We introduce Meraxes, a new, purpose-built semi-analytic galaxy formation model designed for studying galaxy growth during reionization. Meraxes is the first model of its type to include a temporally and spatially coupled treatment of reionization and is built upon a custom (100 Mpc) 3 N -body simulation with high temporal and mass resolution, allowing us to resolve the galaxy and star formation physics relevant to early galaxy formation. Our fiducial model with supernova feedback reproduces the observed optical depth to electron scattering and evolution of the galaxy stellar mass function between z=5 and 7, predicting that a broad range of halo masses contribute to reionization. Using a constant escape fraction and global recombination rate, our model is unable to simultaneously match the observed ionizing emissivity at z < ∼ 6. However, the use of an evolving escape fraction of 0.05-0.1 at z∼6, increasing towards higher redshift, is able to satisfy these three constraints. We also demonstrate that photoionization suppression of low mass galaxy formation during reionization has only a small effect on the ionization history of the inter-galactic medium. This lack of 'selfregulation' arises due to the already efficient quenching of star formation by supernova feedback. It is only in models with gas supply-limited star formation that reionization feedback is effective at regulating galaxy growth. We similarly find that reionization has only a small effect on the stellar mass function, with no observationally detectable imprint at M * >10 7.5 M . However, patchy reionization has significant effects on individual galaxy masses, with variations of factors of 2-3 at z=5 that correlate with environment.

Low‐frequency observatories are currently being constructed with the goal of detecting redshifted 21‐cm emission from the epoch of reionization. These observatories will also be able to detect intensity fluctuations in the cumulative 21‐cm emission after reionization, from hydrogen in unresolved damped Lyα absorbers (such as gas‐rich galaxies) down to a redshift z∼ 3.5. The inferred power spectrum of 21‐cm fluctuations at all redshifts will show acoustic oscillations, whose comoving scale can be used as a standard ruler to infer the evolution of the equation of state for the dark energy. We find that the first generation of low‐frequency experiments (such as MWA or LOFAR) will be able to constrain the acoustic scale to within a few per cent in a redshift window just prior to the end of the reionization era, provided that foregrounds can be removed over frequency bandpasses of ≳8 MHz. This sensitivity to the acoustic scale is comparable to the best current measurements from galaxy redshift surveys, but at much higher redshifts. Future extensions of the first‐generation experiments (involving an order of magnitude increase in the antennae number of the MWA) could reach sensitivities below 1 per cent in several redshift windows and could be used to study the dark energy in the unexplored redshift regime of 3.5 ≲z≲ 12. Moreover, new experiments with antennae designed to operate at higher frequencies would allow precision measurements (≲1 per cent) of the acoustic peak to be made at more moderate redshifts (1.5 ≲z≲ 3.5), where they would be competitive with ambitious spectroscopic galaxy surveys covering more than 1000 deg2. Together with other data sets, observations of 21‐cm fluctuations will allow full coverage of the acoustic scale from the present time out to z∼ 12.

In this paper we present calculations of the UV luminosity function from the Dark-ages Reionization And Galaxy-formation Observables from Numerical Simulations (DRAG-ONS) project, which combines N-body, semi-analytic and semi-numerical modelling designed to study galaxy formation during the Epoch of Reionization. Using galaxy formation physics including supernova feedback, the model naturally reproduces the UV LFs for high-redshift star-forming galaxies from z∼5 through to z∼10. We investigate the luminosity-star formation rate (SFR) relation, finding that variable SFR histories of galaxies result in a scatter around the median relation of 0.1-0.3 dex depending on UV luminosity. We find close agreement between the model and observationally derived SFR functions. We use our calculated luminosities to investigate the luminosity function below current detection limits, and the ionizing photon budget for reionization. We predict that the slope of the UV LF remains steep below current detection limits and becomes flat at M UV −14. We find that 48 (17) per cent of the total UV flux at z∼6 (10) has been detected above an observational limit of M UV ∼−17, and that galaxies fainter than M UV ∼−17 are the main source of ionizing photons for reionization. We investigate the luminosity-stellar mass relation, and find a correlation for galaxies with M UV <−14 that has the form M * ∝10 −0.47MUV , in good agreement with observations, but which flattens for fainter galaxies. We determine the luminosity-halo mass relation to be M vir ∝10 −0.35MUV , finding that galaxies with M UV =−20 reside in host dark matter haloes of 10 11.0±0.1 M at z∼6, and that this mass decreases towards high redshift.

We assess the impact of inhomogeneous reionization on detection of H II regions surrounding luminous high-redshift quasars using planned low-frequency radio telescopes. Our approach is to implement a seminumerical scheme to calculate the three-dimensional structure of ionized regions surrounding a massive halo at high redshift, including the ionizing influence of a luminous quasar. As part of our analysis we briefly contrast our scheme with published seminumerical models. We calculate mock 21-cm spectra along the line of sight towards highredshift quasars, and estimate the ability of the planned Murchison Widefield Array to detect the presence of H II regions. The signal-to-noise ratio for detection will drop as the characteristic bubble size grows during reionization because the quasar's influence becomes less prominent. However, quasars will imprint a detectable signature on observed 21-cm spectra that is distinct from a region of typical intergalactic medium (IGM). At epochs where the mean hydrogen neutral fraction is ≈30 per cent or greater we find that neutral gas in the IGM surrounding a single quasar will be detectable (at a significance of 5σ ) within 100-h integrations in more than 50 per cent of cases. 1000-h integrations will be required to detect a smaller neutral fraction of 15 per cent in more than 50 per cent of cases. A highly significant detection will be possible in only 100 h for a stack of 10 smaller 3 proper Mpc H II regions. The accurate measurement of the global average neutral fraction ( x H I ) will be limited by systematic fluctuations between lines of sight for single H II regions. We estimate the accuracy with which the global neutral fraction could be measured from a single H II region to be 50, 30 and 20 per cent for x H I ≈ 0.15, 0.3 and 0.5, respectively.

Motivated by recent measurements of the number density of faint AGN at high redshift, we investigate the contribution of quasars to reionization by tracking the growth of central supermassive black holes in an update of the Meraxes semi-analytic model. The model is calibrated against the observed stellar mass function at z ∼ 0.6 − 7, the black hole mass function at z 0.5, the global ionizing emissivity at z ∼ 2 − 5 and the Thomson scattering optical depth. The model reproduces a Magorrian relation in agreement with observations at z < 0.5 and predicts a decreasing black hole mass towards higher redshifts at fixed total stellar mass. With the implementation of an opening angle of 80 deg for quasar radiation, corresponding to an observable fraction of ∼23.4 per cent due to obscuration by dust, the model is able to reproduce the observed quasar luminosity function at z ∼ 0.6 − 6. The stellar light from galaxies hosting faint AGN contributes a significant or dominant fraction of the UV flux. At high redshift, the model is consistent with the bright end quasar luminosity function and suggests that the recent faint z ∼ 4 AGN sample compiled by Giallongo et al. (2015) includes a significant fraction of stellar light. Direct application of this luminosity function to the calculation of AGN ionizing emissivity consequently overestimates the number of ionizing photons produced by quasars by a factor of 3 at z ∼ 6. We conclude that quasars are unlikely to make a significant contribution to reionization.