Parametrizations of the global 21-cm signal and parameter estimation from single-dipole experiments
Geraint J. A. Harker (1), Jordan Mirocha (2), Jack O. Burns (2) and, Jonathan R. Pritchard (3) ((1) University College London, (2) University of, Colorado, (3) Imperial College London)

TL;DR
This paper introduces a physically motivated parametrization of the global 21-cm signal for single-dipole experiments, improving bias reduction and physical plausibility in signal extraction from low-frequency radio data.
Contribution
It proposes a new tanh-based parametrization of the 21-cm signal that better captures physical features and reduces biases compared to previous methods.
Findings
Tanh parametrization outperforms previous models in accuracy.
Biases can be reduced by discarding less robust information.
Method is suitable for experiments with 1000-hour observations.
Abstract
One approach to extracting the global 21-cm signal from total-power measurements at low radio frequencies is to parametrize the different contributions to the data and then fit for these parameters. We examine parametrizations of the 21-cm signal itself, and propose one based on modelling the Lyman-alpha background, IGM temperature and hydrogen ionized fraction using tanh functions. This captures the shape of the signal from a physical modelling code better than an earlier parametrization based on interpolating between maxima and minima of the signal, and imposes a greater level of physical plausibility. This allows less biased constraints on the turning points of the signal, even though these are not explicitly fit for. Biases can also be alleviated by discarding information which is less robustly described by the parametrization, for example by ignoring detailed shape information…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
