Reducing and constraining the intrinsic galaxy alignment contamination to weak lensing measurements

Abstract

We present a method for removal of the contaminating effects of intrinsic galaxy alignments, in measurements of cosmic shear from multi-colour weak lensing surveys. The method down-weights pairs which are likely to be close in three dimensions, based on spectroscopic or photometric redshifts. Results are dramatic: the intrinsic contamination of the low-redshift Sloan photometric redshift survey could be 80 times the lensing signal, but this can be almost completely removed, leaving random shot noise errors of the order of 10%. Intrinsic galaxy alignments, although an annoying contaminant for cosmic shear studies, are interesting in their own right, and we therefore present a new observational constraint for their amplitude from the aperture mass B-mode in the Red-Sequence Cluster Survey. The small measured B-mode rules out the published intrinsic alignment models from numerical simulations, which assume no evolution in galaxy clustering.