[Paper Review] CSST WL preparation I: forecast the impact from non-Gaussian covariances and requirements on systematics-control

Ji Yao, Huanyuan Shan+12 more|arXiv (Cornell University)|Apr 10, 2023

Astronomy and Astrophysical Research4 citations

TL;DR

This paper forecasts the impact of non-Gaussian covariances—dominated by super-sample covariance—on cosmic shear constraints from the China Space Station Telescope (CSST), showing a ~1/3 loss in figure-of-merit for matter clustering and ~1/6 for dark energy parameters. It establishes stringent systematics-control requirements (10⁻² to 10⁻³ level) for intrinsic alignment, baryonic feedback, shear bias, and redshift distribution to ensure unbiased cosmology.

ABSTRACT

The precise estimation of the statistical errors and accurate removal of the systematical errors are the two major challenges for the stage IV cosmic shear surveys. We explore their impact for the China Space-Station Telescope (CSST) with survey area $\sim17,500°^2$ up to redshift $\sim4$. We consider statistical error contributed from Gaussian covariance, connected non-Gaussian covariance and super-sample covariance. We find the non-Gaussian covariances, which is dominated by the super-sample covariance, can largely reduce the signal-to-noise of the two-point statistics for CSST, leading to a $\sim1/3$ loss in the figure-of-merit for the matter clustering properties ($σ_8-Ω_m$ plane) and $1/6$ in the dark energy equation-of-state ($w_0-w_a$ plane). We further put requirements of systematics-mitigation on: intrinsic alignment of galaxies, baryonic feedback, shear multiplicative bias, and bias in the redshift distribution, for an unbiased cosmology. The $10^{-2}$ to $10^{-3}$ level requirements emphasize strong needs in related studies, to support future model selections and the associated priors for the nuisance parameters.

Motivation & Objective

• Assess the impact of non-Gaussian covariances—particularly super-sample covariance—on cosmological constraints from CSST cosmic shear surveys.
• Quantify the degradation in signal-to-noise and figure-of-merit due to statistical errors from non-Gaussianity in two-point statistics.
• Establish stringent requirements for systematics mitigation to prevent bias in cosmological parameter estimation.
• Evaluate the role of priors on nuisance parameters in preserving constraining power under systematics.
• Guide future data analysis and survey strategy design to minimize super-sample covariance effects.

Proposed method

• Model the full covariance matrix of cosmic shear two-point functions, including Gaussian, connected non-Gaussian, and super-sample covariance (SSC) contributions.
• Use the CCL (Cosmic Linear Polarization) library to compute non-linear matter power spectra and weak lensing shear power spectra.
• Simulate survey geometry with a realistic redshift distribution $n(z)$ and inhomogeneous galaxy clustering to compute SSC effects.
• Apply Fisher matrix forecasts to estimate signal-to-noise and figure-of-merit degradation under various error sources.
• Assess residual bias in cosmological parameters (e.g., $\sigma_8$-$\Omega_m$, $w_0$-$w_a$) under different levels of unmitigated systematics.
• Test the impact of priors on nuisance parameters by varying their precision and evaluating resulting constraints.

Experimental results

Research questions

• RQ1 To what extent does super-sample covariance degrade the figure-of-merit for cosmological parameters in CSST cosmic shear surveys?
• RQ2 What level of systematics control is required for intrinsic alignment, baryonic feedback, shear multiplicative bias, and redshift distribution bias to avoid significant cosmological bias?
• RQ3 How do priors on nuisance parameters affect the degradation of cosmological constraints under systematics?
• RQ4 How does the inhomogeneous galaxy distribution and survey footprint influence the magnitude of super-sample covariance?
• RQ5 What are the implications of SSC for early-data survey strategy design in CSST to maximize scientific return?

Key findings

• Super-sample covariance reduces the figure-of-merit for matter clustering ($\sigma_8$-$\Omega_m$) by approximately one-third, and for dark energy ($w_0$-$w_a$) by about one-sixth.
• The inclusion of non-Gaussian covariances, especially SSC, significantly degrades signal-to-noise in cosmic shear two-point statistics.
• Systematic error requirements must be at the $10^{-2}$ to $10^{-3}$ level to prevent cosmological bias, which is more stringent than current stage III constraints.
• Weakening priors on nuisance parameters by a factor of ten reduces constraining power by a factor of two, highlighting the need for tight priors.
• High-fidelity simulations and combined observational calibration methods (e.g., self-calibration) are essential for accurate systematics control.
• The survey footprint and inhomogeneous galaxy distribution strongly influence SSC, necessitating optimized survey strategies to minimize its impact.