Building tools for the modelling of epoch of reionisation galaxies in the era of JWST

One of the Holy Grails of modern astronomy is to observe the first generations of stars and how they assembled into galaxies. While recent years have made incredible progress in this endeavour, we have reached somewhat of a cul-de-sac in terms of the sensitivity and wavelength coverage necessary to push further back into reionisation. We also struggle to make substantive statements about the physical properties of the galaxies we do detect at z > 6. The James Webb Space Telescope (JWST) promises to revolutionise our view of this epoch, however modelling must keep up in order to satisfactorily interpret these new observations. In this thesis I chart the development of a Bayesian SED fitting code interrogator. This code is optimised for maximum flexibility in model choices, so as to test between the suite of existing stellar, nebular and dust models, and also to modularly add new models as they are developed. I use this code on a sample of low redshift Extreme Emission Line Sources (EELS), selected from the GAMA survey as analogues to high redshift galaxies. Specifically I use interrogator to show how the various assumptions implicit in modern SED fitting affect the inferred physical properties of these objects, biasing results on even the most basic properties such as stellar mass and star formation rates. Finally, using the bluetides simulation, I propose a novel technique for modelling the stellar spectra of galaxies in the Epoch of Reionisation (EoR). By relaxing the single metallicity approximation commonly used in SED fitting, I show that improvements can be made to the recovery of these properties at the cost of adding additional parameters to the model.