Science1 distinct publisher3 min readPublished
A NASA-funded team simulated the solar system's passage through cold galactic clouds and put dates on when Earth sat outside the Sun's shield. Dust in seabed cores gives that reconstruction a check; the temperature effect has none yet.
The Scientist · Science desk
Compiled by The ScientistSomething wrong?How this is made
The chain of inference has three links and they are not equally strong. The heliosphere is inflated by charged particles streaming outward from the Sun in every direction [12], so its radius is a pressure balance, and a cold dense cloud raising the external pressure is a straightforward fluid-dynamics problem. That is the sturdy link. The second link is atmospheric: in the simulations, exposure to dense galactic hydrogen raised atmospheric water vapour and altered the upper atmosphere, with effects that reached down toward the surface [8]. The third link is the climate attribution, and here the release stops. It reports no temperature change and no radiative forcing for these encounters [16]. The "several degrees" figure that appears nearby belongs to the ice ages themselves [11], not to anything the model assigns to a cloud.
The dust markers are the most testable part of the argument, so it is worth being exact about what they test [7]. They test whether the solar system passed through dusty material at roughly those times. They do not test whether Earth's temperature responded. Reconstructions of these deep-time shifts have been built on Earth's orbit, greenhouse gas concentrations and ice coverage [10], and an external forcing term earns entry by accounting for variance those leave unexplained. That case is not the one being made here.
Take the three windows at their midpoints and they sit 2.5, 6.5 and 13.5 million years ago, gaps of four and seven million years [14]. Uneven spacing is honest but awkward: a periodic galactic signal would be far easier to test against layered sediment records than three one-off events. All three fall inside the last 14 million years, roughly 0.3 percent of the Sun's 4.6-billion-year history [15][13], so this is a claim about the recent galactic neighbourhood rather than the long record. And the release does not say whether "2 to 3 million years ago" describes how long an encounter lasted or how precisely its date is known [18]. The difference between a million-year exposure and a brief one is the whole question of whether any climate response had time to develop.
The companion study runs on an entirely different clock. It addresses how a dimmer young Sun kept early Earth warm, with powerful solar eruptions generating strong greenhouse gases in the early atmosphere [9]. The release names neither its author nor its journal [17], which makes it the thinner half of the pairing. What the two share is a filing category rather than a mechanism, and the category is the interesting part: solar and galactic environment as an input to paleoclimate rather than a backdrop to it. The SHIELD work itself was published on Aug. 21 in Annual Review of Astronomy and Astrophysics [3]. My guess is the trajectory reconstruction holds and the climate attribution stays qualitative for years. The number that would change that is a forcing in watts per square metre at a stated cloud density and heliosphere radius. That is the one I would wait for.
Ranked by verification strength, evidence, and original report placement.
The release identifies the second study only as a separate study led by a NASA scientist, without naming its lead author or publication venue.
A NASA Goddard Space Flight Center release dated September 2, 2026 reports two recent NASA-funded studies suggesting events from the Sun's distant past influenced Earth's climate.
SHIELD stands for Solar Wind with Hydrogen Ion charge Exchange and Large-Scale Dynamics, and is one of NASA's DRIVE (Diversify, Realize, Integrate, Venture, Educate) Science Centers.
The SHIELD study was published on Aug. 21 in Annual Review of Astronomy and Astrophysics and used computer simulations to reconstruct the heliosphere's past trajectory through the galaxy.
Merav Opher, SHIELD's principal investigator at Boston University, and colleagues simulated encounters between the solar system and extremely cold regions filled with gas and dust, and their work indicates the Sun passed through such environments at least three times during the past several million years.
According to the simulations, interstellar cold clouds may have pressed against the heliosphere hard enough to compress it dramatically, at times shrinking it smaller than Earth's orbit and temporarily leaving the planet outside the Sun's protective bubble.
Distinct publishers with included, body-backed reporting in this cluster.
1 article · September 1, 2026
Follow any of these and your For You feed starts watching them — no settings page required.
science
A pole-on magnetar hands vacuum birefringence its first astrophysical candidate1 distinct publisher
science
Ten years of hip scans place coffee's bone-density signal above five cups a day1 distinct publisher
science
Sucralose and stevia reshaped mouse gut chemistry into generations that never drank them1 distinct publisher
science
Myelin repair cells make their own stress hormone, and it seems to set the clock1 distinct publisher
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Peer-reviewed paper, single-release telling
Everything reaches the reader through one document: NASA Goddard's own announcement, reprinted by ScienceDaily. The release is precise in one direction — journal, publication day, principal investigator, institution — and vague in the other: the interstellar dust in seabed cores, Antarctic snow and Moon samples is the reconstruction's only external check, and no one says whose measurements those are. The dates rest on simulation; nothing observed constrains how big the heliosphere was two million years ago.
No uptake visible either way
Nothing in this reporting shows anyone outside SHIELD working with the reconstruction. The August 21 paper is the work surfacing, not other groups taking it up, and the 'digital twin' of the heliosphere Goddard describes is stated as a goal of the center rather than something in use. Citations, a replication attempt or a competing compression model would settle this; none is offered, so we decline to score it.
Shield lost in the headline, 'may' in the text
The body hedges properly — 'may have', 'according to the simulations', 'raise the possibility' — and then the headline states in the past tense that Earth lost the Sun's protective shield. The gap opens widest at the climate joint: water vapor rises inside a model, ice ages appear in the summary line, and no quantity links the two. The three dates are the sturdiest thing on offer, and they at least arrive alongside dust records of matching age.
The funder writes the announcement
NASA funds SHIELD, NASA Goddard wrote this release, and the release argues that the center NASA funds is illuminating ancient ice ages and, eventually, how to spot habitable systems elsewhere — a passage that reads like a case for continued support. ScienceDaily supplies a headline and no scrutiny. None of that makes the simulations wrong; it does mean nobody in the chain had reason to lean on the weak joint.
One telling, one team
A single publisher carrying a single institutional release, with the only independent contribution being arithmetic on the dates it supplies. That supports reporting what Opher's team claims and when they published it; it does not support judging whether a heliosphere crushed inside Earth's orbit survives contact with other modelers, or even whether '2 to 3 million years ago' means a duration or a margin of error.