SIMULATING
THE UNIVERSE
Dark matter distribution
70°
50
A S T R O N O M Y • MAY 2018
60°
50°
40°
30°
20°
10°
0°
350°
340°
–30°
In 1941, five years before digital
computers were invented, Erik
Holmberg of Lund Observatory in
Sweden performed the first simu-
lation of colliding galaxies. What
he lacked in computational power
he made up for with ingenuity.
Because an object’s gravita-
tional pull and apparent bright-
ness both vary inversely with the
square of distance, Holmberg
realized he could use light as a
proxy for gravity. Representing
two galaxies with 37 lightbulbs
each, he measured the brightness
at different locations to deter-
mine the strength and direction
of the gravitational force and
adjusted their motions accord-
ingly. Although crude, Holmberg’s
analog computation provided
insights into the frequency of gal-
axy mergers and suggested that
close encounters could generate
spiral arms.
Astronomers quickly
embraced digital computers as a
tool to simulate events and time-
scales beyond anything accessi-
ble to mere mortals. In the 1970s,
brothers Alar and Jüri Toomre
pioneered numerical simulations
of galaxy mergers by represent-
ing each galaxy as a swarm of
interacting particles whose tra-
jectories were calculated by the
computer. Although the avail-
able computing power at the
time limited the number of par-
ticles that could be followed,
these simulations revealed that
galaxy mergers are common,
and that mergers of spiral gal-
axies can produce things that
look like elliptical galaxies.
As computational power
grew, so did astronomers’ ambi-
tion. Today, thanks to state-of-
the-art computers, it’s possible
to simulate the evolution of
enormous volumes of space in
unprecedented detail using tril-
lions of particles to represent
both luminous and dark matter.
Beginning from some assumed
set of initial conditions, astrono-
mers can run the simulation for-
ward in time to see what the
predicted distribution and
properties of galaxies would be
today and then compare the
results with observations. Such
simulations have become an
invaluable tool for understand-
ing how the universe reached
its present state. — M.W.
– 40°
1 billion light-years
–50°
An international team of scientists used the 570-megapixel Dark Energy Camera at Cerro Tololo
Inter-American Observatory in Chile to record light from 26 million galaxies. Tiny distortions in
their apparent shapes caused by gravitational lensing provide a map of the dark matter in different
regions of space. In this image, red regions have more dark matter than average, while blue regions
have less. Hints of large-scale filamentary features can be seen. The European Space Agency’s
Euclid mission and NASA’s Wide Field Infrared Survey Telescope, both scheduled for launch in 2020, will
measure cosmic shear from space, providing even sharper views of the cosmic dark matter distribution.
CHIHWAY CHANG OF THE KAVLI INSTITUTE FOR COSMOLOGICAL PHYSICS AT THE UNIVERSITY OF CHICAGO AND THE DES COLLABORATION
They noticed that clusters in the Perseus-Pisces
Supercluster are elongated in the same direc-
tion as the filament that bridges them, leading
them to suggest in a 1980 paper published in
Monthly Notices of the Royal Astronomical
Society that “the orientation of clusters in
superclusters is a conspicuous morphological
property of superclusters.”
The alignment of galaxies and clusters over
tens or hundreds of millions of light-years
means the innermost regions of some galaxies
are aligned with their surroundings on scales
larger than 1,000 times the size of a single gal-
axy itself. It suggests that the birth and evolu-
tion of these objects have been strongly
influenced by the cosmic web. But how?
Going with the flow
A computer simulation shows the gossamer-like
structure of the cosmic web. Within this network,
matter flows along filaments, piling up where they
intersect. The simulation shown here, one of the
largest ever done, followed the motion of trillions
of particles as gravity amplified tiny variations
in their initial distribution. Black points show the
locations of dark matter; yellow denotes dense
regions where galaxies and clusters form; and
white indicates voids. The region displayed here is
2.5 billion light-years across, only a portion of the
entire simulated volume. JOACHIM STADEL, UZH
Galaxies, like people, are products of their
environment. Elliptical galaxies, for example,
usually huddle together in groups and clusters,
while spiral galaxies prefer more elbowroom.
Environment clearly plays a role in galaxy
orientations, too. There are two leading theo-
ries for how this happens. One suggests that
galaxies are born aligned with their surround-
ings, while the other assumes that alignment is
something they acquire later.
Galaxies might gain their orientation in
several ways. Big galaxies grow by cannibaliz-
ing smaller ones, a process astronomers euphe-
mistically call merging. But mergers aren’t
haphazard. Computer simulations show that
they occur most frequently along well-defined