Smoothed Particle Inference: A Kilo-Parametric Method for X-Ray Galaxy Cluster Modeling

Abstract

We propose an ambitious new method that models the intracluster medium in clusters of galaxies as a set of X-ray-emitting smoothed particles of plasma. Each smoothed particle is described by a handful of parameters including temperature, location, size, and elemental abundances. Hundreds to thousands of these particles are used to construct a model cluster of galaxies, with the appropriate complexity estimated from the data quality. This model is then compared iteratively with X-ray data in the form of adaptively binned photon lists via a two-sample likelihood statistic and iterated via Markov chain Monte Carlo. The complex cluster model is propagated through the X-ray instrument response using direct sampling Monte Carlo methods. With this approach, the method can reproduce many of the features observed in the X-ray emission in a less assumption-dependent way than traditional analyses, and it allows for a more detailed characterization of the density, temperature, and metal abundance structure of clusters. Multi-instrument X-ray analyses and simultaneous X-ray, Sunyaev-Zeldovich (SZ), and lensing analyses are a straightforward extension of this methodology. Significant challenges still exist in understanding the degeneracy in these models and the statistical noise induced by the complexity of the models.