Logo Logo
Switch language to English
Chung, Andrew S. (2016): The long and winding road: lyman-alpha radiative transfer and galactic outflows. Dissertation, LMU München: Fakultät für Physik



The standard model of cosmology has been extremely successful in explaining observations of the universe. However, bringing the standard model predictions for the distribution of galaxies into agreement with observations relies critically on the invocation of feedback processes to regulate galaxy formation via galactic outflows. As such, there is intense interest in understanding and modelling the underlying physical processes involved in these outflows. In this thesis we present two new numerical Lyα radiative transfer codes, and apply them in a two-pronged approach to understanding galactic outflows driven by stellar feedback. In our first approach we identify the first systematic failing of basic ‘shell model’ outflows – an inability to produce the ‘blue bumps’ seen in the spectra of recent observations. We then present, and test with numerical radiative transfer simulations, a minor extension to the standard shell model paradigm which leads naturally to the production of blue bumps via Fermi-like acceleration of Lyα photons. In this way we paper over the cracks that were starting to show in shell models, and allow them to remain consistent with observations for the present time. However, we also cast one eye to the future, where we expect that at some point we will be forced to abandon such simple shell models. We therefore pursue an approach whereby we attempt to numerically simulate the hydrodynamical process of galaxy formation in a cosmological context. As a further step we then perform numerical radiative transfer to derive the observable properties of the formed galaxy. In this way we are able to test galactic outflow models against observations of similar systems. We find that a combination of the three observables which we simulate (Lyα emission, absorption, and spectral shape) provides a strong constraint on current outflow models, and ultimately motivates the development of better models.