Title:The Limits of Cosmic Shear

Abstract:In this paper we discuss the commonly-used approximations for two-point cosmic shear statistics. We discuss the four most prominent assumptions in this statistic: the flat-sky, tomographic, Limber and configuration-space approximations, that the vast majority of cosmic shear results to date have used simultaneously. Of these approximations we find that the flat-sky approximation suppresses power by >1% on scales of l<100 and the standard Limber approximation implementation enhances power by >5% on scales l<100; in doing so we find an l-dependent factor that has been neglected in analyses to date. To investigate the impact of these approximations we reanalyse the CFHTLenS 2D correlation function results. When using all approximations we reproduce the result that measurements of the matter power spectrum amplitude are in tension with measurements from the CMB Planck data: where a conditional value of sigma8=0.789 +/- 0.015 is found from CFHTLenS and sigma8=0.830 +/- 0.015 from Planck. When we do not use the Limber and flat-sky approximations we find a conditional value of sigma8=0.801 +/- 0.016 from the CFHTLenS data, significantly reducing the tension between Planck CMB results and lensing results from CFHTLenS. When including the additional effect of expected photometric redshift biases we find sigma8=0.839 +/- 0.017 which is consistent with Planck. We also discuss the impact on CMB lensing. For Euclid, LSST, and WFIRST and any current or future survey none of these approximations should be used.