Baryon acoustic oscillations from HI intensity mapping: the importance of cross-correlations in the monopole and quadrupole

TL;DR

This study demonstrates that combining multipole expansion and cross-correlation techniques significantly enhances the detection of baryon acoustic oscillations in HI intensity mapping data, even with beam limitations.

Contribution

It introduces a multipole expansion approach and highlights the benefits of cross-correlating HI intensity maps with galaxy surveys to improve BAO detection.

Findings

Achieved ~4.5σ BAO detection in auto-correlation with beam effects.

Cross-correlation increased BAO detection to ~6σ.

Including quadrupole doubles BAO detection significance.

Abstract

Cosmological parameter estimation in the post-reionisation era via neutral hydrogen radio emission ($\text{\hi}$), is among the key science goals of the forthcoming SKA Observatory (SKAO). This paper explores detection capability for baryon acoustic oscillations (BAO) with a suite of \(100\) simulations introducing the main limitations from foreground contamination and poor angular resolution caused by the radio telescope beam. Such broad single-dish beam representing a serious challenge for BAO detection with \hi\ intensity mapping, we investigate a multipole expansion approach as a means for mitigating such limitations. We also showcase the gains made from cross-correlating the \hi\ intensity mapping data with an overlapping spectroscopic galaxy survey, aiming to test potential synergies between the SKA Project and other future cosmological experiments at optical/near-infrared wavelengths. For our ${\sim}\,4\,000\,\mathrm{deg}^2$ data set at $z\,{=}\,0.9$, replicating the essential features of an SKAO \hi\ intensity mapping survey, we were able to achieve a ${\sim}\,4.5\sigma$ detection of BAO features in auto-correlation despite the dominant beam effect. Cross-correlation with an overlapping galaxy survey can increase this to a ${\sim}\,6\sigma$ detection. Furthermore, including the power spectrum quadrupole besides the monopole in a joint fit can approximately double the BAO detection significance. Despite not implementing a radial-only $P(k_\parallel)$ analysis in favour of the three-dimensional $P(\bm{k})$ and its multipoles, we were still able to obtain robust constraints on the radial Alcock-Paczynski parameter, whereas the perpendicular parameter remains unconstrained and prior dominated due to beam effects.