The Impact of Different Neutrino Transport Methods on Multidimensional Core-collapse Supernova Simulations

TL;DR

This study compares three approximate neutrino transport methods in multidimensional core-collapse supernova simulations, analyzing their effects on supernova dynamics, neutrino properties, and computational costs to inform future modeling choices.

Contribution

It provides a detailed comparison of M1, IDSA, and ASL neutrino transport schemes within the same simulation framework, highlighting their advantages, disadvantages, and impacts on supernova outcomes.

Findings

Transport schemes have minor effects on PNS radius, mass, and accretion rate.

Neutrino luminosities and shock radii differ by 10-20%.

Differences are more pronounced in gain region properties and PNS convection.

Abstract

Neutrinos play a crucial role in the core-collapse supernova (CCSN) explosion mechanism. The requirement of accurately calculating the transport of neutrinos makes simulations of the CCSN mechanism extremely challenging and computationally expensive. Historically, this stiff challenge has been met by making approximations to the full transport equation. In this work, we compare CCSN simulations in one- and two-dimensions with three approximate neutrino transport schemes, each implemented in the {\tt FLASH} simulation framework. We compare a two moment M1 scheme with an analytic closure (M1), the Isotropic Diffusion Source Approximation (IDSA), and the Advanced Spectral Leakage (ASL) method. We identify and discuss the advantages and disadvantages of each scheme. For each approximate transport scheme, we use identical grid setups, hydrodynamics, and gravity solvers to investigate the transport effects on supernova shock dynamics and neutrino quantities. We find that the transport scheme has a small effect on the evolution of protoneutron star (PNS) radius, PNS mass, and the mass accretion rate. The neutrino luminosities, mean energies, and shock radii have a $\sim$ 10-20\% quantitative difference but the overall qualitative trends are fairly consistent between all three approximations. We find larger differences in the gain region properties, including the gain region mass and the net heating rate in the gain region, as well as the strength of PNS convection in the core. We investigate the progenitor, nuclear equation of state, and stochastic perturbation dependence of our simulations and find similar magnitudes of impact on key quantities. We also compare the computational expense of the various approximations.