A fully data-driven algorithm for accurate shear estimation

• Published in: Astronomy and astrophysics (Berlin) Vol. 656; p. A135
• Main Author: Hoekstra, Henk
• Format: Journal Article
• Language: English
• Published: Heidelberg EDP Sciences 01.12.2021
• Subjects: Algorithms; Anisotropy; Bias; Cosmology; Ellipticity; Galaxies; Large scale structure of the universe; Optimization; Point spread functions; Shear
• ISSN: 0004-6361; 1432-0746; 1432-0746
• DOI: 10.1051/0004-6361/202141670

Weak lensing by large-scale structure is a powerful probe of cosmology if the apparent alignments in the shapes of distant galaxies can be accurately measured. We study the performance of a fully data-driven approach, based on M ETA D ETECTION , focusing on the more realistic case of observations with an anisotropic point spread function (PSF). Under the assumption that PSF anisotropy is the only source of additive shear bias, we show how unbiased shear estimates can be obtained from the observed data alone. To do so, we exploit the finding that the multiplicative shear bias obtained with M ETA D ETECTION is nearly insensitive to the PSF ellipticity. In practice, this assumption can be validated by comparing the empirical corrections obtained from observations to those from simulated data. We show that our data-driven approach meets the stringent requirements for upcoming space- and ground-based surveys, although further optimisation is possible.