arXiv2026
The distribution of radio continuum galaxies is a useful, fast, and accessible probe of the matter distribution in the Universe, enlightening us about the Universe's initial conditions, the physics of dark matter, and the nature of the mysterious dark energy. However, radio continuum galaxies alone cannot easily be localised in the radial direction, and cross-identification of host sources from optical catalogues is challenging across wide area surveys. Moreover, there are several redshift-dependent properties of radio galaxy populations that all need accurate modelling to make reliable inferences about fundamental physics. These include accurate measurements of the redshift distribution of radio sources ($dN/dz$), the coupling between radio galaxies and the underlying matter distribution (quantified by the galaxy bias, $b(z)$), and the true flux distribution $N(S,z)$ of the radio sources (magnification bias). The amount of encoded cosmological information depends on the survey properties and the level of homogeneity across its footprint. In this chapter, we demonstrate the cosmological potential of a 20,000 sq. deg survey with the SKAO in AA4 configuration, using 10,000 hours of observations. Such a survey will reach $\mathcal{O}(μ\mathrm{Jy/beam)}$ sensitivities and detect $\mathcal{O}$(300-400 million) radio sources, the largest sample of radio continuum galaxies to date. This surpasses the number of sources assumed for the previous SKA cosmology Red Book. We predict the angular clustering of such a survey, using mocks accounting for potential telescope systematics, and discuss which data corrections may be needed when these systematics cannot be accurately modelled.}