Theory at a Crossroads: An In-depth Look at Simulations of Galaxy Interactions

Abstract

While cosmological simulations capture a wealth of information regarding how galaxies evolve in their large-scale environment, idealized simulations can achieve high spatial, temporal, and mass resolution. The subgrid physics models which lead to the successes of cosmological simulations are developed and perfected using idealized simulations. Idealized simulations remain valuable but need to be updated according to the findings of cosmological simulations and modern observations. The primary concern of this thesis is to explore the limitations of idealized simulations and provide suggestions to improve the methodology. Observations have shown that the gas discs of spiral galaxies are always the same size or larger than stellar discs. Despite this, most idealized simulations of galaxy interactions employ equal-sized discs. I present a series of experiments which investigate the consequences of this assumption: the magnitude and efficiency of inflow is affected by a confluence of structural and orbital parameters. Idealized simulations are informed by observational catalogues. These typically use the projected separation and tidal features to identify merging systems, both of which are subject to biases. To assess these biases, I create a sample of interacting pairs from IllustrisTNG. I generate mock observations of the simulated pairs and use both observational techniques and the full cosmological data to determine that ∼45% of these pairs are visually identifiable as interacting. In this work, I show that local merger samples constructed from stellar features are likely to be incomplete and biased toward certain environments. I then use the merger sample to perform a series of tests that assess the validity of the Keplerian (ideal) approximation. Many aspects are consistent with cosmological simulation, however accretion onto the halo provides a non-negligible amount of mass and momentum which has significant effects on galaxies’ trajectories. I provide distributions of infall conditions as a primer for future idealized simulations, and additionally present a case study that tests the proposed methodology. Under certain circumstances, the idealized prescription is able to predict orbital parameters such as the time of first pericenter.