The Hydrostatic Mass Bias and the $σ_8$ Tension: A Multi-Probe Forecast for Stage-IV/V Surveys

Abstract

The hydrostatic mass bias ($b_{\mathrm{HSE}}$) is a leading systematic uncertainty in cluster cosmology and a principal source of degeneracy with $σ_8$ and $Ω_m$. We investigate the capability of Stage-IV CMB and optical surveys to calibrate $b_{\mathrm{HSE}}$ using tomographic cross-correlations between the thermal Sunyaev--Zel'dovich (tSZ) effect, galaxy clustering, and weak lensing. We perform a Fisher forecast incorporating realistic survey noise, foreground modeling for clustered CIB and radio sources, and full marginalization over cosmological and astrophysical nuisance parameters, including per-bin galaxy bias perturbations, photometric redshift shifts, intrinsic alignments, and baryonic feedback modeled with HMCode2020. With optimized tomographic binning (10 lens and 5 source bins for LSST; 6 lens and 5 source bins for CSST), we forecast marginalized constraints of $0.98\%$ for SO+LSST, $1.60\%$ for CMB-S4+LSST, and $2.40\%$ for CMB-S4+CSST. Tomography improves $b_{\mathrm{HSE}}$ precision by factors of approximately three relative to non-tomographic analyses, reflecting the role of redshift information in breaking the $b_{\mathrm{HSE}}$--$σ_8$ degeneracy. Optical-only probes provide no direct constraint on $b_{\mathrm{HSE}}$, whereas inclusion of tSZ-containing spectra enables percent-level calibration under realistic systematic assumptions. The results demonstrate that multi-probe tomographic analyses with Stage-IV surveys can achieve robust control of hydrostatic mass bias, strengthening cluster-based constraints on structure growth.

Figures (7)

• Figure 1: Stacked panels showing the sensitivity of each probe and cross-probe combination of the angular power spectra ($\mathcal{D}_\ell= \ell(\ell+1)\mathcal{C}_\ell/(2\pi)$) for auto- and cross-correlations of the thermal Sunyaev-Zel’dovich (tSZ, $y$), galaxy clustering ($g$), and cosmic shear ($\gamma$) to the amplitude of matter density fluctuations ($\sigma_8$), with all other cosmological parameters (e.g., $\Omega_m, h, n_s$) held fixed. APS and CPS represents angular power spectrum and cross-power spectrum, respectively.
• Figure 2: Normalized redshift distributions for LSST Year 10 and CSST. All distributions satisfy $\int n(z)\,dz = 1$, where $n(z)$ denotes the normalized redshift distribution (total: $n(z)$, tomographic bins: $n(z|\mathrm{bin})$). Distributions include photometric redshift uncertainties convolved via the error-function formalism of Ma06: $\sigma_z = 0.05(1{+}z)$ (LSST sources), $\sigma_z = 0.03(1{+}z)$ (LSST lenses), and CSST values from mission forecasts ($\sigma_z = 0.05(1{+}z)$ sources, $\sigma_z = 0.03(1{+}z)$ lenses; CSST). Top row (Overplotted Totals): Direct comparison of total distributions. Left: CSST sources extend to $z \approx 2.8$ with stronger high-$z$ counts than LSST (spanning $z \approx 0$--3.5). Right: Both lens distributions peak at $z \approx 0.5$, with CSST extending to $z \approx 1.8$ (vs. LSST’s focus on $z \approx 0$--1.5). Middle row (LSST Binned): Left: 5 equal-count source bins (edges from 1% peak threshold, ensuring uniform statistical weight). Right: 10 fixed-width lens bins ($z = 0.2$--1.2, $\Delta z = 0.1$) consistent with LSST DESC conventions DESCSRD. Bottom row (CSST Binned): Left: 5 source bins optimized for tSZ--shear and cosmic shear cross-correlations. Right: 6 lens bins optimized for galaxy--tSZ and clustering cross-correlations, leveraging CSST’s superior photo-$z$ precision. LSST distributions use parameters from the Core Cosmology Library (CCL) Chisari19 while CSST distributions are obtained from simulations following Xiong25).
• Figure 3: Analysis pipeline for forecasting hydrostatic mass bias ($b_{\mathrm{HSE}}$) constraints. The workflow progresses from survey specification through tomographic/non-tomographic analysis of multi-probe power spectra to final marginalized cosmological constraints. Tomographic analyses achieve high-precision $b_{\mathrm{HSE}}$ constraints for SO+LSST (0.98% precision), with competitive results for CMB-S4+LSST (1.60%) and CMB-S4+CSST (2.40%).
• Figure 4: Constraints on hydrostatic mass bias ($b_{\mathrm{HSE}}$) from Fisher matrix forecasts of CMB-optical cross-correlations, comparing CMB-S4 + CSST, CMB-S4 + LSST, and SO + LSST. Main panel: $68\%$ and $95\%$ confidence contours in the $\sigma_8$--$b_{\mathrm{HSE}}$ plane. Solid/dashed curves denote tomography (CSST: 6+5 bins, LSST: 10+5 bins)/no-tomography configurations. Color coding: CMB-S4 + CSST (blue), CMB-S4 + LSST (orange), SO + LSST (purple). The black star marks the fiducial cosmology ($\sigma_8 = 0.811$, $b_{\mathrm{HSE}} = 0.2$), consistent with the CMB-S4 Science Book′s adopted cluster mass model. Inset: Relative $b_{\mathrm{HSE}}$ precision (1$\sigma$ uncertainty as percentage of fiducial value). SO + LSST tomography achieves 0.98% precision on $b_{\mathrm{HSE}}$, corresponding to sub-percent precision, and representing a $79.7\%$ reduction in the marginalized 1$\sigma$ uncertainty relative to its no-tomography case (4.80% $\rightarrow$ 0.98%).
• Figure 5: Multi-parameter constraints from cross-correlation analyses combining CMB and optical surveys. The corner plot shows 1$\sigma$ and 2$\sigma$ confidence contours for five selected cosmological and astrophysical parameters: the matter fluctuation amplitude $\sigma_8$, hydrostatic mass bias $b_{\mathrm{HSE}}$, matter density $\Omega_m$, Hubble constant $h$, and intrinsic alignment amplitude $A_{\mathrm{IA}}$. Four survey configurations are compared: CMB-S4 + CSST without tomography (blue dashed), CMB-S4 + CSST with tomography (blue solid), CMB-S4 + LSST with tomography (orange solid), and SO + LSST with tomography (purple solid). The significant tightening of constraints with tomography demonstrates the power of redshift binning in breaking parameter degeneracies, particularly for $b_{\mathrm{HSE}}$ which shows precision improvements from 29% (CMB-S4+CSST) to 79.7% (SO+LSST) with tomographic analyses.