Weak Lensing Measurements: A Revisited Method and Application to Hubble Space Telescope Images

Abstract

The weak distortions produced by gravitational lensing in the images of background galaxies provide a method to measure directly the distribution of mass in the universe. However this technique requires high precision measurements of the lensing shear and cautious corrections for systematic effects. Kaiser, Squires, & Broadhurst (1995) proposed a method to calibrate the ellipticity-shear relation in the presence of Point Spread Function (PSF) anisotropies and camera distortions. We revisit the KSB method and show that both the PSF and the camera distortions can be corrected for using source moments, as opposed to ellipticities. We clarify the applicability of some of the approximations made in this method. We derive expressions for the corrections which only involve the galaxy moments. We derive an explicit relation between the shear and the average ellipticity. We discuss the shortcomings of the method, and test its validity using numerical simulations. As an application of the method, we repeat the analysis of the HST WFPC2 camera performed by Hoekstra et al. (1998). We confirm the presence of sizable (10%) PSF ellipticities at the edge of the WFPC2 chips. We also show that the PSF ellipticity varies by as much as 2% over time. We use these measurements to correct the shape of galaxies in the HST Survey Strip (``Groth'' Strip). By considering the dependence of the ellipticities on object size, we show that, after corrections, the residual systematic uncertainty for galaxies with radii greater than 0.15 arcsec, is about 0.4%, when averaged over each chip. We discuss how these results provide good prospects for measuring weak lensing by large-scale structure with deep HST surveys.