Validation of the DESI-DR1 3x2-pt analysis: scale cut and shear ratio tests

TL;DR

This paper validates a joint DESI-DR1 3x2pt analysis by propagating uncertainties in the non-linear matter power spectrum, galaxy bias, and baryon feedback to design scale cuts that keep cosmological biases below $0.3\sigma$ in the $S_8$-$\Omega_m$ plane, and by employing shear-ratio tests to check internal consistency. Using DESI-DR1 with KiDS-1000, DES-Y3, HSC-Y1, and HSC-Y3, the authors develop a consistent analytic covariance and models for cosmic shear, galaxy clustering, and galaxy-galaxy lensing, including IA and magnification terms. They find that a scale cut of $(R_{\rm ggl},R_{\rm clus})=(6,8)\,h^{-1}\mathrm{Mpc}$ satisfies the bias criteria across surveys (with KiDS-1000 requiring a minor shaping adjustment), and that shear-ratio measurements coherently fit the same model as cosmic shear while improving constraints on source redshift distributions and IA parameters. The SR results remain robust across alternative non-linear models and cosmologies, demonstrating the framework’s readiness for the forthcoming full $3\times2$pt cosmological parameter analysis and underscoring the SR test as a valuable tool for future large surveys.

Abstract

Combined survey analyses of galaxy clustering and weak gravitational lensing (3x2-pt studies) will allow new and accurate tests of the standard cosmological model. However, careful validation is necessary to ensure that these cosmological constraints are not biased by uncertainties associated with the modelling of astrophysical or systematic effects. In this study we validate the combined 3x2-pt analysis of the Dark Energy Spectroscopic Instrument Data Release 1 (DESI-DR1) spectroscopic galaxy clustering and overlapping weak lensing datasets from the Kilo-Degree Survey (KiDS), the Dark Energy Survey (DES), and the Hyper-Suprime-Cam Survey (HSC). By propagating the modelling uncertainties associated with the non-linear matter power spectrum, non-linear galaxy bias and baryon feedback, we design scale cuts to ensure that measurements of the matter density and the amplitude of the matter power spectrum are biased by less than 30% of the statistical error. We also test the internal consistency of the data and weak lensing systematics by performing new measurements of the lensing shear ratio. We demonstrate that the DESI-DR1 shear ratios can be successfully fit by the same model used to describe cosmic shear correlations, and analyse the additional information that can be extracted about the source redshift distributions and intrinsic alignment parameters. This study serves as crucial preparation for the upcoming cosmological parameter analysis of these datasets.

Figures (11)

• Figure 1: The footprints on the sky of weak lensing surveys: KiDS-1000 (red), DES-Y3 (green), HSC-Y1 (light blue), and HSC-Y3 (dark blue), and their overlap with the DESI-DR1 dataset (dark grey). The two hemispheres are centred on the North Galactic Pole (NGP) and South Galactic Pole (SGP) regions. The overall footprint of the DESI survey is shaded as light grey.
• Figure 2: The average angular density per unit redshift of the DESI-DR1 lens samples used in our analysis, spanning three Bright Galaxy Survey (BGS) and three Luminous Red Galaxy (LRG) tomographic bins.
• Figure 3: The weighted angular density per unit redshift of the weak lensing source datasets used in our analysis, spanning different tomographic samples for the KiDS-1000, DES-Y3, HSC-Y1 and HSC-Y3 datasets. The "spikes" in the redshift distributions, which appear for some surveys, are caused by sample variance when these distributions are estimated from calibration samples which only span small areas of sky.
• Figure 4: The parameter bias between our baseline and contaminated data vectors assuming a scale cut of $6 \, h^{-1} \, \mathrm{Mpc}$ for galaxy-galaxy lensing and $8 \, h^{-1} \, \mathrm{Mpc}$ for galaxy clustering. For cosmic shear, we use the fiducial scale cuts of each weak lensing survey, except for a minor update to KiDS-1000 as noted in the Section \ref{['sec:method_valid']}. The purple and green shaded regions show the $0.3 \, \sigma$ contour for the 2D marginalized constraints in the sub-space of $(\Omega_\mathrm{m}, S_8)$ for fits to the baseline and contaminated data vectors, respectively. The different rows show results for KiDS-1000, DES-Y3, HSC-Y1 and HSC-Y3, and the columns display fits to the $1 \times 2$-pt data (cosmic shear only), $2 \times 2$-pt data (galaxy-galaxy lensing and projected clustering), and to the full $3 \times 2$-pt data vector.
• Figure 5: The measurements of the angle-averaged shear ratio from DESI-DR1 lenses and KiDS, DES-Y3, HSC-Y1 and HSC-Y3 sources. The averaged ratios are displayed for all the unique combinations of lens and source bins, ordered in groups corresponding to a given lens sample. We overplot the best-fitting model (green solid line) from our baseline analysis, and additionally a model in which we set the values of IA, $\Delta z$, $m$ and magnification equal to zero (the blue solid line, which we refer to as the fiducial model).