Unbiased constraints on reionization model parameters in presence of the foreground wedge

TL;DR

This paper evaluates the effectiveness of clustering wedges in accurately constraining reionization model parameters from 21 cm observations, demonstrating they provide unbiased estimates and help break parameter degeneracies.

Contribution

It demonstrates that clustering wedges yield unbiased constraints on EoR parameters and effectively break degeneracies, validating their use in foreground-affected 21 cm analyses.

Findings

Clustering wedges provide unbiased parameter constraints.

Uncertainties with wedges are comparable to foreground-free data.

Wedges help break degeneracies between model parameters.

Abstract

A possible way to disentangle the redshifted 21 cm signal of neutral hydrogen (HI) in the Epoch of Reionization (EoR) from the much larger astrophysical foregrounds is to restrict the analysis to a wedge-shaped region in the Fourier space. The foregrounds are confined to this wedge because of their smooth spectral properties which allows one to estimate the HI power spectrum in the foreground-free portion, known as the reionization window. The estimate of the spherically averaged power spectrum in the window, however, differs from the true value because of the line of sight anisotropies in the signal. The difference can be estimated and taken into account for a given reion- ization model by expanding the power spectrum in terms of the Legendre polynomials. This corrected power spectrum, called as the clustering wedges, can then be used for comparing the model predictions with the observations. In this work, we carefully ex- amine whether the clustering wedges are appropriate for such model comparisons and whether they provide truly unbiased constraints on the EoR parameters. We use the excursion-set based semi-numerical simulations, coupled with MCMC-based statistical methods, for our analysis. We find that the clustering wedges not only yield faithful constraints, the statistical uncertainties on the parameter values too are comparable to those obtained using the data without the presence of any foregrounds. We also find that the clustering wedges are helpful in breaking the degeneracies between different EoR model parameters.