A tale of two populations: the stellar mass of central and satellite galaxies

We develop a new empirical methodology to study the relation between the stellar mass of galaxies and the mass of their host subhaloes. Our approach is similar to abundance matching, and is based on assigning a stellar mass to each subhalo within a large cosmological N-body simulation. However, we significantly extend the method to include a different treatment for central and satellite galaxies, allowing the stellar mass of satellite galaxies to depend on both the host subhalo mass, and on the halo mass. Unlike in previous studies, our models are constrained by two observations: the stellar mass function of galaxies, and their autocorrelation functions (CFs). We present results for similar to 10(6) different successful models, showing that the uncertainty in the host subhalo mass reaches a factor of similar to 10 for a given stellar mass. Our results thus indicate that the host subhalo mass of central and satellite galaxies is poorly constrained by using information coming solely from the abundance and CFs of galaxies. In addition, we demonstrate that the fraction of stellar mass relative to the universal baryon fraction can vary between different models, and can reach similar to 0.6 for a specific population of low-mass galaxies. We conclude that additional observational constraints are needed, in order to better constrain the mass relation between haloes and galaxies. These might include weak lensing, satellite kinematics or measures of clustering other than the CFs. Moreover, CFs based on larger surveys with a better estimate of the systematic uncertainties are needed.