Colliding Galaxies Move From the Lab to the Movies

Our most powerful telescopes can see billions of years back in time, but only in our imaginations can we fully envision the birth and death of our universe. To help tell this story, Chris Hayward, a theoretical astrophysicist at the Center for Computational Astrophysics, develops state-of-the-art computer simulations to explore the interplay of the smallest and largest scales of physics and their influence on the structure and evolution of the cosmos. From the birth and sometimes cataclysmic death of stars to the tug of distant galaxies on the Milky Way and its neighbors, Hayward’s simulations encapsulate multiple kinds of physics and physical scales to explain what we see today.

Struck by the beauty and deeper meaning of the visualizations, the writer and director Terrence Malick included one of them in his new film Voyage of Time, a labor of love that traces the 13.8-billion-year history of the universe, including the evolution of Earth and of life, and then goes on to forecast the future, after the Earth and its life are long gone. The Imax version of the film, narrated by Brad Pitt, was released on Oct. 7.

Hayward had a chance to see the film at its Los Angeles premiere and recently spoke with us about the experience. An edited version of the interview follows.

Your visualizations are aesthetically stunning, but they also offer insight into the history and evolution of the universe. What do they tell us? 

One of the challenges in astrophysics is that you have a large range of scales, and multiple kinds of physics, such as gravity, fluid dynamics, magnetic fields and radiation, which are all important and coupled. To start with, there’s the scale of the whole universe, over which very distant galaxies exert gravitational forces on the Milky Way, and the Milky Way tugs on galaxies far, far away. Those gravitational interactions influence how matter clumps together.

Then, on astrophysically small scales, there’s the physics of how gas cools and is compressed in shocks. At high densities, stars form and then die, sometimes in violent explosions called supernovas, which release energy and momentum that subsequently affect the gas out of which the star was born. So there are feedback loops, where the small scales couple up to the larger scales. Much of modern theoretical galaxy formation is done with simulations to reveal the interplay between these different types of physics and different types of scales.

What physics were you investigating with the galaxy merger simulation visualized in Voyage of Time?

One of the key developments in that galaxy merger simulation, which Phil Hopkins now at the California Institute of Technology, other collaborators, and I described in 2013, was the treatment of stellar feedback processes. We were looking specifically at feedback that occurs when stars form, and we were working on understanding how those processes affect the gas around the stars. We implemented a new model that described these interactions, and we wanted to see what happens in a galaxy collision. Such mergers generate a lot of compression of gas, which generally leads to a burst in star formation. If, for example, you visualize the gas in these simulations, you see what looks like lots of bombs going off. It’s very dramatic and offers insight into how star formation affects galaxy evolution.

Because understanding the physics of these feedback processes is crucial to solving the puzzle of galaxy formation, we developed a large-scale collaborative project, Feedback in Realistic Environments, that focuses on modeling these feedback processes, studying their implications, and comparing our predictions with observations to test our understanding.

How do your visualizations compare with astronomers’ observations?

Sometimes when we make visualizations, we try our best to mimic what astronomers would observe with telescopes, such as the Hubble Space Telescope. A large fraction of my research involves calculating how our simulated galaxies would appear in real observations so that we can do apples-to-apples comparisons with real data. However, we also sometimes create visualizations that do not correspond to specific observables because we can still gain insight from the simulations. For example, our telescopes cannot image dark matter, which constitutes the majority of the matter in the universe, but we can model it in our simulations and create visualizations to understand its behavior.

Why do you transform your simulations into visualizations?

High-quality visualizations are a great tool for capturing the audience’s attention. They have the advantage of conveying qualitative aspects of our computer simulations in an easy-to-parse fashion. For example, when we visualize the gas in our simulations, we use color to indicate different temperatures. Then, by watching the visualization, we can understand physical processes, such as how exploding stars can heat gas and expel it from a galaxy.

How did your visualizations end up in Terrence Malick’s Voyage of Time?

The film’s visual-effects coordinator contacted Phil Hopkins after seeing some visualizations on his website. Phil asked if anyone in our group would be interested in working with him and with the Voyage of Time people to adapt the visualizations for the film, and I volunteered because I thought it would be a great opportunity to share my research with the public.

Both of you attended the Los Angeles premiere of Voyage of Time. How did you react to seeing your scientific work on the big screen? 

Originally, the visual-effects coordinator requested that we re-render three different visualizations at higher resolution. We then iterated dozens of times on each visualization, spending tens of thousands of computer processor hours and weeks of our time. In the end, only one visualization made the cut and is in the actual Imax film. It appears in the first few minutes, which cover the time from the Big Bang to the formation of the solar system and Earth. Most of the film is focused on the formation and evolution of life.

When I saw it, I felt a lot of pride. I was especially proud when seeing my name in the credits. I also felt a lot of relief that we actually made the cut — which I didn’t know for certain until I saw the film — because Terrence Malick is known for being extremely exacting.

How scientifically accurate are the visualizations in the film compared with the computer simulations?

The galaxy collision visualization included in the film is very scientifically accurate: The colors encode information about the ages and chemical compositions of the stars in the galaxies and how the light from the stars is affected by cosmic dust. I’m glad that we were able to provide a visualization that is both sufficiently visually appealing for a Hollywood film and scientifically accurate.

What did you hope to achieve by sharing your visualizations with Terrence Malick and his team?

As scientists, we have an ethical obligation to share our knowledge with the general public because, first, the whole reason that we do science is to advance humanity’s knowledge, and second, more practically speaking, taxpayers fund much of our work, so we owe it to them to communicate our results to them in a digestible fashion. The opportunity to share our results via a high-profile Hollywood feature film, created by a renowned director, is a rare chance to share our research with a broad swath of the public, and we jumped at the chance to do so.

It is my hope that nonscientists will be motivated by the film to learn more about modern science, which I think is one of the most valuable endeavors undertaken by humanity.

Visualization of galaxy merger (Courtesy of Chris Hayward/CCA and Phil Hopkins/Caltech)

In this preliminary visualization of a galaxy collision, later adapted for the film, two spiral galaxies slam into each other again and again until they finally merge.