# Impact of Foregrounds on HI Intensity Mapping Cross-Correlations with Optical Surveys

## TL;DR

This paper investigates how 21cm foreground removal affects HI intensity mapping cross-correlations with optical surveys, finding minimal impact for spectroscopic surveys but significant degradation for photometric surveys with high redshift uncertainties, and proposes solutions for real data application.

## Contribution

It introduces a reconstruction method to mitigate line-of-sight amplitude changes caused by foreground cleaning, improving cross-correlation reliability with photometric surveys.

## Key findings

- Foreground cleaning minimally affects spectroscopic cross-correlations.
- Photometric surveys with high redshift uncertainties experience significant signal degradation.
- Reconstruction techniques can restore accurate redshift distributions in cross-correlation analyses.

## Abstract

The future of precision cosmology could benefit from cross-correlations between intensity maps of unresolved neutral hydrogen (HI) and more conventional optical galaxy surveys. A major challenge that needs to be overcome is removing the 21cm foreground emission that contaminates the cosmological HI signal. Using N-body simulations we simulate HI intensity maps and optical catalogues which share the same underlying cosmology. Adding simulated foreground contamination and using state-of-the-art reconstruction techniques we investigate the impacts that 21cm foregrounds and other systematics have on these cross-correlations. We find that the impact a FASTICA 21cm foreground clean has on the cross-correlations with spectroscopic optical surveys with well-constrained redshifts is minimal. However, problems arise when photometric surveys are considered: we find that a redshift uncertainty {\sigma}_z {\geq} 0.04 causes significant degradation in the cross power spectrum signal. We diagnose the main root of these problems, which relates to arbitrary amplitude changes along the line-of-sight in the intensity maps caused by the foreground clean and suggest solutions which should be applicable to real data. These solutions involve a reconstruction of the line-of-sight temperature means using the available overlapping optical data along with an artificial extension to the HI data through redshift to address edge effects. We then put these solutions through a further test in a mock experiment that uses a clustering-based redshift estimation technique to constrain the photometric redshifts of the optical sample. We find that with our suggested reconstruction, cross-correlations can be utilized to make an accurate prediction of the optical redshift distribution.