The Evolution of 21-cm Structure (EOS): public, large-scale simulations of Cosmic Dawn and Reionization

TL;DR

This paper introduces the EOS project, providing large-scale, public 21-cm cosmological simulations of the Cosmic Dawn and Reionization, to aid in data analysis and interpretation of upcoming observations.

Contribution

It presents the 2016 EOS data release with the largest public 21-cm simulations, including calibrated prescriptions and predictions for reionization duration and observational detectability.

Findings

Reionization duration estimated between 2.7 and 5.7 in redshift.

Large-scale 21-cm power can vary by up to a factor of 10 depending on the model.

Forecasts indicate SKA1-low and HERA can detect the cosmic signal with sufficient integration.

Abstract

We introduce the Evolution of 21-cm Structure (EOS) project: providing periodic, public releases of the latest cosmological 21-cm simulations. 21-cm interferometry is set to revolutionize studies of the Cosmic Dawn (CD) and epoch of reionization (EoR), eventually resulting in 3D maps of the first billion years of our Universe. Progress will depend on sophisticated data analysis pipelines, which are in turn tested on large-scale mock observations. Here we present the 2016 EOS data release, consisting of the largest (1.6 Gpc on side with a 1024^3 grid), public 21-cm simulations of the CD and EoR. We include calibrated, sub-grid prescriptions for inhomogeneous recombinations and photo-heating suppression of star formation in small mass galaxies. We present two simulation runs that approximately bracket the contribution from faint unseen galaxies. From these two extremes, we predict that the duration of reionization (defined as a change in the mean neutral fraction from 0.9 to 0.1) should be between 2.7 < Delta z < 5.7. The large-scale 21-cm power during the advanced EoR stages can be different by up to a factor of ~10, depending on the model. This difference has a comparable contribution from: (i) the typical bias of sources; and (ii) a more efficient negative feedback in models with an extended EoR driven by faint galaxies. We also make detectability forecasts. With a 1000h integration, HERA and SKA1-low should achieve a signal-to-noise of ~few-hundreds throughout the EoR/CD, while in the maximally optimistic scenario of perfect foreground cleaning, all instruments should make a statistical detection of the cosmic signal. We also caution that our ability to clean foregrounds determines the relative performance of narrow/deep vs. wide/shallow surveys expected with SKA1. Our 21-cm power spectra, simulation outputs and visualizations are publicly available.