The reliability of the kinematical evidence for dark matter: the effects of non-sphericity, substructure and streaming motions

Using cosmological N-body simulations of dark matter haloes, we study the effects of nonsphericity, substructure and streaming motions in reproducing the structure and internal kinematics of clusters of galaxies from kinematical measurements. Fitting a Navarro, Frenk & White (NFW) model to the 3D density profile, we determine the virial mass, concentration parameter and velocity anisotropy of the haloes, and then calculate the profiles of projected velocity moments, as they would be measured by a distant observer. Using these mock data, we apply a Jeans analysis for spherical objects to reproduce the line-of-sight velocity dispersion and kurtosis profiles and fit the three parameters. We find that the line-of-sight velocity dispersion and kurtosis profiles of a given halo can vary considerably with the angle of view of the observer. We show that the virial mass, concentration parameter and velocity anisotropy of the haloes can be reproduced satisfactorily independently of the halo shape, although the virial mass tends to be underestimated, the concentration parameter overestimated, and the recovered anisotropy is typically more tangential than the true one. The mass, concentration and velocity anisotropy of haloes are recovered with better precision when their mean velocity profiles are near zero.