2004/10/31 by A. Challinor, Anthony Challinor, G. Chon +1
Mathematics · Physics and Astronomy · #Anisotropy #Astrophysics #Astrophysics and Cosmic Phenomena #Computational physics #Cosmic background radiation #Cosmic microwave background #Cosmology and Gravitation Theories #Covariance #Covariance function #Estimator #Mathematics #Optics #Physics #Polarization (electrochemistry) #Radio Astronomy Observations and Technology #Spectral density #Statistics #astro-ph
paper · pdf · doi:10.1111/j.1365-2966.2005.09076.x
published as Mon.Not.Roy.Astron.Soc.360:509-532,2005 · 26 pages, 10 figures (some reduced-resolution postscript; better version of Fig. 4 also included with submission but not embedded). Minor changes (and typos fixed) to match published (MNRAS) version
openalex publication_date 2005/06/21 · arxiv created 2005/06/23 · arxiv updated 2009/12/01 · openalex created_date 2016/06/24 · openalex updated_date 2026/08/05
Quadratic methods with heuristic weighting (e.g. pseudo-Cl or correlation function methods) represent an efficient way to estimate power spectra of the cosmic microwave background (CMB) anisotropies and their polarization. We construct the sample covariance properties of such estimators for CMB polarization, and develop semi-analytic techniques to approximate the pseudo-Cl sample covariance matrices at high Legendre multipoles, taking account of the geometric effects of mode coupling and the mixing between the electric (E) and magnetic (B) polarization that arise for observations covering only part of the sky. The E–B mixing ultimately limits the applicability of heuristically weighted quadratic methods to searches for the gravitational-wave signal in the large-angle B-mode polarization, even for methods that can recover (exactly) unbiased estimates of the B-mode power. We show that for surveys covering 1 or 2 per cent of the sky, the contribution of E-mode power to the covariance of the recovered B-mode power spectrum typically limits the tensor-to-scalar ratio that can be probed with such methods to ∼0.05.