The Quantification of Incertitude in Astrophysical Simulation Codes

Abstract:  We present a study of methodologies for the propagation of epistemic uncertainty, also known as incertitude, in complex astrophysical simulations. We chose the community simulation instrument MESA (Modules for Experiments in Stellar Astrophysics) and simulated the evolution of stars with a ZAMS mass of one solar mass. We explored the case of incertitude in stellar wind parameters, specifically parameters employed to model stellar winds during the red giant and asymptotic giant branch phases of evolution. These parameters are inputs to MESA, and we chose uncertainty intervals for each. Treating MESA as a ``black box,\" we applied two incertitude propagation techniques, Cauchy deviates and quadratic response surface models, to obtain bounds for white dwarf masses at the cessation of thermonuclear burning. These methodologies are applicable to other computational incertitude propagation problems.