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.

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.

We investigate high-redshift galaxy sizes using a semi-analytic model constructed for the Dark-ages Reionization And Galaxy-formation Observables from Numerical Simulation project. Our fiducial model, including strong feedback from supernovae and photoionization background, accurately reproduces the evolution of the stellar mass function and UV luminosity function. Using this model, we study the size-luminosity relation of galaxies and find that the effective radius scales with UV luminosity as R e ∝ L 0.25 at z∼5-9. We show that recently discovered very luminous galaxies at z∼7 (Bowler et al. 2016) and z∼11 (Oesch et al. 2016) lie on our predicted size-luminosity relations. We find that a significant fraction of galaxies at z > 8 will not be resolved by JWST, but GMT will have the ability to resolve all galaxies in haloes above the atomic cooling limit. We show that our fiducial model successfully reproduces the redshift evolution of average galaxy sizes at z > 5. We also explore galaxy sizes in models without supernova feedback. The no-supernova feedback models produce galaxy sizes that are smaller than observations. We therefore confirm that supernova feedback plays an important role in determining the size-luminosity relation of galaxies and its redshift evolution during reionization.

Using Hubble data, including new grism spectra, Oesch et al. recently identified GNz11, an M UV =-21.1 galaxy at z=11.1 (just 400 Myr after the big bang). With an estimated stellar mass of ∼10 9 M , this galaxy is surprisingly bright and massive, raising questions as to how such an extreme object could form so early in the Universe. Using Meraxes, a semi-analytic galaxy-formation model developed as part of the Dark-ages Reionization And Galaxy-formation Observables from Numerical Simulations (DRAGONS) programme, we investigate the potential formation mechanisms and eventual fate of GN-z11. The volume of our simulation is comparable to that of the discovery observations and possesses two analogue galaxies of similar luminosity to this remarkably bright system. Existing in the two most massive subhaloes at z=11.1 (M vir =1.4×10 11 M and 6.7×10 10 M ), our model analogues show excellent agreement with all available observationally derived properties of GN-z11. Although they are relatively rare outliers from the full galaxy population at high-z, they are no longer the most massive or brightest systems by z=5. Furthermore, we find that both objects possess relatively smooth, but extremely rapid mass growth histories with consistently high star formation rates and UV luminosities at z>11, indicating that their brightness is not a transient, merger-driven feature. Our model results suggest that future wide-field surveys with the James Webb Space Telescope may be able to detect the progenitors of GN-z11 analogues out to z∼13-14, pushing the frontiers of galaxy-formation observations to the early phases of cosmic reionization and providing a valuable glimpse of the first galaxies to reionize the Universe on large scales.

We investigate the clustering properties of Lyman-break galaxies (LBGs) at z ∼ 6 -8.Using the semi-analytical model Meraxes constructed as part of the Dark-ages Reionization And Galaxy-formation Observables from Numerical Simulation (DRAGONS) project, we predict the angular correlation function (ACF) of LBGs at z ∼ 6 -8. Overall, we find that the predicted ACFs are in good agreement with recent measurements at z ∼ 6 and z ∼ 7.2 from observations consisting of the Hubble eXtreme Deep Field (XDF), the Hubble Ultra-Deep Field (HUDF) and Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey (CANDELS) field. We confirm the dependence of clustering on luminosity, with more massive dark matter haloes hosting brighter galaxies, remains valid at high redshift. The predicted galaxy bias at fixed luminosity is found to increase with redshift, in agreement with observations. We find that LBGs of magnitude M AB(1600) < −19.4 at 6 z 8 reside in dark matter haloes of mean mass ∼ 10 11.0 -10 11.5 M , and this dark matter halo mass does not evolve significantly during reionisation.