Bayesian Analysis of Galaxy Clusters
Determining physical information from astrophysical observations often requires creating a hypothetical model and applying it to the data. For galaxy clusters we can use models of Sunyaev-Zeldovich radio or bremsstrahlung X-ray emission from the intracluster medium (ICM), or strong lensing of background galaxies to determine cluster properties and profiles like mass and pressure. In X-ray this is typically done by deriving de-projected 3D profiles from fits to the 2D surface brightness map.
Bayes-X is an existing tool that avoids the assumptions of deprojection by generating 3D models of the ICM and using them to simulate observations from a 3D source. Bayesian methods are applied to these simulations to fit the model to observed data and thus determine cluster profiles and properties. In this thesis Bayes-X is tested against real and simulated data and re-implemented to improve performance, usability and ease of development. This new version is then extensively tested to determine if it is a reliable analysis tool. This includes tests against simulated clusters where structure is well known and can be directly compared against results. Possible problems and limitations identified during testing are discussed and mitigations proposed where possible.
Recommendations are made for future development and improvements outside of the scope of this thesis.