Preprint: DESI spectra identify 10,000+ candidate cold gas inflows into galaxies

A new arXiv preprint reports what could be the largest direct observational sample yet of cold gas flowing into galaxies, using data from the Dark Energy Spectroscopic Instrument, or DESI. In a search of 15.6 million galaxies at redshift less than 0.6, the authors say they identified more than 10,000 likely inflows — a scale far beyond past studies based on tens or hundreds of objects.

That matters because astronomers have generally had a much easier time spotting gas being driven out of galaxies than gas falling in. Yet that incoming gas is a basic part of galaxy growth, replenishing the material needed to keep forming stars. As the authors write in the abstract, “Direct observational constraints on how galaxies acquire their gas remain remarkably limited.”

The paper, “Peering down the barrel with DESI DR2: 10 000+ inflows at z < 0.6 reveal how galaxies accrete cold gas,” was posted to arXiv on May 4 as arXiv:2605.02999v1 and marked as submitted to Astronomy & Astrophysics. Lead author S. Weng, or Simon Weng, and 41 co-authors used spectra from DESI Data Release 2 to look for sodium absorption, known as Na I D, in galaxies’ own light. This “down-the-barrel” approach uses absorption features seen against a galaxy itself to infer whether gas along our line of sight is moving toward the galaxy or away from it. The team says it used Bayesian evidence ratios to test whether the absorption required extra interstellar gas components beyond a galaxy’s systemic component.

According to the abstract, the researchers built a catalog of 50,088 galaxies with moderate evidence and 27,420 galaxies with strong evidence for down-the-barrel absorption. They say the inferred absorption components split into three broad groups: about 50% at flow velocities below minus 50 kilometers per second, about 30% within 50 kilometers per second of a galaxy’s systemic velocity, and about 20% above 50 kilometers per second. That last group underpins the paper’s headline claim of “10,000+ inflows,” though the classification and interpretation are, at this stage, the authors’ own and have not yet been vetted through peer review.

The abstract highlights two main scientific results. First, the authors report “strong evidence for a large population of low-velocity, infalling absorbers” moving at about 20 kilometers per second in edge-on galaxies, which they say is consistent with radial inflows predicted in simulations. Second, in early-type galaxies, they say inflow velocity tracks stellar velocity dispersion more strongly than stellar mass, which they interpret as a sign that some inflows may be linked to accreting satellites. Taken together, the preprint argues that galaxies at redshift less than 0.6 may be gaining gas through multiple pathways.

If that result holds up, the significance is less about a single dramatic object than about finally having a sample large enough to sharpen observational constraints on the baryon cycle, the flow of ordinary matter into and out of galaxies. But for now, the work should be treated as preliminary. The study is an arXiv preprint with 42 authors, not a peer-reviewed paper, and its claims will need to withstand outside scrutiny before they can be considered established.