# Modelling the observed luminosity function and clustering evolution of Lyman-$\alpha$ emitters: growing evidence for late reionization

## TL;DR

This paper models high-redshift Lyman-alpha emitters to understand their luminosity and clustering evolution, providing evidence for late reionization consistent with recent CMB and Ly alpha forest observations.

## Contribution

It introduces an empirical model calibrated to observations, simulates LAE populations in large cosmological volumes, and explores the impact of the IGM and CGM on Ly alpha transmission, supporting late reionization.

## Key findings

- Late reionization is supported by the model.
- Significant Ly alpha attenuation occurs in the CGM and infalling gas.
- Predictions made for LAE observations at z=6.6, 7.0, 7.3.

## Abstract

We model the high redshift (z > 5) Lyman-$\alpha$ emitting (LAE) galaxy population using the empirical rest-frame equivalent width distribution. We calibrate to the observed luminosity function and angular correlation function at z = 5.7 as measured by the SILVERRUSH survey. This allows us to populate the high-dynamic-range Sherwood simulation suite with LAEs, and to calculate the transmission of their Ly $\alpha$ emission through the inter-galactic medium (IGM). We use this simulated population to explore the effect of the IGM on high-redshift observations of LAEs, and make predictions for the narrowband filter redshifts at z = 6.6, 7.0 and 7.3. Comparing our model with existing observations, we find a late reionization is suggested, consistent with the recent low optical depth derived from the cosmic microwave background (CMB) by the Planck collaboration and the opacity fluctuations in the Ly $\alpha$ forest. We also explore the role of the circum-galactic medium (CGM) and the larger volume of gas which is infalling onto the host halo versus the IGM in attenuating the Ly $\alpha$ signal, finding that a significant fraction of the attenuation is due to the CGM and infalling gas, which increases towards the end of reionization, albeit with a large scatter across the mock LAE population.