Comparative testing of subgrid models for fast neutrino flavor conversions in core-collapse supernova simulations

We investigate key methodologies of Bhatnagar-Gross-Krook subgrid modeling for neutrino fast flavor conversions (FFCs) in core-collapse supernova based on spherically symmetric Boltzmann radiation hydrodynamics simulations. We first examine time integration methods (explicit, implicit, or semi-implicit) and time step control for the subgrid term and then compare various approaches in the literature approximating FFCs in two aspects: (i) angular-dependent survival probability of neutrinos versus simple equipartition condition with a certain baryon mass density threshold and (ii) four-species treatment versus three-species assumption (${\ensuremath{\nu}}_{x}={\overline{\ensuremath{\nu}}}_{x}$). We find that the equipartition condition is reasonable for outgoing neutrinos, but large deviations emerge in the ingoing neutrinos, that has an influence on matter profiles. We also find that the three-species model, in which flavor conversions evolve toward erasing electron neutrino lepton number crossings, behave differently from the four-species models, where heavy leptonic neutrino numbers are appropriately treated in FFC subgrid modeling. In four-species models, we commonly observe noticeable differences between ${\ensuremath{\nu}}_{x}$ and ${\overline{\ensuremath{\nu}}}_{x}$, highlighting the limitation in three-species treatments to study impacts of flavor conversion on neutrino signals. Our result also suggests that FFC models yield lower neutrino heating rate and smaller shock radii compared to cases with no FFC, in agreement with earlier studies employing quantum kinetic neutrino transport. This work provides valuable information toward robust implementation of FFC subgrid model into classical transport and serves as a pilot study for future multidimensional simulations.