Astronomers have long suspected that supernova explosions stir the gas inside galaxies into the turbulent, churning state that pervades the interstellar medium, but proving it across an entire galaxy has remained stubbornly out of reach. Now, a team using China’s Five-Hundred-Meter Aperture Spherical Radio Telescope together with the Jansky Very Large Array has delivered the most compelling evidence yet. By surveying the neutral atomic hydrogen in the Andromeda galaxy, the nearest giant spiral to the Milky Way, the researchers catalogued 118 expanding shells known as superbubbles and showed that the energy these structures inject into the gas matches, in both magnitude and spatial pattern, the energy that turbulence dissipates throughout the galactic disk. The finding, published in Nature Astronomy, closes a decades-old gap between theory and observation in one of astrophysics’ most fundamental energy-budget problems.
Turbulence is not a cosmetic feature of galaxies; it is a controlling one. The random, supersonic motions of interstellar gas set the thickness of galactic disks, regulate how molecular clouds collapse into new stars, and determine how efficiently matter converts into luminous stellar populations. Without a continuous supply of energy, turbulence in a galaxy like Andromeda would decay on timescales of only a few tens of millions of years, far shorter than the age of the disk itself. Something must constantly re-energize the gas, and supernovae have topped the list of candidate drivers since the earliest theoretical treatments of the multiphase interstellar medium. Alternative mechanisms, including gravitational instabilities in the rotating disk and the magnetorotational instability, have also been proposed, and disentangling their contributions observationally has proven exceptionally difficult.
The key to the new result lies in superbubbles, the gigantic cavities that clusters of massive stars carve out of the neutral hydrogen gas around them. When a group of hot, short-lived stars forms together, the combined winds and subsequent core-collapse supernovae of its members blow a common expanding shell into the surrounding medium. These shells, sometimes hundreds of light-years across, act as fossil records of clustered stellar feedback: their sizes, expansion velocities, and ambient gas densities encode how much kinetic energy the parent star cluster deposited into the disk. Because supernovae in galaxies tend to occur in clusters rather than in isolation, superbubbles are the natural conduits through which stellar feedback feeds galactic-scale turbulence, making a complete census of them the decisive observational test.
Until now, no such dynamically complete, galaxy-wide census existed. Earlier surveys of hydrogen holes and shells in Andromeda and other nearby galaxies, dating back to work in the 1980s, identified cavities in the gas but lacked the sensitivity and velocity resolution to measure expansion reliably across a whole disk, leaving the energy budget unsettled. The new study overcame this limitation by combining the extraordinary sensitivity of FAST, the world’s largest single-dish radio telescope, with the fine angular resolution of the Jansky Very Large Array. The combination was essential: single-dish data recover the diffuse, large-scale hydrogen emission that interferometers miss, while interferometric data resolve the fine structure of individual shells. An image-fusion technique merged the two data sets into a single, dynamically complete hydrogen data cube of Andromeda, capturing structures on all relevant scales simultaneously.
From this combined data set, the team identified 118 superbubbles distributed across the entire disk of M31, each appearing as a coherent shell in both space and velocity. The dynamical ages of the shells extend up to roughly 40 million years, a range that matches theoretical expectations for the duration of supernova activity within a star cluster, since the most massive members explode within a few million years while lower-mass stars detonate tens of millions of years later. This consistency between the observed age distribution and stellar-evolution models strengthens the interpretation that the shells are indeed powered by clustered supernovae rather than by other processes such as infalling clouds or galactic-scale instabilities.
The crucial quantitative step was an energy accounting performed two independent ways. First, from the measured sizes, expansion velocities, and surrounding gas densities of the superbubbles, the researchers calculated the rate at which supernovae inject kinetic energy into the neutral medium. The inferred injection rates span 10^49 to 10^51.5 erg per cubic kiloparsec per million years. Second, from the same hydrogen data, they measured the turbulent velocity dispersion of the gas and derived how quickly turbulent energy dissipates at each location in the disk, using established scaling relations for supersonic, magnetized turbulence. If supernovae truly power the turbulence, these two independently determined rates should agree, both in total magnitude and in how they vary from place to place across the galaxy.
They do agree, and remarkably well. The kinetic-energy-injection rates inferred from the superbubble population closely match the turbulence dissipation rates derived from the gas kinematics, not only in overall magnitude but also in their spatial distribution across the Andromeda disk. Regions where the shells inject more energy are precisely the regions where the gas exhibits the strongest turbulent motions. This point-by-point correspondence is far more constraining than a global average, because it rules out a coincidence in which supernovae supply the right amount of energy somewhere in the galaxy while a different mechanism actually drives the local turbulence. The result demonstrates that clustered supernova feedback alone is sufficient to sustain galactic-scale turbulence in a giant spiral galaxy.
The implications extend well beyond Andromeda. Because M31 is the nearest giant spiral and a close analogue of the Milky Way, the result provides the strongest direct evidence to date that our own galaxy’s turbulent interstellar medium is likewise maintained by the death throes of massive stars. Turbulence, in turn, feeds back into star formation: it both prevents gas from collapsing too quickly into stars and concentrates density enhancements that seed new star-forming clouds. A galaxy’s structure and evolutionary trajectory therefore depend on this feedback loop, and models of galaxy formation and evolution can now anchor their prescriptions for supernova-driven turbulence to a directly measured, observationally verified energy budget rather than to theoretical assumption alone.
The study also showcases the power of combining complementary radio facilities. FAST’s collecting area delivers sensitivity to faint, extended hydrogen emission at a level no other instrument can reach, while the Very Large Array contributes the sub-arcminute resolution needed to resolve individual shells hundreds of parsecs away in a neighboring galaxy. The data, the 118-object superbubble catalogue, and the analysis code have been released publicly, allowing other researchers to scrutinize the classification, refine the energy estimates, and extend the method to additional galaxies. As similar combined surveys target more spirals, astronomers will be able to test whether supernova-driven turbulence dominates universally or whether gravitational and magnetic mechanisms take over in particular environments, such as low-star-forming outer disks or violently interacting systems.
For decades, the image of galaxies as serene, slowly rotating pinwheels has coexisted with the reality that their gas is in constant, violent motion, churned by forces whose origin remained unproven. With 118 superbubbles now mapped across Andromeda and their energy output shown to balance the turbulent dissipation of the entire disk, that origin is no longer a hypothesis but a measurement. The explosions of massive stars, it turns out, are not merely the spectacular endings of stellar lives; they are the beating heart that keeps entire galaxies stirred, structured, and capable of forming the next generation of stars.
Subject of Research: Observational evidence that clustered supernova feedback, traced by hydrogen superbubbles, sustains galactic-scale turbulence in the Andromeda galaxy.
Article Title: Supernova origin of galactic turbulence revealed by superbubbles
Article References: Meng, F., Tsai, C.-W., Wu, J., Jiao, S., Mac Low, M.-M., Zhang, Z.-Y., Saintonge, A., Li, H., Li, Z., Wang, J., Wang, L., Xu, H., Yang, Y., Zhang, K., Li, R., & Li, D. (2026). Supernova origin of galactic turbulence revealed by superbubbles. Nature Astronomy. https://doi.org/10.1038/s41550-026-02981-9
Image Credits: AI Generated
DOI: 10.1038/s41550-026-02981-9
Keywords: supernovae, superbubbles, galactic turbulence, Andromeda galaxy, M31, neutral hydrogen, FAST telescope, Jansky Very Large Array, interstellar medium, star formation, stellar feedback, Nature Astronomy
Cite Scienmag News
Grant Pearson. (September 20, 2026). Superbubbles Reveal Supernovae as the Engine of Galactic Turbulence. Scienmag. https://scienmag.com/superbubbles-reveal-supernovae-as-the-engine-of-galactic-turbulence/
Grant Pearson. "Superbubbles Reveal Supernovae as the Engine of Galactic Turbulence." Scienmag, 20 September 2026, https://scienmag.com/superbubbles-reveal-supernovae-as-the-engine-of-galactic-turbulence/. Accessed 20 September 2026.
Grant Pearson. "Superbubbles Reveal Supernovae as the Engine of Galactic Turbulence." Scienmag. September 20, 2026. https://scienmag.com/superbubbles-reveal-supernovae-as-the-engine-of-galactic-turbulence/

