Development of statistical tools for studies of the astrophysical rapid neutron capture process

Abstract:
The rapid neutron capture process (r-process) is a complex nucleosynthesis mechanism for the creation of heavy nuclei, which occurs under extreme astrophysical conditions, such as binary neutron star mergers and some types of core-collapse supernovae. An accurate understanding of the r-process is crucial for explaining the abundances of roughly half the elements heavier than iron in the solar system. Not only are the predictions of the r-process abundance pattern affected by the thermodynamical conditions of such astrophysical events, significant uncertainty also arises from the properties of thousands of neutron-rich nuclides involved in the process. While many of the neutron-rich nuclei may become experimentally accessible in the near future, it is essential to quantify the uncertainty originating from theoretical descriptions of atomic nuclei and identify key nuclear physics inputs of the numerical simulations of the r-process.

In this thesis, several statistical methods have been applied to tackle the uncertainty of nuclear physics inputs in the studies of the r-process nucleosynthesis. Variance-based sensitivity analysis method identifies influential nuclear physics inputs in a statistically rigorous manner and probes their effect on elemental abundance patterns. Ensemble Bayesian model averaging method provides a simple framework for combining competing theoretical nuclear physics models based on experimental data and quantifying their uncertainty. Furthermore, emulation of r-process abundance calculations has been performed using artificial neural networks, which dramatically speeds up the calculations of abundance patterns, potentially allowing for scaling up various statistical analyses. While the effectiveness of these methods has been shown for the specific features of the observed solar abundance pattern and nuclear physics observables, they are readily applicable to broader aspects of the studies of the r-process nucleosynthesis.