Multi-scale Simulations of the Universe

Team: Justin Read (PI); Oscar Agertz (CoI; project lead); Ramon Rey-Raposo

Galaxy formation is one of the most important unsolved problems in modern astrophysics. One of the key outstanding puzzles has become known as the over-cooling problem. Numerical simulations produce galactic discs that are much denser, and rotate far more quickly than real galaxies in the Universe. This likely owes to poorly constrained processes on small scales (~< 1 parsec) – star formation and stellar feedback – that return energy and momentum to the galactic interstellar medium (ISM). Even though the spatial resolution in cosmological galaxy formation simulations today has reached the scale of massive giant molecular clouds (GMCs) (10-100 pc, e.g. Agertz et al. 2009), star formation physics remains 'sub-grid' (i.e. below the resolution limit of the simulation). This forces important, but poorly understood, concepts such as the efficiency of star formation and stellar feedback to become a set of tuneable free parameters (Scannapieco et al. 2012), placing the predictive power of simulations in serious doubt. A similar problem exists in the field of star formation (see state-of-the-field review by Padoan et al. 2013). Here, the largely unknown large scale galactic initial conditions of turbulent molecular clouds hampers our understanding of how massive star forming regions form, evolve and ultimately affect the host galaxy via stellar feedback. Galaxy and star formation are therefore not separable, but rather two aspects of the same multi-scale problem, which so far have been treated as separate phenomena.

In this project we aim to significantly improve upon our understanding of star and galaxy formation by adopting a novel “zoom-in” approach, where state-of-the-art numerical models of star forming regions will be resolved in their proper galactic context, hence connecting the scales for the first time.