Bayesian Inference and Global Sensitivity Analysis for Ambient Solar Wind Prediction

A comprehensive uncertainty quantification framework that allows for ensemble predictions with reduced uncertainties of the Potential Field Source Surface (PFSS), Wang-Sheeley-Arge (WSA), and Heliospheric Upwind eXtrapolation (HUX) ambient solar wind model chain. The framework proceeds in the following steps: First, we identify the problem space and parameters that have uncertainties. Second, through global sensitivity analysis, we find out which of these parameters significantly impacts the variance of the quantity of interest. Third, we infer the full distribution of the most important parameters through Bayesian inference. Fourth, we make ensemble predictions with these newfound parameter distributions