The Evolution of Substructure in Galaxy, Group and Cluster Haloes III: Comparison with Simulations

Abstract

In a previous paper, we described a new method for including detailed information about substructure in semi-analytic models of halo formation based on merger trees. In this paper, we compare the predictions of our model with results from self-consistent numerical simulations. We find that in general the two methods agree extremely well, particularly once numerical effects and selection effects in the choice of haloes are taken into account. As expected from the original analyses of the simulations, we see some evidence for artificial overmerging in the innermost regions of the simulated haloes, either because substructure is being disrupted artificially or because the group-finding algorithms used to identify substructure are not detecting all the bound clumps in the highest-density regions. Our analytic results suggest that greater mass and force resolution may be required before numerical overmerging becomes negligible in all current applications. We discuss the implications of this result for observational and experimental tests of halo substructure, such as the analysis of discrepant magnification ratios in strongly lensed systems, terrestrial experiments to detect dark matter particles directly, or indirect detection experiments searching for positrons, gamma-rays, neutrinos or other dark matter decay products.