Helioseismic and neutrino data-driven reconstruction of solar properties

Abstract

In this work we use Bayesian inference to quantitatively reconstruct the solar properties most relevant to the solar composition problem using as inputs the information provided by helioseismic and solar neutrino data. In particular, we use a Gaussian process to model the functional shape of the opacity uncertainty to gain flexibility and become as free as possible from prejudice in this regard. With these tools we first readdress the statistical significance of the solar composition problem. Furthermore, starting from a composition unbiased set of standard solar models we are able to statistically select those with solar chemical composition and other solar inputs which better describe the helioseismic and neutrino observations. In particular, we are able to reconstruct the solar opacity profile in a data driven fashion, independently of any reference opacity tables, obtaining a 4% uncertainty at the base of the convective envelope and 0.8% at the solar core. When systematic uncertainties are included, results are 7.5% and 2% respectively. In addition we find that the values of most of the other inputs of the standard solar models required to better describe the helioseismic and neutrino data are in good agreement with those adopted as the standard priors, with the exception of the astrophysical factor $S_{11}$ and the microscopic diffusion rates, for which data suggests a 1% and 30% reduction respectively. As an output of the study we derive the corresponding data driven predictions for the solar neutrino fluxes.