Testing PSF Interpolation in Weak Lensing with Real Data

Abstract

Reconstruction of the point spread function (PSF) is a critical process in weak lensing measurement. We develop a real-data based and galaxy-oriented pipeline to compare the performances of various PSF reconstruction schemes. Making use of a large amount of the CFHTLenS data, the performances of three classes of interpolating schemes -polynomial, Kriging, and Shepard -are evaluated. We find that polynomial interpolations with optimal orders and domains perform the best. We quantify the effect of the residual PSF reconstruction error on shear recovery in terms of the multiplicative and additive biases, and their spatial correlations using the shear measurement method of Zhang et al. (2015). We find that the impact of PSF reconstruction uncertainty on the shear-shear correlation can be significantly reduced by cross correlating the shear estimators from different exposures. It takes only 0.2 stars (SNR ∼ > 100) per square arcmin on each exposure to reach the best performance of PSF interpolation, a requirement that is satisfied in most of the CFHTlenS data.

Weak Lensing for Precision Cosmology

“…The second is interpolating to other positions so the PSF model can be used to measure galaxy photometry and shapes (for discussion of several PSF interpolation methods, see e.g. Bergé et al 2012;Gentile, Courbin & Meylan 2013;Kitching et al 2013;Lu et al 2017). Challenges in PSF modeling LSST (Baumer, Davis & Roodman 2017) and interpolation differ for ground-and space-based imaging.…”

Testing Shear Recovery with Field Distortion

“…As already demonstrated in Lu et al (2017), for the CFHTLenS data, a chip-wise pixel-by-pixel spatial interpolation of the PSF power spectra with the 1st or 2nd order polynomial functions is the best way of reconstructing the PSF field for Fourier Quad. Our discussion here therefore only focuses on how to select out bright stars (typically with SNR ∼ > 100) from bright sources.…”

The three-year shear catalog of the Subaru Hyper Suprime-Cam SSP Survey

“…It is important to ensure that residual systematic errors are well below the statistical error floor so that any physical constraints obtained from the weak lensing measurements are not biased. Observationally there are several sources of systematic effects inherent in characterizing galaxy shapes, even in a statistical sense: (i) "noise bias" due to the non-linear impact of noise on shear estimation (Refregier et al 2012;Zhang & Komatsu 2011); (ii) "model bias" due to imperfect assumptions about galaxy morphology (e.g., Bernstein 2010); (iii) "weight bias" caused by shear-dependent weighting (e.g., Fenech Conti et al 2017); (iv) "selection bias" originating from an improper treatment of selection effects around cuts (e.g., Mandelbaum et al 2005); (v) systematics related to blending of galaxy light profiles (e.g., Li et al 2018;Sheldon et al 2020); (vi) mis-estimation of the point-spread function (PSF; e.g., Lu et al 2017); and (vii) other systematics from detector non-idealities -e.g., "tree rings", "edge distortions" (Plazas et al 2014), and brighter-fatter effects (Antilogus et al 2014)and from the atmosphere -e.g., differential chromatic refraction (DCR; Plazas & Bernstein 2012). There are other astrophysical uncertainties such as photometric redshift errors, intrinsic alignments of galaxy shapes and the impact of baryonic effects (Mandelbaum 2018).…”