Making it Big: The Effect of Dust on Galaxy Morphology – by Lamiya Mowla

Galaxy morphology is one of the fundamental and oldest observational tools used to study the formation and evolution of galaxies. Decades of observations from the ground and thousands of orbits of extragalactic imaging by the Hubble Space Telescope (HST) have constructed a picture for the relation between the sizes and masses of galaxies, which has been used to piece together the puzzle of galaxy evolution over the last 10 Gyrs. However, it is impossible to complete this puzzle using observations alone, making cosmological hydrodynamical simulations with multi-scale models of physical processes a critical deciphering tool. Despite the success of simulations in largely matching observed galaxy scaling relations, they have yet to reproduce the observed size evolution of galaxies; the sizes of galaxies are not consistent with observations, nor across different simulations. Is this because the simulations are making galaxies with the wrong sizes? Or is it because they have not yet incorporated the effect of the major culprit that is distorting our observations of galaxies – dust? In this talk I will present the effect of dust attenuation on galaxy morphology using state-of-the-art cosmological hydrodynamical simulation SIMBA with dust radiative transfer package Powderday. If true, this might remarkably change the picture of the galaxy size-mass relation as painted by HST. Resolved stellar population synthesis modelling, augmented by upcoming JWST observation, will be needed to verify our revised view of the size-mass relation of galaxies.

Searching for dark matter in the Lyman-alpha forest – by Keir Rogers

The search for dark matter offers the possibility for rich new physics that deepens our understanding of the Universe. The ultra-light axion is a compelling candidate that is motivated, e.g., by the string theory “axiverse” and as a possible solution to the so-called “small-scale crisis” of the cold dark matter model. I will present new, robust bounds on the axion mass that improve by an order of magnitude over previous studies and significantly exclude the canonical mass scale of 10^-22 eV. The bounds exploit new spectroscopic observations of high-redshift (z > 5) quasars and absorption features formed in the intergalactic medium (IGM) — called the “Lyman-alpha forest”.

In order to address the challenge of robustly marginalising the uncertain astrophysics of the IGM, I will present a new framework for dark matter bounds using cosmological data. This utilises a novel “emulator” for the effect of dark matter models on the power spectrum. I will discuss how the application of active learning (Bayesian emulator optimisation) can ensure robust, converged parameter inference from a limited number of expensive simulations.