Simulating the Milky Way

This image of the Eris simulation shows the stars in the galaxy as observers would see it. Blue colors are regions of recent star formation, while redder regions are associated with older stars. The spiral arms are typically star-forming, and the central bulge is basically “red and dead.” Credit: J. Guedes and P. Madau

After nine months of number-crunching on a powerful supercomputer, a
beautiful spiral galaxy matching our own Milky Way emerged from a
computer simulation of the physics involved in galaxy formation and
evolution. The simulation by researchers at the University of
California, Santa Cruz, and the Institute for Theoretical Physics in
Zurich solves a longstanding problem that had led some to question the
prevailing cosmological model of the universe.

"Previous efforts to form a massive disk galaxy like the Milky Way had
failed, because the simulated galaxies ended up with huge central
bulges compared to the size of the disk," said Javiera Guedes, who
recently earned her Ph.D. in astronomy and astrophysics at UC Santa
Cruz and is first author of a paper on the new simulation, called
"Eris." The paper has been accepted for publication in the
Astrophysical Journal.

The Eris galaxy is a massive spiral galaxy with a central "bar" of
bright stars and other structural properties consistent with galaxies
like the Milky Way. Its brightness profile, bulge-to-disk ratio,
stellar content, and other key features are all within the range of
observations of the Milky Way and other galaxies of the same type. "We
dissected the galaxy in many different ways to confirm that it fits
with observations," Guedes said.

According to coauthor Piero Madau, professor of astronomy and
astrophysics at UC Santa Cruz, the project required a large investment
of supercomputer time, including 1.4 million processor-hours on NASA´s
state-of-the-art Pleiades supercomputer, plus additional supporting
simulations on supercomputers at UCSC and the Swiss National
Supercomputing Center. "We took some risk spending a huge amount of
supercomputer time to simulate a single galaxy with extra-high
resolution," Madau said.

The results support the prevailing "cold dark matter" theory, in which
the evolution of structure in the universe is driven by the
gravitational interactions of dark matter ("dark" because it can´t be
seen, and "cold" because the particles are moving slowly). Gravity
acted initially on slight density fluctuations present shortly after
the Big Bang, pulling together the first clumps of dark matter, which
grew into larger and larger clumps through the hierarchical merging of
smaller progenitors. The ordinary matter that forms stars and planets
(less than 20 percent of the matter in the universe) has fallen into
the "gravitational wells" created by large clumps of dark matter,
giving rise to galaxies in the centers of dark matter halos.

For the past 20 years, however, efforts to reproduce this process in
computer simulations have failed to generate massive disk galaxies
that look anything like the Milky Way, with its spiral arms in a large
flat disk around a small central bulge made up of old stars. A
realistic simulation of star formation was the key to Eris´s success,
Madau said.

This comparison shows the Eris simulation (top) and the Milky Way (bottom). Credit: S. Callegari, J. Guedes, and the 2MASS collaboration

"Star formation in real galaxies occurs in a clustered fashion, and to
reproduce that out of a cosmological simulation is hard," Madau said.
"This is the first simulation that is able to resolve the high-density
clouds of gas where star formation occurs, and the result is a Milky
Way type of galaxy with a small bulge and a big disk. It shows that
the cold dark matter scenario, where dark matter provides the
scaffolding for galaxy formation, is able to generate realistic
disk-dominated galaxies."

To perform the Eris simulation, the researchers began with a
low-resolution simulation of dark matter evolving to form the haloes
that host present-day galaxies. Then they chose a halo with an
appropriate mass and merger history to host a galaxy like the Milky
Way and "rewound the tape" back to the initial conditions. Zooming in
on the small region that evolved into the chosen halo, they added gas
particles and greatly increased the resolution of the simulation. High
resolution means tracking the interactions of a huge number of

"The simulation follows the interactions of more than 60 million
particles of dark matter and gas. A lot of physics goes into the code
— gravity and hydrodynamics, star formation and supernova explosions
— and this is the highest resolution cosmological simulation ever
done this way," said Guedes, who is currently a postdoctoral
researcher at the Swiss Federal Institute of Technology in Zurich (ETH

The high resolution allowed for a more precise recipe for star
formation. In a low-resolution simulation, with gas densities averaged
out over relatively large areas, the threshold density for star
formation has to be set so low that stars tend to form in diffuse gas
throughout the galaxy. In the Eris simulation, the star-formation
threshold allowed stars to form only in high-density regions,
resulting in a more realistic distribution of stars.

An important consequence is that when stars explode as supernovae
within these localized, high-density regions, the energy injected into
the interstellar medium blows a lot of gas out of the galaxy.
"Supernovae produce outflows of gas from the inner part of the galaxy
where it would otherwise form more stars and make a large bulge,"
Madau said. "Clustered star formation and energy injection from
supernovae are making the difference in this simulation."