Calibrating baryonic effects in cosmic shear with external data in the LSST era

TL;DR

This work tackles the bias introduced by baryonic feedback on weak-lensing cosmology by coupling LSST-like cosmic shear with external gas tracers (X-ray gas fractions and stacked kSZ) within the Baryo n Correction Model (BCM) framework implemented in the Baccoemu emulator. It quantifies the calibration precision required for BCM parameters, finding that $\sigma(\log_{10} M_c) \lesssim 0.1$ and $\sigma(\log_{10} \eta) \lesssim 0.1$ are needed to avoid significant degradation of $S_8$, with $10-20\%$ level constraints on these parameters corresponding to practical targets. The study shows near-term X-ray data can constrain bound gas with high precision and near-term kSZ data can constrain ejected gas; long-term datasets further tighten these limits, and a joint WL+X-ray+kSZ analysis can reduce degradation factors on $S_8$, $n_s$, and $h$ to around $\sim 1.1$–$1.2$, effectively restoring the WL cosmological power. Nonetheless, some BCM parameters (e.g., stellar and inner-bound gas structure) remain poorly constrained by these tracers, indicating the need for additional probes (tSZ, X-ray cross-correlations) to fully self-calibrate baryonic effects. Overall, multi-tracer external calibration promises substantial gains for Stage-IV WL surveys, enabling robust cosmological inferences in the presence of complex baryonic physics.

Abstract

Cosmological constraints derived from weak lensing (WL) surveys are limited by baryonic effects, which suppress the non-linear matter power spectrum on small scales. By combining WL measurements with data from external tracers of the gas around massive structures, it is possible to calibrate baryonic effects and, therefore, obtain more precise cosmological constraints. In this study, we generate mock data for a Stage-IV weak lensing survey such as the Legacy Survey of Space and Time (LSST), X-ray gas fractions, and stacked kinetic Sunyaev-Zel'dovich (kSZ) measurements, to jointly constrain cosmological and astrophysical parameters describing baryonic effects (using the Baryon Correction Model - BCM). First, using WL data alone, we quantify the level to which the BCM parameters will need to be constrained to recover the cosmological constraints obtained under the assumption of perfect knowledge of baryonic feedback. We identify the most relevant baryonic parameters and determine that they must be calibrated to a precision of $\sim 10$-$20\%$ to avoid significant degradation of the fiducial WL constraints. We forecast that long-term X-ray data from $\sim 5000$ clusters should be able to reach this threshold for the parameters that characterise the abundance of hot virialised gas. Constraining the distribution of ejected gas presents a greater challenge, however, but we forecast that long-term kSZ data from a CMB-S4-like experiment should achieve the level of precision required for full self-calibration.

Figures (13)

• Figure 1: The bound gas fraction and its associated error from mock measurements of $A_{\mathrm{bgas}}$ and $B_{\mathrm{bgas}}$ based on the pivot masses and redshifts from recent X-ray studies, with $M_{\mathrm{bgas}}^{\mathrm{piv}} = 10^{13} M_{\odot}$. The error bars are derived from currently available measurements summarised in Grandis:2023. The black curve shows the Baccoemu parametrisation for the bound gas fraction (Equation \ref{['eq:f_bgas']}) for the fiducial cosmology used to generate the mock measurements. The blue and grey curves show the sensitivity of the model to variations in $\log_{10} M_{\mathrm{c}}$ and $\log_{10} \beta$, respectively. The horizontal band shows the mean cosmological value of the bound gas fraction, defined as $f_{\mathrm{bgas}} = \Omega_{\mathrm{b}} / \Omega_{\mathrm{m}} - f_*$. The variation in the value of $f_{\mathrm{bgas}}$ across the band reflects differences in the stellar fraction, $f_*$, associated with each individual mock data point.
• Figure 2: The CMB temperature shift due to the kSZ effect as a function of the aperture radius $\theta_{\mathrm{d}}$ of the CAP filter and for different feedback strengths. The model profile is convolved with the ACT DR5 f90 beam profile with FWHM = 2.1 arcmin. The error bars are derived from the baseline noise level of the Simons Observatory.
• Figure 3: The marginalised posteriors for $\Omega_{\mathrm{m}}$, $S_8 = \sigma_8 \sqrt{\Omega_{\mathrm{m}}/0.3}$, $h$, $n_{\mathrm{s}}$, and $\Sigma m_\nu$ obtained from LSST-like weak lensing data up to multipoles of $\ell = 2000$. We compare the posteriors obtained under the marginalisation over baryonic effects (black) to the case where the baryonic parameters are kept fixed at the values used to generate the mock data (red). The inner and outer contours show the 95$\%$ and $68\%$ confidence levels, respectively. We marginalise over intrinsic alignments, photometric redshift uncertainties, and multiplicative shape biases in both cases.
• Figure 4: The marginalised posteriors on the baryonic parameters for LSST-like cosmic shear data alone. The inner and outer contours show the 68$\%$ and $95\%$ confidence levels, respectively. We marginalise over intrinsic alignments, photometric redshift uncertainties, and multiplicative shape biases.
• Figure 5: The errors on $S_8$, $h$, and $n_{\mathrm{s}}$, denoted as $\sigma(S_8)$, $\sigma(h)$, and $\sigma(n_{\mathrm{s}})$, respectively, as a function of the error on each baryonic parameter, denoted as $\sigma(\log_{10} \Theta)$ where $\Theta = \{M_{\mathrm{c}}, \eta, \beta, M_1, \theta_{\mathrm{inn}}, \theta_{\mathrm{out}}, M_{\mathrm{inn}}\}$. Here, the triangular and circular points represent the level to which we can constrain each parameter using near-term and long-term data from external tracers of the gas distribution in and around haloes, respectively. These constraints come from X-ray and kSZ data for $M_{\mathrm{c}}$ and $\eta$, respectively.