Tidal alignments as a contaminant of redshift space distortions

We investigate the effect of orientation-dependent selection effects on galaxy clustering in redshift space. It is found that if galaxies are aligned by large-scale tidal fields, then these selection effects give rise to a dependence of the observed galaxy density on the local tidal field, in addition to the well-known dependences on the matter density and radial velocity gradient. This alters the galaxy power spectrum in a way that is different for Fourier modes parallel to and perpendicular to the line of sight. These tidal galaxy alignments can thus mimic redshift space distortions, and thus result in a bias in the measurement of the velocity power spectrum. If galaxy orientations are affected only by the local tidal field, then the tidal alignment effect has exactly the same scale and angular dependence as the redshift space distortions in the linear regime, so it cannot be projected out or removed by masking small scales in the analysis. We consider several toy models of tidal alignments and orientation-dependent selection, normalize their free parameter (an amplitude) to recent observations, and find that they could bias the velocity amplitude f(z)G(z) by 5--10 per cent in some models, although most models give much smaller contamination. We conclude that tidal alignments may be a significant systematic error in redshift space distortion measurements that aim to test general relativity via the growth of large-scale structure. We briefly discuss possible mitigation strategies.