Extension of the Advanced Spectral Leakage scheme for neutron star merger simulations. (arXiv:2102.08882v1 [astro-ph.HE])
<a href="http://arxiv.org/find/astro-ph/1/au:+Gizzi_D/0/1/0/all/0/1">Davide Gizzi</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Lundman_C/0/1/0/all/0/1">Christoffer Lundman</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+OConnor_E/0/1/0/all/0/1">Evan O&#x27;Connor</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Rosswog_S/0/1/0/all/0/1">Stephan Rosswog</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Perego_A/0/1/0/all/0/1">Albino Perego</a>

We extend a neutrino transport approximation, called Advanced Spectral
Leakage (ASL), with the purpose of modeling neutrino-driven winds in neutron
star mergers. Based on a number of snapshots we gauge the ASL parameters by
comparing against both the two-moment (M1) scheme implemented in the FLASH code
and the Monte Carlo neutrino code Sedonu. The ASL scheme contains three
parameters, the least robust of which results to be a blocking parameter for
electron neutrinos and anti-neutrinos. The parameter steering the angular
distribution of neutrino heating is re-calibrated compared to the earlier work
(arXiv:1906.11494). We also present a new, fast and mesh-free algorithm for
calculating spectral optical depths, which, when using Smoothed Particle
Hydrodynamics (SPH), makes the neutrino transport completely particle-based. We
estimate a speed-up of a factor of $gtrsim 100$ in the optical depth
calculation when comparing to a grid-based approach. In the suggested
calibration we recover luminosities and mean energies within $25%$. A
comparison of the rates of change of internal energy and electron fraction in
the neutrino-driven wind suggests comparable accuracies of ASL and M1, but a
higher computational efficiency of the ASL scheme. We estimate that the ratio
between the CPU hours spent on the ASL neutrino scheme and those spent on the
hydrodynamics is $lesssim 0.8$ per timestep when considering MAGMA2
(arXiv:1911.13093) as source code for the Lagrangian hydrodynamics, to be
compared with a factor of 10 from the M1 in FLASH.

We extend a neutrino transport approximation, called Advanced Spectral
Leakage (ASL), with the purpose of modeling neutrino-driven winds in neutron
star mergers. Based on a number of snapshots we gauge the ASL parameters by
comparing against both the two-moment (M1) scheme implemented in the FLASH code
and the Monte Carlo neutrino code Sedonu. The ASL scheme contains three
parameters, the least robust of which results to be a blocking parameter for
electron neutrinos and anti-neutrinos. The parameter steering the angular
distribution of neutrino heating is re-calibrated compared to the earlier work
(arXiv:1906.11494). We also present a new, fast and mesh-free algorithm for
calculating spectral optical depths, which, when using Smoothed Particle
Hydrodynamics (SPH), makes the neutrino transport completely particle-based. We
estimate a speed-up of a factor of $gtrsim 100$ in the optical depth
calculation when comparing to a grid-based approach. In the suggested
calibration we recover luminosities and mean energies within $25%$. A
comparison of the rates of change of internal energy and electron fraction in
the neutrino-driven wind suggests comparable accuracies of ASL and M1, but a
higher computational efficiency of the ASL scheme. We estimate that the ratio
between the CPU hours spent on the ASL neutrino scheme and those spent on the
hydrodynamics is $lesssim 0.8$ per timestep when considering MAGMA2
(arXiv:1911.13093) as source code for the Lagrangian hydrodynamics, to be
compared with a factor of 10 from the M1 in FLASH.

http://arxiv.org/icons/sfx.gif