http://arxiv.org/abs/1705.09285
Numerical simulations have become a necessary tool to describe the complex interactions among the different processes involved in galaxy formation and evolution, unfeasible via an analytic approach. The last decade has seen a great effort by the scientific community in improving the sub-grid physics modelling and the numerical techniques used to make numerical simulations more predictive. Although the recently publicly available code GIZMO has proven to be successful in reproducing galaxy properties when coupled with the model of the MUFASA simulations and the more sophisticated prescriptions of the FIRE setup, it has not been tested yet using delayed cooling supernova feedback, which still represent a reasonable approach for large cosmological simulations, for which detailed sub-grid models are prohibitive. In order to limit the computational cost and to be able to resolve the disc structure in the galaxies we perform a suite of zoom-in cosmological simulations with rather low resolution centred around a sub-L* galaxy with a halo mass of $3\times 10^{11}\,\rm M_\odot$ at $z=0$, to investigate the ability of this simple model, coupled with the new hydrodynamic method of GIZMO, to reproduce observed galaxy scaling relations (stellar to halo mass, stellar and baryonic Tully-fisher, stellar mass-metallicity and mass-size). We find that the results are in good agreement with the main scaling relations, except for the total stellar mass, larger than that predicted by the abundance matching technique, and the effective sizes for the most massive galaxies in the sample, which are too small.
A. Lupi, M. Volonteri and J. Silk
Mon, 29 May 17
-90/35
Comments: 15 pages, 14 figures and 1 table; accepted for publication on MNRAS