Shear recovery accuracy in weak-lensing analysis with the elliptical Gauss-Laguerre method

We implement the elliptical Gauss-Laguerre (EGL) galaxy-shape measurement method proposed by Bernstein & Jarvis and quantify the shear recovery accuracy in weak-lensing analysis. This method uses a deconvolution fitting scheme to remove the effects of the point-spread function (PSF). The test simulates >10(7) noisy galaxy images convolved with anisotropic PSFs and attempts to recover an input shear. The tests are designed to be immune to statistical ( random) distributions of shapes, selection biases, and crowding, in order to test more rigorously the effects of detection significance (signal-to-noise ratio [S/N]), PSF, and galaxy resolution. The systematic error in shear recovery is divided into two classes, calibration ( multiplicative) and additive, with the latter arising from PSF anisotropy. At S/N > 50, the deconvolution method measures the galaxy shape and input shear to similar to 1% multiplicative accuracy and suppresses > 99% of the PSF anisotropy. These systematic errors increase to similar to 4% for the worst conditions, with poorly resolved galaxies at S/N similar or equal to 20. The EGL weak-lensing analysis has the best demonstrated accuracy to date, sufficient for the next generation of weak-lensing surveys.