FPFS Shear Estimator: Systematic Tests on the Hyper Suprime-Cam Survey First Year Data

TL;DR

This paper applies the FPFS shear estimator to Hyper Suprime-Cam survey data, demonstrating its low bias and consistency with other methods, while thoroughly testing for systematics and calibration uncertainties.

Contribution

It provides a systematic validation of the FPFS shear estimator on real survey data, including bias calibration, PSF residual analysis, and comparison with re-Gaussianization methods.

Findings

FPFS shear estimator has less than 1% multiplicative bias without blending.

Shear calibration uncertainties are well-characterized and controlled.

Weak lensing measurements from FPFS and re-Gaussianization catalogs are consistent.

Abstract

We apply the Fourier Power Function Shapelets (FPFS) shear estimator to the first year data of the Hyper Suprime-Cam survey to construct a shape catalog. The FPFS shear estimator has been demonstrated to have multiplicative bias less than $1\%$ in the absence of blending, regardless of complexities of galaxy shapes, smears of point spread functions (PSFs) and contamination from noise. The blending bias is calibrated with realistic image simulations, which include the impact of neighboring objects, using the COSMOS Hubble Space Telescope images. Here we carefully test the influence of PSF model residual on the FPFS shear estimation and the uncertainties in the shear calibration. Internal null tests are conducted to characterize potential systematics in the FPFS shape catalog and the results are compared with those measured using a catalog where the shapes were estimated using the re-Gaussianization algorithms. Furthermore, we compare various weak lensing measurements between the FPFS shape catalog and the re-Gaussianization shape catalog and conclude that the weak lensing measurements between these two shape catalogs are consistent with each other within the statistical uncertainty.