Cosmic magnification on multi-catalogue Herschel submillimetre galaxies

TL;DR

This work broadens the cosmological utility of magnification bias by incorporating multiple Herschel submillimetre catalogues (HSPSC, HerMES, HeVICS, HerS) overlapping SDSS spectroscopic lenses, beyond the traditionally used H-ATLAS data. It introduces KDE-based random catalogs to enable robust cross-correlation measurements in inhomogeneous submillimetre surveys and employs MCMC to jointly constrain both Halo Occupation Distribution parameters and cosmological parameters ($Ω_m$, $σ_8$, $h$). The study achieves the first detection of magnification bias for SMGs outside H-ATLAS and finds results for individual catalogues with large uncertainties, while a joint analysis is consistent with $Λ$CDM and aligns more closely with Planck 2018 than prior non-tomographic results. The analysis highlights the current limitation imposed by narrow sub-mm sky coverage and argues that wider, lensing-optimised surveys are needed to access larger scales and tighten cosmological constraints.

Abstract

{Submillimetre galaxies (SMGs) are excellent background sources for magnification-bias studies, but the limited sky coverage in the submillimetre (sub-mm) band constrains their statistical power. Beyond H-ATLAS, Herschel produced additional sub-mm catalogues, though not optimised for spatial statistical lensing analyses.} {Our goal is to refine cosmological constraints from SMG magnification bias by exploiting the full sub-mm sky surveyed by Herschel.} {We expanded the SMG sample by incorporating other Herschel catalogues overlapping SDSS spectroscopic lenses. Random catalogues were generated via kernel density estimation to compute cross-correlations, and Markov Chain Monte Carlo methods were applied to infer astrophysical and cosmological parameters for each catalogue and for the combined dataset.} {We report the first detection of magnification bias in SMGs beyond H-ATLAS, reinforcing the robustness of this observable. Individual Herschel catalogues yield reasonable central values for $Ω_m$ and $σ_8$, although with large uncertainties. The combined analysis, dominated by the more powerful H-ATLAS sample, gives results consistent with $Λ$CDM: $Ω_m = 0.30^{+0.05}_{-0.07}$, $σ_8 = 0.80 (+/- 0.07)$, and $h < 0.80$, in better agreement with \textit{Planck} 2018 than previous non-tomographic studies.} {SMGs are promising tracers for magnification bias, but the narrow sub-mm coverage remains a major limitation. Wider surveys optimised for lensing would enable cross-correlations on larger scales, yielding tighter cosmological constraints.}

Figures (11)

• Figure 1: Redshift distributions of the foreground lenses (SDSS, in grey) and the different background SMG samples: HSPSC is depicted in blue, HerMES in pink, HeVICS in brown and HerS in green. The redshift distribution across different fields within a given catalog is very robust.
• Figure 2: CCF measurements using GAMA survey as lenses and H-ATLAS SMGs as background sources. The data obtained with the new methodology that uses KDE to reproduce the random catalogue is represented in orange. Results from CUE24, where no such treatment was applied, are depicted in dark grey (CCF using the sky zones G09, G12, G15 and SGP) and light grey (CCF eliminating G15 from the analysis).
• Figure 3: CCF estimated using a homogeneous random catalogue (older methodology, shown as empty circles) and a KDE-based random catalogue (filled circles), for the different datasets under study. The lenses are SDSS low-redshift galaxies, and the background sources are SMGs. In the first and second panels, HSPSC data are shown in blue (all available regions) and in purple (the “base” dataset, restricted to regions exclusive to HSPSC). The third panel displays HerMES data in pink, the fourth HeVICS in brown, and the fifth HerS in green. For comparison, results from CUE24 are overlaid in each panel as grey crosses. An offset has been applied to the KDE-based results along the x-axis for clarity.
• Figure 4: Marginalised posterior distributions and probability contours for the cosmological parameters in the in the GAMA/H-ATLAS case. The CCF obtained using KDE for the random catalogue is shown in orange, while the results from CUE24 are depicted in dark grey (including all four sky regions: G09, G12, G15, and SGP) and in light grey (excluding G15 from the analysis).
• Figure 5: Marginalised posterior distribution of the HOD parameters derived from MCMC runs. Upper panel: GAMA and H-ATLAS data. The results obtained with the new methodology, which employs KDE to generate the random catalogue, are shown in orange. In contrast, the results from CUE24, where a homogeneous random catalogue was used, are depicted in dark grey (for the CCF including all sky regions: G09, G12, G15, and SGP) and in light grey (for the CCF excluding G15 from the analysis). Bottom panel: Marginalised posterior distribution of the HOD parameters derived from MCMC runs on the cross-correlation using SDSS as lenses and HSPSC extended (light blue), base (violet), and HerMES (pink) catalogues as sources.