Galaxies going MAD: the Galaxy-Finder Comparison Project

With the ever-increasing size and complexity of fully self-consistent simulations of galaxy formation within the framework of the cosmic web, the demands upon object finders for these simulations have simultaneously grown. To this extent we initiated the Halo-Finder Comparison Project that gathered together all the experts in the field and has so far led to two comparison papers, one for dark matter field haloes, and one for dark matter subhaloes. However, as state-of-the-art simulation codes are perfectly capable of not only following the formation and evolution of dark matter but also accounting for baryonic physics, i.e. gas hydrodynamics, star formation, stellar feedback, etc., object finders should also be capable of taking these additional physical processes into consideration. Here we report - for the first time - on a comparison of codes as applied to the Constrained Local UniversE Simulation (CLUES) of the formation of the Local Group which incorporates much of the physics relevant for galaxy formation. We compare both the properties of the three main galaxies in the simulation (representing the Milky Way, Andromeda and M33) and their satellite populations for a variety of halo finders ranging from phase space to velocity space to spherical overdensity based codes, including also a mere baryonic object finder. We obtain agreement amongst codes comparable to (if not better than) our previous comparisons - at least for the total, dark and stellar components of the objects. However, the diffuse gas content of the haloes shows great disparity, especially for low-mass satellite galaxies. This is primarily due to differences in the treatment of the thermal energy during the unbinding procedure. We acknowledge that the handling of gas in halo finders is something that needs to be dealt with carefully, and the precise treatment may depend sensitively upon the scientific problem being studied.