Description
Radio cosmology today is driven by high-precision instruments like LOFAR, MeerKAT, and HERA, which are continually improving limits on the 21cm power spectrum. The next-generation instrument SKA will be the largest radio interferometric array, enabling high dynamic range radio maps. However, vast data volumes expand radio inference challenges, with foregrounds up to five orders brighter than the signal. Additionally, ARCADE 2 and LWA suggest an anomalous Radio Synchrotron Background (RSB) excess. This excess aligns with CMB radiation at high frequencies but deviates from a blackbody spectrum at lower frequencies, suggesting an additional astrophysical or exotic source. While various explanations have been proposed, such as modified populations of faint extragalactic sources or new physical processes, there is still no consensus on the origin of this potential signal, making the RSB excess an open question in radio cosmology.
In this talk, I will present a novel Bayesian formalism based on Gibbs sampling that draw statistical samples from a model of the full sky (and its covariance) involving large numbers of free parameters, and conditioned on complex interferometric data. The formalism is based on a visibility response operator that incorporates all the instrument degrees of freedom. It also automatically deals with any gaps in the data. By injecting a model that includes an RSB excess component, we can investigate the possibility of characterising this signal using large interferometric arrays, such as the ones used for 21cm cosmology.
