The current leading model for cosmology is the ΛCDM model, which assumes a Universe filled with Cold Dark Matter (CDM), and a dark energy component (Λ). The model succesfully predicts the non-linear growth of cosmic structures from the initial density perturbations observed in the Cosmic Microwave Background, to the large scale web of galaxies in the present-day Universe. The ΛCDM predictions agree very well with observations at large cosmic scales but is difficult ot falsify the theory at these scales. We therefore turn to small cosmic scales, the regime where dwarf galaxies live.
Since dwarf galaxies are relatively faint, detailed observations of these systems are limited to dwarf galaxies in our own Local Group, a population which consists mostly of satellites of our own Milky Way and Andromeda. Modelling these systems is extremely challenging, as the properties of dwarf galaxies satellites will be affected by the presence of their parent galaxy. A full model therefore should take into account enough large scale information to accurately model the growth of the parent halo, and should at the same time have a high enough resolution to resolve the small scale physics that governs the evolution of the satellite. Furthermore, it is crucial that these models include an accurate treatment of the hydrodynamics of the interstellar and intergalactic medium, as the complex interplay between star formation and stellar feedback with the surrounding gas is responsible for the properties of the resulting satellite.
To enable future model predictions for the properties of low-mass Local Group dwarf galaxies, I have developed a new simulation code, Shadowfax, which uses a novel moving-mesh hydrodynamical integration scheme. This scheme is more accurate than traditional hydrodynamical schemes used for simulations of galaxy formation and evolution, and is Lagrangian in nature, making it ideally suited for simulations with a high dynamic range in masses and sizes. During my presentation, I will explain the basics of the moving-mesh scheme, and demonstrate its validity with results obtained using Shadowfax. I will conclude with a short note on scaling, and will explain how I plan to make Shadowfax ready for simulations on high-performance computing systems.