Understanding the results of galaxy-galaxy lensing using galaxy-mass correlation in numerical simulations

McKay et al. have recently used measurements of weak galaxy-galaxy lensing in the Sloan Digital Sky Survey to estimate the cross-correlation between galaxies and the projected dark matter density field. They have derived a relation between aperture mass within a radius of 260 h (-1) kpc, M (260), and lens galaxy luminosity that does not depend on galaxy luminosity, type or environment. In this paper, we study the cross-correlation between galaxies and dark matter using galaxy catalogues constructed from a high-resolution N -body simulation of a LambdaCDM universe. We show that the simulation reproduces the observational results reasonably well, except that the predicted mass-to-light ratio is about a factor of 2 too high. In the simulation, M (260) is approximately equal to the halo virial mass for L * galaxies. M (260) overestimates the virial mass for fainter galaxies and underestimates it for brighter galaxies. We use the simulations to show that under certain circumstances the halo virial mass may be recovered by fitting a Navarro-Frenk-White (NFW) profile to the projected galaxy-mass correlation function. If we apply our method to the observations, we find that L * galaxies typically reside in haloes of similar to2 x 10(12) h (-1) M., consistent with halo masses estimated from the observed Tully-Fisher relation. In the simulations, the halo virial mass scales with galaxy luminosity as L (1.5) for central galaxies in haloes and for galaxies in low-density regions. For satellite galaxies, and for galaxies in high-density regions, there is no simple relation between galaxy luminosity and halo mass and care must be exercised when interpreting the lensing results.