Astrophysical Modeling and Inference Limitations in Pulsar Timing Arrays: A Case Study

Recent pulsar timing array (PTA) experiments have found evidence for a nano-hertz gravitational wave background (GWB), prompting extensive investigation into its possible origins. While many studies have performed parameter estimation for proposed sources, few assess the statistical viability of their models for inference from PTA data. A complete analysis of the source of the GWB needs to understand the degeneracy and saturation point of its model parameters— where more data no longer improves constraints—as well as the extent of prior-dominance of the parameters in the Bayesian analyses. In this talk, I highlight my study of the statistical viability of a widely used semi-analytic population model of supermassive black hole binaries. The study has strong implications for the limitations of astrophysical modeling and inferences, framing the expectations for the near and far future of PTAs.