It typically covers 13.7 billion years.Dr. Jeremy Bailin produces movies. If you watch one, you should concentrate during his signature flash-forward scene.
Bailin, a University of Alabama astrophysicist, creates and analyzes computer simulations of galaxy formation. Using supercomputers, high-powered machines capable of extremely rapid calculations, it can take a few months just to run the computations necessary to produce a 3-minute movie like this one (right).
The use of astronomical simulations has become common over the last 25 years, researchers say. They are the theoretical complement to direct observations.
Through simulations, astronomers like Bailin and his UA departmental colleague, Dr. Dean Townsley, say they gain insight into objects in space that is not possible through observations alone, even with the most powerful telescope.
Telescopes provide views from only their vantage points, and they show observers a snapshot of an object at a specific time. With a click of a mouse, Bailin can tilt a simulated galaxy, viewing it from different angles, giving a three-dimensional perspective.
“I show this to my friends who observe galaxies,” says Bailin, while using his mouse to tilt a galaxy on a monitor in his office, “and their jaws drop open, and they say, ‘I want to do that with my galaxies.’ And they can’t.”
Townsley faces a different set of challenges in his study of one kind of supernovae – a type of exploding star that, for several weeks, outshines the galaxy where it’s located before dimming. The supernovae in which the UA astronomer is most interested receive their energy from white dwarfs, the final dim stages of low- and medium-mass stars.
Researchers want to better understand how supernovae work, including the process involved in the transition from white dwarf to supernova, and the origin of the white dwarf.
“With observation, it’s comparatively easy to observe the supernovae,” Townsley says, “but observing the thing that came before is very hard.”
Scientists don’t know where the supernovae explosion will occur before seeing it, so it’s difficult to know where to train their telescopes to see the white dwarf that precedes it. And the lack of light produced by white dwarfs greatly adds to the challenge.
“We can use what we understand from stellar physics to do a simulation of what we think will happen,” Townsley says.
Using calculations and simulations, theoretical astrophysicists, like Townsley, develop multiple hypotheses to predict how the processes that lead to supernovae may have unfolded. By comparing the explosions of their simulated systems with telescopic views of the actual explosions, the researchers can see which one of the hypothetical origins appears most viable.
“It’s all about how stars actually work … how the burning in stars works, how these processes actually happen,” Townsley says.
“Another really great thing we can do with simulations is that we can watch galaxies, as an example, evolve,” Bailin says.
Galaxies – collections of stars, gas, dust and dark matter, grouped together – change over long time scales.
“A really fast event in the history of a galaxy lasts a hundred million years,” Bailin says. “So, you can’t sit around and watch a galaxy evolve.”
Although you can observe galaxies at different stages of development, you cannot infer from those observations how they changed, says Bailin, a researcher in UA’s College of Arts and Sciences. Think of it, he suggests, as if you were an alien and you were asked to learn about the aging process of humans by looking at an old group photograph of individuals of differing ages, ethnicities and genders.
Without knowing anything about the biology of aging, you would not know which of the older people, if any, the young person might grow to resemble.
“When you look at anything in astronomy, you are actually seeing it as it was a little bit in the past,” Bailin says. “If you’re looking at a really, really, really distant galaxy, you are actually seeing it as it was when it was pretty young because that light took a long time to get to you. So, you’re seeing it as it was a long time ago.”
Looking at distant galaxies provides clues to what a young galaxy can look like. Nearer and nearer galaxies may suggest how increasingly older galaxies appear.
“But,” Bailin says, “you can’t connect the dots, because you don’t know which young galaxies turn into which middle age galaxies which turn into which old galaxies. “It’s not like all galaxies look the same. There is such a huge variety that this is a really serious problem.”
It’s one simulations attempt to address.
Dr. Owain Snaith, a UA post-doctoral researcher who works with Bailin, says simulations and direct observations are intertwined.
“You always have to do simulations with reference to observations,” Snaith says. “Neither of them exists in a vacuum. They complement each other. And, they both have to move together. Improved observations allow us to test the models. Better models allow us to interpret the observations.”
But before astrophysicists can begin drawing preliminary conclusions from a simulation model, they first must develop it using principles of physics. There are three primary components scientists must account for in developing galaxy simulations, Bailin says.
Dark matter, invisible material that is the main mass in the universe is also the most important mass in the simulation. Diffuse gas, which begins as hydrogen and helium gas, is the second component. The gas becomes denser and turns into stars, the third piece.
With a couple of clicks, Bailin launches a movie of one of his simulations – a galaxy known as g1536 – and describes what transpires on-screen over the next few minutes. The movie depicts how the simulated galaxy evolves from soon after the universe’s birth to the present day – a period of some 13.7 billion years, scientists say.
The most intense action scenes in this movie come early in the galaxy’s life.
“The green fuzzy stuff is the gas, and the individual points – the circles – are the star particles that represent the stars in the galaxy,” Bailin says. “One of the things we see when galaxies form is they don’t form as one object. Small galaxies form first, and they come together and merge together to form larger objects. You see that constantly throughout the simulation.”
This particular simulation is of a spiral galaxy: a galaxy having a flat disc in its center with spiral arms coiling and extending outward, giving a pinwheel-like appearance.
“Gravity is the physical process driving all of this,” Bailin says. “The more mass you have, the stronger gravity is, and the shorter distance you have, the stronger gravity is. So, when things are dense, when you have a lot of mass in a small volume, gravity acts faster.”
One point of emphasis for Bailin is a suite of 16 simulations he co-developed with Dr. Greg Stinson, now of the Max Planck Institute for Astrophysics in Heidelberg, Germany.
Called MUGS, or the McMaster Unbiased Galaxy Simulations, the movies don’t attempt to simulate the birth of actual, observed galaxies. Instead, like many simulations, these 16 movies are statistical representations of how some of the galaxies around us would have formed. Bailin plans to grant other researchers easy access to the simulations by soon posting all 16 online.
“If you have 16 simulations, which we do, you can start to figure out what is general to galaxy formation, as a whole, and what is a quirk of history (in an individual galaxy).”
Principles governing gravity, hydrodynamics, ultraviolet background— light from the formation of stars and quasars—and radiative cooling, which transpires as gases emit light – all must be plugged into the formula, Bailin says.
To produce the movies, the scientists break up the universe into individual pieces and individual time steps, Bailin says. At each time step, they calculate all the forces acting on each of the particles in the simulation. They then advance time and show how piece changes according to the forces acting upon it. The simulation is repeated until it reached the present day.
“Even though we have a blob of gas that is a hundred thousand times the mass of the sun, we think we understand how stars should form inside that blob in a given amount of time,” Bailin says.
So, how accurate are the simulations in comparison to reality? Overall, Bailin says he’s confident of their accuracy.
“There are clearly ways the simulation does not mimic reality,” Bailin said. “We know our red galaxies are not red enough. We know our disc galaxies don’t have big enough discs. There’s no question we’re wrong in those places.
“But the overall way the galaxies form, this hierarchical merging, where small pieces come together into bigger and bigger things, I would be very surprised if that turned out to be wrong in a big picture way. I would believe that we have some of the quantitative details wrong, but, big picture, there is too much evidence now for that.”
Bailin says he was first exposed to galaxy simulation in graduate school, and he still enjoys watching the finished product.
“There is something fun about just being able to watch a movie of it,” Bailin says. We analyze them to death, but you can gain a surprising amount of physical intuition just by watching.”
Townsley says one of the intriguing things about supernovae is that humans, and most everything else in existence, are comprised of materials that originated from them.
“Most of the iron in the Earth, for example, comes from thermonuclear supernovae,” Townsley says. “Most of the silicon comes from core-collapse supernovae. Both of them contribute to all elements heavier than hydrogen and helium. At the beginning of the universe, there was, essentially, just hydrogen and helium, and everything else is made in supernovae.”
So, finding tools, like simulations, to better understand such significant objects is motivational, Townsley says.
“One of the huge revolutions in physics and science was the invention of calculus,” Townsley says. “It was then possible to compute things, to make quantitative predictions. It was huge in astronomy because you could now actually compute how the sky changes, the behavior of the planets, the behavior of the moons around the planets. Before calculus, this wasn’t possible.
“Now, not only do we have calculus, but we can do a huge amount of computations because we have very large computing machines. We don’t have to just approximate what is happening and then calculate the end result. You can actually simulate the detailed physics.
“It’s a second revolution, and it’s still going on.”
Dr. Bailin is an assistant professor and Dr. Townsley is an associate professor within UA’s department of physics and astronomy.