Dark Energy Survey Year 6 Results: Weak Lensing and Galaxy Clustering Cosmological Analysis Framework

TL;DR

The DES Year 6 paper develops a robust weak lensing and galaxy clustering analysis framework, covering cosmic shear, $2\times2pt$, and $3\times2pt$ in $\Lambda$CDM and $w$CDM, with careful treatment of theoretical systematics including baryonic feedback, galaxy bias, intrinsic alignments, lens magnification, and redshift calibration. It introduces a comprehensive modeling pipeline (matter power spectrum via HMCode, bias via Eulerian perturbation theory and HEFT, IA models NLA/TATT-4, lens magnification, and mode-projected redshifts) and a fully analytic covariance (CosmoCov) incorporating Gaussian, non-Gaussian, SSC, and survey geometry effects. The authors implement scale cuts to control small-scale modeling uncertainties, validated with Cardinal and synthetic datasets, and perform Bayesian inference with Nautilus, providing forecasts and demonstrating strong gains over DES Year 3, while highlighting projection effects in multi-probe analyses. The work establishes a practical framework and methodological blueprint for upcoming photometric surveys (e.g., LSST, Euclid, Roman), emphasizing the balance between robust modeling, computational efficiency, and scalable validation to maximize cosmological returns from weak lensing and clustering data.

Abstract

We present the methodology for the weak lensing and galaxy clustering analyses of the Dark Energy Survey (DES) Year 6 data set. In this work, we design and validate the analysis pipeline for the cosmic shear, galaxy clustering plus galaxy$-$galaxy lensing ($2 \times 2$pt), and the joint analysis in the $3 \times 2$pt. Our framework accounts for key theoretical uncertainties, such as baryonic feedback and galaxy bias, incorporating both linear and non-linear models. We apply scale cuts in regimes where theoretical modeling becomes unreliable. The robustness of the pipeline is validated using mock data and simulations, confirming unbiased cosmological constraints and highlighting the importance of posterior projection effects in the validation process. As a result, we deliver robust and validated analysis pipelines for cosmic shear, $2 \times 2$pt, and $3 \times 2$pt in $Λ$CDM and $w$CDM scenarios, including a well-defined set of scales suitable for real data analysis, a robust prescription for theoretical systematics, and the theoretical covariance of the signal. This comprehensive methodology also lays the groundwork for future galaxy surveys such as the Vera C. Rubin Observatory Legacy Survey of Space and Time.

Figures (22)

• Figure 1: Estimated redshift distributions for the lens or clustering catalog, divided into six redshift bins (pinks), presented in giannini2025darkenergysurveyyear and the lensing efficiency kernel for each of the four source redshift bins (blues), presented in yin2025, demonstrating that the DES Year 6 data is most sensitive between $z=0.1 - 1.0$.
• Figure 2: Comparison of the accuracy of Hm20 and Hf dark-matter-only matter correlation function predictions against the Eemu estimates over the redshift range $0 < z < 1.5$. The comparison is shown for two different cosmological models: the fiducial cosmology adopted in this work ($\Omega_{\rm m} = 0.31$, $\sigma_8 = 0.76$) and an alternative cosmology within the emulator range ($\Omega_{\rm m} = 0.33$, $\sigma_8 = 0.88$). Smallest scale reached with the linear galaxy bias analysis is $R_{\rm Lin} = 6$ Mpc/$h$, while for the non-linear galaxy bias analysis $R_{\rm NL} = 4$ Mpc/$h$. For the cosmic shear analysis, the smallest scale considered is $R_{\rm shear} = 2.8$ Mpc/$h$.
• Figure 3: Impact of different modeling assumptions on the covariance matrix, shown as the ratio of the standard deviation extracted from the modified covariance to that from the reference covariance. The reference covariance is computed using a Halofit dark matter-only matter power spectrum, neglecting intrinsic alignments, and including survey geometry effects. The ratio is plotted as a function of the data point index, from left to right: cosmic shear, galaxy--galaxy lensing, and galaxy clustering. Solid translucent lines display errorbars for the 1000 data points in the data vector, while circles denote the 660 points entering in the linear galaxy bias analysis after applying scale cuts.
• Figure 4: Cosmic shear 2PCFs measured in four redshift bins and 20 logarithmic angular bins over the range $2.5 < \theta < 250$ arcmin. The figure shows the noiseless mock signal for three scenarios: the low-baryon (dark matter only) case with Eemu matter power spectrum (black), the fiducial case obtained with Hm20 and $\log_{10} T_{\mathrm{AGN}}$ = 7.7 (orange), and the high-baryon scenario (Bahamas 8.0) with Eemu (green). The lower panels display the residuals of each signal with respect to the low-baryon case. The grey shaded region and vertical dashed lines indicate data points excluded from the analysis due to unmodeled baryonic feedback and nonlinearities in the matter power spectrum, corresponding to the optimized cosmic shear analysis under the NLA and TATT intrinsic alignment parameterizations, respectively. The pairs of indices on each panel denote the the correlated source redshift bins.
• Figure 5: Galaxy-galaxy lensing 2PCFs estimates in four source redshift bins and six lens bins, measured in 20 logarithmic angular bins over the range $2.5 < \theta < 250$ arcmin. The figure shows the noiseless mock signals both with linear (blue) and non-linear galaxy bias (orange) contributions following the fiducial Fast-pt formalism. Additionally, alternative Aemulus-Heft galaxy bias signals mimicking the MagLim (yellow) and Redmagic (red) samples, are displayed, along with the Cardinal (magenta) measurement. The lower panels show the residuals of each signal with respect to the non-linear galaxy bias baseline case. The grey shaded region and vertical dashed line indicate data points excluded from the analysis due to unmodeled galaxy bias, baryonic feedback, and non-linearities in the matter power spectrum, corresponding to scales of 6 and 4 Mpc$/h$. The pairs of numbers within each panel denote the indices of the correlated source and lens redshift bins.