Investigating the Contribution of Extended Radio Sources to the Epoch of Reionisation Power Spectrum

Abstract

We investigate the contribution of extended radio sources such as Centaurus A, and Galactic supernova remnants (SNRs) to our ability to detect the statistical $21\,\rm{cm}$ signal from the Epoch of Reionisation (EoR) with the Murchison Widefield Array (MWA). These sources are typically ignored because they are in highly attenuated parts of the MWA primary beam, however in aggregate these sources have apparent flux densities of $10\,\rm{Jy}$ on angular scales we expect to detect the $21\,\rm{cm}$ signal. We create bespoke multi-component 2D Gaussian models for Galactic SNRs and for Centaurus A, and simulate the visibilities for two MWA snapshot observations. We grid those visibilities and then Fourier transform them with respect to frequency, averaging them both spherically and cylindrically to produce the 1D and 2D power spectra. We compare the simulated 1D power spectra to the expected $21\,\rm{cm}$ power spectrum. We find that although these extended sources are in highly attenuated parts of the MWA primary beam pattern, collectively they have enough power ($\sim10^4-10^5\,\rm{mK^2}\,\it{h^{-3}} \,\rm{Mpc^{3}}$) on EoR significant modes ($|\mathbf{k}| \leq 0.1 h \rm{Mpc}^{-1}$) to prohibit detection of the $21\,\rm{cm}$ signal ($10^4\,\rm{mK^2}\,\it{h^{-3}} \,\rm{Mpc^{3}}$). We find that $50-90\%$ of sources must be removed in order to reduce leakage to a level of $10-20\%$ of the $21\,\rm{cm}$ power spectrum on EoR significant modes. The effects of widefield extended sources will have implications on the detectability of the $21\,\rm{cm}$ signal for the MWA and with the future Square Kilometre Array (SKA).