Abstract :
Magnetic fields are a fundamental component of the interstellar medium (ISM), with their dy- namical role in disk galaxies and their coupling to turbulence remaining a key open question. In this thesis, we investigate galactic magnetism through high-resolution, galaxy-scale magne- tohydrodynamical simulations conducted with the adaptive-mesh refinement code RAMSES.
We developed simulations of Milky Way–like galaxies that include self-gravity, star forma- tion, and supernova feedback, together with non-equilibrium chemistry relevant for following H2 formation and therefore more accurately modeling star formation. Two contrasting initial magnetic field morphologies were considered: a fully ordered (toroidal) configuration and a fully random field. Using these models, we studied the dynamical importance of magnetic fields through the plasma β parameter and the magnetic field strength–density (B–ρ) relation.
We find that unlike the classical picture of a flat B–ρ relation at low densities that turns into a power-law at high densities, in our simulations the B–ρrelation exhibits large intrinsic scatter and strong temporal and spatial variability, with slopes that are neither universal nor static. Disordered magnetic fields are more efficiently amplified by turbulent motions than ordered fields, leading to localized enhancements in magnetic pressure. On galactic scales, however, these differences remain modest and are largely obscured by turbulence driven by stellar feedback, differential rotation, and other large-scale gas flows.
We also examined how the properties of turbulent flows changed by changing the scale on which the mean or laminar flow is defined. We did this using a spherical filtering method, decomposing the magnetic and velocity fields into mean and fluctuating components while sys- tematically varying the averaging scale. We analyzed the resulting energy ratios and power spectra as functions of time and averaging radius, and showed that the inferred turbulent en- ergy fractions increased with averaging scale. We found that, independently of the averaging scale, turbulent kinetic energy exceeds turbulent magnetic energy by one to two orders of mag- nitude, indicating that turbulence in these systems is primarily driven by gas motions rather than magnetic forces. These results have direct implications for observational techniques that rely on scale-dependent field decompositions.
Overall, this thesis demonstrates that the apparent dynamical importance of magnetic fields in disk galaxies depends sensitively on spatial scale and temporal variability. While magnetic fields influence local ISM structure and dense gas evolution, they do not dominate the global energy budget in the systems studied, highlighting the need for scale-aware diagnostics when interpreting galactic magnetism.
