Kavli Affiliate: Nickolay Y. Gnedin
| First 5 Authors: Nickolay Y. Gnedin, , , ,
| Summary:
I compare the power spectra of the radiation fields from two recent sets of
fully-coupled simulations that model cosmic reionization: "Cosmic Reionization
On Computers" (CROC) and "Thesan". While both simulations have similar power
spectra of the radiation sources, the power spectra of the photoionization rate
are significantly different at the same values of cosmic time or the same
values of the mean neutral hydrogen fraction. However, the power spectra of the
photoionization rate can be matched at large scales for the two simulations
when the matching snapshots are allowed to vary independently. I.e., on large
scales, the radiation field in two simulations proceeds through the same
evolutionary stages, but the timing of these stages is different in different
simulations and is not parameterized by an easily interpretable physical
quantity like the mean neutral fraction or the mean free path. On small scales,
large differences are present and remain partially unexplained.
Both CROC and Thesan use the Variable Eddington Tensor approximation for
modeling radiative transfer, but adopt different closure relations (optically
thin OTVET versus M1). The role of this key difference is tested by using
smaller simulations with a new cosmological simulation code that implements
both closure relations in a controlled environment (the same hydro, cooling,
and gravity solvers and the star formation recipe). In these controlled tests,
both the M1 closure and the OTVET ansatz follow the expected behavior from a
simple analytical approximation, demonstrating that the differences in the
2-point function of the radiation field induced by the choice of the Eddington
tensor are not dominant.
| Search Query: ArXiv Query: search_query=au:”Nickolay Y. Gnedin”&id_list=&start=0&max_results=3