H i intensity mapping with MeerKAT: primary beam effects on foreground cleaning

TLDR: In this paper, the authors show that the coupling between beam sidelobes and the foreground structure can complicate the cleaning of the MeerKAT beam and show that this effect is constrained to a narrow region in parallel space and can be reduced if the maps are carefully re-smoothed to a common lower resolution.

Abstract: Upcoming and future neutral hydrogen Intensity Mapping surveys offer a great opportunity to constrain cosmology in the post-reionization Universe, provided a good accuracy is achieved in the separation between the strong foregrounds and the cosmological signal. Cleaning methods rely on the frequency smoothness of the foregrounds and are often applied under the assumption of a simplistic Gaussian primary beam. In this work, we test the cleaning in the presence of a realistic primary beam model with a non trivial frequency dependence. We focus on the Square Kilometre Array precursor MeerKAT telescope and simulate a single-dish wide area survey. We consider the main foreground components, including an accurate full sky point source catalogue. We find that the coupling between beam sidelobes and the foreground structure can complicate the cleaning. However, when the beam frequency dependence is smooth, we show that the cleaning is only problematic if the far sidelobes are unexpectedly large. Even in that case, a proper reconstruction is possible if the strongest point sources are removed and the cleaning is more aggressive. We then consider a non-trivial frequency dependence: a sinusoidal type feature in the beam width that is present in the MeerKAT beam and is expected in most dishes, including SKA1-MID. Such a feature, coupling with the foreground emission, biases the reconstruction of the signal across frequency, potentially impacting the cosmological analysis. Interestingly, such contamination is present at a lower level even when no point sources are included and the beam is Gaussian, showing that this frequency ripple can be problematic even within the main lobe. We show that this effect is constrained to a narrow region in $k_\parallel$ space and can be reduced if the maps are carefully re-smoothed to a common lower resolution.