Investigating Gas Turbulence in Galactic Discs

Abstract: The interstellar medium of both local and distant disc galaxies is observed to be supersonically turbulent and the turbulent motion of the gas is thought to be important for galaxy formation and evolution, particularly as a key ingredient for star formation. Consequently, this generates a striking relation between the star formation rate (SFR) and gas velocity dispersion ($\sigma_{\rm g}$; a measurement on the level of turbulence). However, there is no consensus on the origin of this turbulence and its relation with star formation is not yet understood. The two main features of this relation is a plateau of $\sigma_{\rm g}\sim 10\ {\rm km\ s^{-1}}$ for SFR $\lesssim2-3\ {\rm \,M}_\odot\,{\rm yr}^{-1}$ and a rapid increase of $\sigma_{\rm g}\gtrsim 100\ {\rm km\ s^{-1}}$ for SFR $\gtrsim2-3\ {\rm \,M}_\odot\,{\rm yr}^{-1}$. These very turbulent gas discs are mainly found in high redshift galaxies, which are observed on the spatial scale of several kpc and are, thus, poorly resolved. This also makes it difficult to accurately correct for beam smearing, which is an observational effect that increases the velocity dispersion due to mixing with the velocity of the disc rotation. Furthermore, they are mainly observed using H$\alpha$ as a tracer, which is dynamically different from e.g. neutral and molecular hydrogen. In this thesis, I determine which of these factors are crucial for shaping the observed features of the $\sigma_{\rm g}-$SFR relation. Furthermore, I investigate the possible origins of gas turbulence in disc galaxies, focusing on two origins popular in literature: gravitational instability and stellar feedback. To achieve all this, I perform state-of-the-art hydro+\textit{N}-body simulations of entire galactic discs with a range of characteristics and feedback processes. I find simulated and observed $\sigma_{\rm g}-$SFR relations to be in excellent agreement. However, the simulated galaxies only reach $\sigma_{\rm g}\lesssim 50\ {\rm km\ s^{-1}}$ while observational data suggest $\sigma_{\rm g}\lesssim 150\ {\rm km\ s^{-1}}$. I show that by considering the warm gas phase, $\sigma_{\rm H\alpha} \sim 130\ {\rm km\ s^{-1}}$ can be reproduced, which highlights the heterogeneity of the data, as different tracers have different kinematics. Furthermore, I demonstrate how even low levels of beam smearing can lead to severe overestimates of the observed turbulent velocity dispersions, a notion that calls into question results from high redshift galaxy observations. I show that the \sigmasfr remains unchanged when removing stellar feedback and that gas turbulence in disc galaxies is driven by a marginally stable disc and that a galactic disc is naturally drawn towards marginal instability. Finally, I present two new analytic equations to predict the level of turbulence in galaxies from the warm ionised gas phase, using simple equipartition arguments, and in the case of a gravitationally unstable disc, by applying a multi-component Toomre's $Q$ equation. The good match of these equations with simulation data encourage the development of new analytic models based on the warm ionised phase and tools such as Toomre's $Q$.