Shear calibration biases in weak lensing surveys

Abstract

We investigate biases induced by the conversion between the observed image shape to shear distortion in current weak lensing analysis methods. Such overall calibration biases cannot be detected by the standard tests such as E/B decomposition or calibration with stars. We find that the non-Gaussianity of point spread function has a significant effect and can lead to up to 15 per cent error on the linear amplitude of fluctuations sigma_8 depending on the method of analysis. This could explain some of the discrepancies seen in recent amplitude determinations from weak lensing. Using an elliptical Laguerre expansion method we develop a re-Gaussianization method which reduces the error to calibration error of order 1 per cent even for poorly resolved galaxies. We also discuss a new type of shear selection bias which results in up to roughly 8 percent underestimation of the signal. It is expected to scale with redshift, inducing errors in the growth factor extraction if not properly corrected for. Understanding and correcting for such effects is crucial if weak lensing is to become a high precision probe of cosmology.