First Light And Reionization Epoch Simulations (FLARES) – I. Environmental dependence of high-redshift galaxy evolution

Abstract

We introduce the First Light And Reionisation Epoch Simulations (Flares), a suite of zoom simulations using the Eagle model. We resimulate a range of overdensities during the Epoch of Reionisation (EoR) in order to build composite distribution functions, as well as explore the environmental dependence of galaxy formation and evolution during this critical period of galaxy assembly. The regions are selected from a large (3.2 cGpc)3 parent volume, based on their overdensity within a sphere of radius 14 h−1 cMpc. We then resimulate with full hydrodynamics, and employ a novel weighting scheme that allows the construction of composite distribution functions that are representative of the full parent volume. This significantly extends the dynamic range compared to smaller volume periodic simulations. We present an analysis of the galaxy stellar mass function (GSMF), the star formation rate distribution function (SFRF) and the star forming sequence (SFS) predicted by Flares, and compare to a number of observational and model constraints. We also analyse the environmental dependence over an unprecedented range of overdensity. Both the GSMF and the SFRF exhibit a clear double-Schechter form, up to the highest redshifts (z = 10). We also find no environmental dependence of the SFS normalisation. The increased dynamic range probed by Flares will allow us to make predictions for a number of large area surveys that will probe the EoR in coming years, carried out on new observatories such as Roman and Euclid.

First Light And Reionisation Epoch Simulations (FLARES) II: The Photometric Properties of High-Redshift Galaxies

“…There is also a similar trend at = 10, but is harder to draw conclusions from, since the Oesch et al ( 2018) data contains only 5 galaxies. Some of this tension can be −8 attributed to the slightly lower normalisation (∼ 0.3 dex) of the SFR function of the E reference volume or F at intermediate SFR (1 < SFR < 10 M yr −1 ) as noted for high-redshift galaxies in Katsianis et al (2017); Lovell et al (2021a, also see Furlong et al 2015) when compared to observed values. Vijayan et al (2021) also showed that the unobscured SFR density of F galaxies at ∈ [5, 7] showed slightly lower normalisation (∼ 0.2 dex) in comparison with the unobscured value from Bouwens et al (2020), even though the dust model was explicitly calibrated to match the UV LF, indicating that either the star formation rates are generally lower in the simulation, or the chemical enrichment rate (and thus the derived dust content) is higher, giving rise to higher attenuation than expected in these model galaxies.…”

“…This relation can also be used to understand the increasing dust temperatures with redshift (explicit redshift evolution is shown in Figure 10). Lovell et al (2021a) has already shown that there is a systematic increase in the normalisation of the sSFR of the F galaxies at constant stellar mass. This would imply that the redshift dependence of the dust temperature can be attributed to the increasing sSFR.…”

“…= 4.7 and have a radius of 14 cMpc/h, spanning a wide range of overdensities, from = −0.479 → 0.970 (see Table A1 in Lovell et al 2021a). We have deliberately selected a large number of extreme overdense regions ( 16) to obtain a large sample of massive galaxies.…”

A machine learning approach to mapping baryons on to dark matter haloes using the eagle and C-EAGLE simulations

“…One means of circumventing this limitation is to use multiple zoom simulations of differing environments, and weight the relative abundance of each simulation based on its relative total matter overdensity. This technique was first demonstrated with the GIMIC simulations (Crain et al 2009), and recently used in the Flares simulations to make predictions for the abundance of galaxies during the epoch of reionisation (Lovell et al 2021;Vijayan et al 2021). Another drawback is that zooms cannot be used to self-consistently predict aspects of the large scale structure, such as the clustering of galaxies, since they are by construction non-representative, small volume regions of the Universe.…”

“…This will allow for the construction of extremely large lightcones (as demonstrated in Hearin et al 2020, using their empirical modelling plus simulationcalibrated approach), necessary for making predictions for wide field surveys from the upcoming Roman and Euclid space-based observatories(Potter et al 2017). To this end, in future work we will explore predictions during the epoch of reionisation, where we will leverage the Flares simulations(Lovell et al 2021). A unique aspect of Flares is that it consists of resimulations of a range of overdensities, providing training data in extreme over-and under -dense environments, which may aid predictions of galaxy properties across all environments.…”