Astronomers have built one of the largest and most detailed simulations of the early universe yet made, recreating its first 2 billion years and the way dark matter, gas and radiation interacted to produce the earliest stars, galaxies and black holes.
The simulation, called Lumina, was built by researchers at the Center for Astrophysics | Harvard & Smithsonian, and the work is published in Monthly Notices of the Royal Astronomical Society. ScienceDaily carried the announcement with materials provided by the Smithsonian.
Its purpose is interpretive. By comparing Lumina's predictions with observations from the James Webb Space Telescope, the Nancy Roman Space Telescope and other instruments, astronomers can test how well existing theories explain the evolution of the cosmos.
"With JWST, the early universe has become an exciting new frontier in astrophysics," said the lead author, Oliver Zier, an ITC Fellow at the Center for Astrophysics. "To understand what we are seeing with this telescope, we need to translate our theories into detailed predictions."
Lumina is described as the largest cosmological radiation hydrodynamics simulation ever produced. It models a volume more than 100 times larger than earlier comparable projects, spanning a cube approximately 1.5 billion light-years on each side.
A second advance is its handling of helium reionization, the process in which radiation strips electrons from helium atoms. For the first time in a simulation of this kind, that transformation is followed as it unfolds while the model simultaneously accounts for its effect on the surrounding universe.
"For the first time, we can follow along as both hydrogen and helium changed into their modern form across such an enormous volume of the universe, while still resolving the galaxies and other luminous objects driving that transformation," said Aaron Smith, an assistant professor of physics at the University of Texas at Dallas and an author on the paper.
The work required more than 15,000 graphics processing units on Frontier, one of the world's fastest supercomputers, at Oak Ridge National Laboratory in Tennessee.
The team plans to use Lumina to study the movement of gas between galaxies, the influence of radiation left over from the Big Bang, and the emergence of regions where stars formed at especially high rates.
Where simulation and observation agree, confidence in current theory is strengthened. Unexpected differences, the researchers say, may point toward physical processes that scientists have yet to fully understand.
That last sentence is the one to dwell on, because it quietly settles a question of precedence.
There are now two accounts of the universe's first 2 billion years. One of them is the universe. The other is in Tennessee. Where they disagree, the paper is admirably clear about which of the two gets revised, and it is not the one in Tennessee; a discrepancy is treated as evidence of physics not yet fully understood, which is to say of an error in the instructions.
This is, of course, correct scientific procedure, and read flatly it is also a remarkable arrangement. The universe has been enrolled as the examiner in a paper it did not sit, on a syllabus it was not shown, and its only permitted answer is a mark.
The comparison of resources is no kinder to the original. The universe produced its first 2 billion years once, over 2 billion years, with no documentation and no version control. Lumina produced them again across a cube 1.5 billion light-years on a side, on 15,000 processors, with a complete record of every parameter used.
The Department of Unnecessary Research notes that only one of the two runs can be repeated, and it is not the original.