There is a long and rewarding history in astronomy of studying the motions of stars, gas, and planets to learn about the distribution of mass in the universe. Observations of nearby galaxies rotating faster than otherwise expected are one of the pillars of evidence for the cold dark matter cosmological model. As increasingly precise predictions are made about the small scale distribution of mass in the universe, we need tools to be able to weigh our beliefs about the mass content of galaxies and how they can be updated by increasingly richer and more complete observations. In this thesis I present a series of applications of a Bayesian hierarchical model for the equilibrium dynamics of spherical galaxies. Some key features of the model are: a natural mechanism for propagating the systematic uncertainty of parameters such as orbital anisotropy, distance, and stellar mass-to-light ratio into estimates for the dark matter content of galaxies, the ability to jointly model multiple tracer populations with heterogeneous data, and a clear way to incorporate and examine prior assumptions about galaxy and dark matter scaling relations.

Globular Clusters (GCs) are powerful tools for understanding the formation of galaxies. GCs are located in the halos of galaxies and, due to their age and density, have borne witness to the major formation events of a galaxy's lifetime. One may study these objects using a wide array of techniques and datasets, including wide-field ground-based imaging, deep space-based imaging, and spectroscopy. All approaches involve tradeoffs, and in this work we consider a variety of ways to study GC systems in imaging data. We examine a wide-field HST/ACS mosaic of the nearby lenticular galaxy NGC 3115, selecting a high-quality GC sample using the superior resolution of the ACS data. We find strong color bimodality in the GC system of NGC 3115 and examine a number of trends in the properties of the GC system. Next, we consider the situation where one is limited to ground-based imaging, where contaminants to the GC population are a major concern. We detail a novel statistical methodology in which we treat the GC population and the contaminant population as a mixture model, and evaluate the model in a Bayesian context. We demonstrate the performance of the model on mock data, and note some areas where current analysis of GC systems may be missing information using traditional selection techniques. We also apply this Bayesian methodology to a subset of SLUGGS survey galaxies with high-quality photometry from either the MegaCam instrument on the Canada France-Hawaii Telescope or the SuprimeCam instrument on the Subaru Telescope. In most cases, the mixture model recovers the GC system well, often finding the traditional bimodality and providing well-calibrated statistical uncertainties for the global parameters of the GC system. Finally, we examine the object NGC 3628 UCD1, a star cluster slightly more massive than the largest GCs. We identify that UCD1 is located in a stellar stream around the galaxy NGC 3628, and therefore is in the process of being accreted. We characterize UCD1 both in wide-field SuprimeCam imaging and in Keck/ESI spectroscopy, and identify a number of interesting parallels between UCD1 and omegaCen, the largest Milky Way GC.

Background

Methods/design

Discussion

Trial registration

Unlabelled

Conclusion