Stars form in dense pockets of interstellar gas, but those pockets are not static nurseries — they are transient, churning structures that assemble, form stars, and disperse on timescales far shorter than the galaxies that host them. A new theoretical study led by University of Maryland astrophysicist Zuzanna Kocjan, working with Vadim A. Semenov, uses hydrodynamic simulations to pin down exactly how short-lived that process is, and the numbers are striking: individual star-forming gas clumps survive for only about 0.4 to 1 million years before feedback from young stars disperses them.
The paper, titled "The Rhythm of the ISM: Tracing the Timescales of Gas Evolution and Star Formation Across Galactic Environments," is posted on arXiv (2602.02657), where its comments field describes it as submitted to The Astrophysical Journal. It was re-featured by the American Astronomical Society's AAS Nova research highlight service on July 28, 2026, under the title "The 'Rhythm' of the Interstellar Medium," which — along with author Vadim Semenov's own publications page — describes the paper as accepted to The Astrophysical Journal.
Three Clocks, One Galaxy
The core contribution of the work is a framework, not just a number. Kocjan and Semenov built simulations spanning three very different types of galaxies — a dwarf galaxy, a Milky Way-mass galaxy, and a gas-rich starburst analog — and used them to define three distinct timescales that together describe the life cycle of star-forming gas:
- Supply timescale: how long it takes for gas to be gathered into a star-forming region in the first place. According to the paper, star-forming regions form on a timescale close to the galaxy's vertical turbulent crossing time — the time it takes turbulent motions to shuttle gas across the thickness of the galactic disk. The paper reports this assembly process takes 3 to 30 million years, decreasing at higher gas surface density as turbulent velocities rise and the disk thins.
- Removal timescale: how long stellar feedback — the radiation, stellar winds, and supernova explosions from newborn stars — takes to blow the gas clump apart. This is where the headline number comes from: individual star-forming gas structures live for just 0.4 to 1 million years before feedback disperses them, only weakly increasing with gas density.
- Depletion timescale: how long it takes for the gas reservoir to be consumed by star formation itself, rather than dispersed. This is the slowest of the three clocks, ranging from 200 to 2,000 million years depending on gas density — with denser regions depleting faster because they form stars more efficiently.
Put together, the framework connects two scales that are normally studied separately: the kiloparsec-scale dynamics of an entire galactic disk, and the much smaller, much faster evolution of an individual star-forming region embedded within it. The supply timescale is set by galaxy-scale turbulence; the removal timescale is set by the physics of stellar feedback acting locally; and the depletion timescale bridges the two, reflecting how efficiently gas converts to stars once it's assembled.
Why the Speed Matters
The sub-million-year lifetime for individual gas clumps is short even by astronomical standards — many times shorter than the roughly 3-to-30-million-year supply timescale that assembles the gas in the first place, and hundreds to thousands of times shorter than the depletion timescale that would be needed to convert most of it into stars. That brevity means any single star-forming region observed today is essentially a snapshot of one link in a fast-moving chain: gas gathers, forms stars, and gets dispersed by those same stars before much of it is converted into stellar mass. The AAS Nova write-up notes that denser gas regions are both more efficient at forming stars and deplete faster, tying the removal and depletion timescales together as functions of local gas density rather than treating them as fixed properties of a galaxy as a whole.
This matters for how astronomers interpret observations of star-forming regions across different galaxy types. A dwarf galaxy, a Milky Way-mass galaxy, and a starburst galaxy don't just have different total amounts of gas — according to the simulations, they operate on different versions of all three clocks, which affects how quickly their interstellar medium cycles between diffuse gas, dense star-forming clumps, and newly formed stars.
Who Did the Work
Kocjan, the lead author, is affiliated with the University of Maryland. Semenov's own publications page lists the paper under work "led by co-advised students" and marks it as accepted to The Astrophysical Journal. The arXiv preprint runs 22 pages with 13 figures and is released under a CC BY 4.0 license, meaning the full paper and its figures are freely reusable with attribution.
Why It Matters
Star formation is the process that turns raw interstellar gas into the stars, planets, and eventually the elements that make up everything from galaxies to living organisms — but it has long been difficult to model consistently because it happens on wildly different scales at once, from galaxy-wide gas dynamics down to the collapse of individual molecular clouds. By defining three explicit, physically distinct timescales — supply, removal, and depletion — and showing how each depends on galactic environment and local gas density, this study gives astronomers a common vocabulary and a testable framework for comparing star formation across galaxy types. The short, sub-million-year lifetime of individual star-forming clumps in particular is a strong constraint: it means feedback from young stars acts far faster than gas depletion, which has implications for how efficiently galaxies convert their gas reservoirs into stars over cosmic time, and for how observational surveys — which typically capture only a snapshot of any given region — should be interpreted against the full life cycle the simulations describe.
