm log_{10} (M_{\rm HI}/M_*)$, as a function of stellar mass, serving as a scaling relation for inferring the neutral hydrogen column density. Measurements from the XGASS survey 10.1093/mnras/sty089 are shown in gray. Bottom: Cold gas metallicity (gas-phase oxygen abundance, $12+\rm log_{10}(Z_{\rm cold})$) as a function of stellar mass compared with several observational mass-metallicity relations including 2004ApJ...613..898TErb_2006mannucci2009lsd2011ApJ...729..140F2014ApJ...795..165Ssanders2015mosdef as depicted.
• Figure 5: Left: The derived IGM transmission as a function of wavelength, color-coded for various redshifts. The vertical dashed line indicates the rest frame Ly$\alpha$ line center wavelength ($\lambda = 1215.7 \,\text{\AA}$). Dashed curves represent the transmission function originally provided by Laursen:2011ApJ, while solid curves are the ones adjusted by assuming that the mean transmission function blueward of the Ly$\alpha$ line follows the empirical relation represented in Eq. (\ref{['eq:T_z']}). Right: A comparison between observations and the model used in this study, illustrating transmitted flux blueward of the Ly$\alpha$ line with respect to redshift. The solid line corresponds to Eq. (\ref{['eq:T_z']}).