Scientists precisely simulate turbulence in the Galaxy — it doesn’t behave like they thought

From the ocean’s rolling swells to the bumpy ride of a jetliner, turbulence is everywhere. It breaks large waves into smaller ones, cascading energy across scales. It is ubitquitous throughout our Galaxy and the broader Universe, shaping the behavior of plasma, stars, and magnetic fields. Yet despite its ubiquity, turbulence remains one of the greatest unsolved problems in physics.

Now, by developing the world’s largest-ever simulations of magnetized turbulence, an international team of scientists has measured — with unprecedented precision — how turbulent energy moves across a vast range of scales. The result: it doesn’t match with long-standing theories.

James Beattie, a postdoctoral researcher at Princeton University's Department of Astrophysical Sciences and a fellow at the Canadian Institute for Theoretical Astrophysics at University of Toronto, led the study along with Amitava Bhattacharjee of Princeton, and colleagues at the Australian National University, Heidelberg University and the Leibniz Supercomputing Center.

By simulating galactic-type turbulence in exquisite detail, the researchers found significant departures from the models that have guided astrophysical theory for decades. The team explicitly observed that magnetic fields alter the way energy cascades through the space between stars in our Galaxy — known as the interstellar medium — suppressing small-scale motions and enhancing certain wave-like disturbances known as Alfvén waves. The findings could reshape how scientists understand the turbulent structure of the Galaxy, the transport of high-energy particles, and even the turbulent birth of stars.

In practical terms, understanding and properly modeling turbulence and the production of highly energetic particles can shed light on how to safely navigate space, at a time when commercial space flight is growing and attracting the interest of civilians and celebrities alike.

‘‘The research has implications for predicting and monitoring space weather to better understand the plasma environment around satellites and future space missions, and also the acceleration of highly energetic particles, which damage everything, and could endanger human beings in space,” said Bhattacharjee, a co-author on the new paper and Professor of Astrophysical Sciences at Princeton.

“A lot of these fundamental plasma turbulence questions are objects of missions now launched by NASA and have implications for understanding the origin of cosmic magnetic fields. Simulations like these would give us insights into how to interpret satellite and ground-based measurements,” said Bhattacharjee.

Simulating Turbulence like Never Before

There is still no complete mathematical framework for predicting how energy moves from large to small scales: across oceans, in the atmosphere, or through the plasma and dust between stars. In space, the problem is even more complex than on Earth due to magnetization, requiring vast computational resources to model. The team’s work relied on the equivalent of 140,000 computers running in parallel. 

“To put these massive simulations into perspective: if we had started one on a single laptop when humans first domesticated animals, it would just be finishing now,” said Beattie. “Luckily, utilizing the amazing resources from the Leibniz Supercomputing Centre, we can distribute the workload across thousands of computers to accelerate the calculations.”

“We are a step closer to uncovering the true nature of astrophysical and space turbulence, from chaotic plasma near Earth to the vast motions within our Galaxy and beyond,” said Beattie, “The dream is to discover universal features in turbulence across the Universe, and we’ll continue pushing the limits of the next-generation of simulations to test that idea.”

The new work will be published in the journal Nature Astronomy on May 13, 2025. In addition to Beattie and Bhattacharjee, co-authors include Christoph Federrath of the Australian National University, Ralf S. Klessen of Heidelberg University, and Salvatore Cielo of the Leibniz Supercomputing Center of the Bavarian Academy of Sciences and Humanities.