Paper check · Uranus’s moon Ariel
A study of the fractured surface of Uranus’s moon Ariel concludes it once needed a subsurface ocean, possibly more than 170 km deep [1 ¶1]. The ocean is a reasonable inference from a model, not a detection. The 170 km figure is the model’s thin-ice solution, and it needs an orbit far more stretched than Ariel’s today.
Our analysis marks our own reasoning and calculations, and the verdicts are our assessment. We couldn’t read the paper itself [2]; what it found comes here from the Planetary Science Institute’s release, the team’s 2024 conference abstract and the coverage.
~40%
of Ariel’s surface was imaged, all of it in the south, in the one close look: Voyager 2 in 1986 [3 ¶7][1 ¶9][4].
≥ 0.04
The orbital eccentricity the model needs, 30–40 times today’s. A 2026 study finds that one candidate cause, a 2:1 resonance with the moon Umbriel, reaches only about 0.02 [1 ¶6][5].
~85%
Our estimate of how much of the thickness of Ariel’s water layer a 170 km ocean would take up if its rock is near 3 g/cm³, leaving about 27 km of ice on top.
The claims about Ariel’s ocean
The release’s headline [1]
In the team’s stress models, the fractures form only with an ocean beneath the ice: thin ice over a very big ocean, or a smaller ocean with a more stretched orbit [1 ¶7]. It is an inference from a model, from images of part of the surface and from an orbit about 0.04 eccentric or more, not a detection. See section 2.
The release, and headlines such as ScienceDaily’s [1 ¶1][6]
This is the model’s thin-ice solution, which needs the least stretched orbit, at least 0.04. Unhedged headlines, such as “Ariel Had A 170km Deep Sub-Surface Water Ocean,” overstate it [7]. See section 1.
The release [1 ¶6]
This is what the model needs to crack the ice, not an observed value. A 2026 study finds that one candidate cause, a 2:1 resonance with Umbriel, raises Ariel’s eccentricity only to about 0.02 [5]. See section 3.
The release’s opening line, as what “growing evidence suggests” [1 ¶1]
The study is about the past. Models suggest any ocean left today would likely be a residual layer under 30 km thick, and only measurements by a spacecraft at Uranus, such as magnetometry, could find it [8 Abstract][9 Abstract]. See section 5.
1 · What 170 km would mean
The release says new research in Icarus found that Ariel’s ocean “may have once been over 100 miles (170 kilometers) deep”, against the Pacific’s average of 4 kilometres [1 ¶1][2][10]. Drawn to scale, that is most of the moon’s outer shell of water and ice.
What is left above a 170 km ocean
The headline case [1 ¶1]
Our assumption, from Ariel’s density
The Pacific’s 4 km average depth, a sliver, beside 170 km, at the same scale
2 · Reading an ocean from cracks
Ariel has old craters beside young, smooth terrain that may come from cryovolcanism, and fractures, ridges and grabens “at scales larger than almost anywhere else in the Solar System” [1 ¶3]. The team asked what interior and orbit could have cracked it this way [1 ¶4].
Fractures, ridges and grabens (blocks of crust that have dropped below their surroundings) [1 ¶3][1 ¶5].
On an eccentric orbit, the tidal bulge Uranus raises on Ariel grows and shrinks, and rocks back and forth, as the moon’s distance from the planet changes [1 ¶5].
Find the orbit and interior whose stresses could have cracked the surface this way. The answer requires an ocean [1 ¶7][12].
Thin ice, very deep ocean
Thicker ice, shallower ocean
“In order to create those fractures, you have to have either a really thin ice on a really big ocean, or a higher eccentricity and a smaller ocean,” said co-author Alex Patthoff. “But either way, we need an ocean to be able to create the fractures that we are seeing on Ariel’s surface” [1 ¶7].
The deep ocean goes with the least eccentric orbit, at least 0.04, and a shallower ocean needs an even more stretched one [1 ¶7][5]. The team’s 2024 abstract presented the thin-ice case as its result: “Our modeling suggests a thin (≤ 30 km) ice shell during a period when Ariel’s eccentricity was ~0.045” [12]. It framed Miranda the same way, with a crust of 30 km or less over an ocean of 100 km or more [15 Abstract].
With Ariel’s density and rock near 3 g/cm³, a 170 km ocean would take up about 85% of the water layer’s thickness, leaving roughly 25–35 km of ice, in line with that abstract.
3 · A small number with a big effect
“While 0.04 may not sound dramatic, eccentricity can strengthen the effects of tidal stresses,” the release says, and “to the eye, the orbit will still resemble a circle” [1 ¶6].
Tides weaken with the cube of the distance, so at an eccentricity of 0.04 the tidal stretch would swing by roughly ±12% every orbit: 13% stronger than at the average distance at the closest point, 11% weaker at the farthest. That flexes the ice every 2.5 days, the length of Ariel’s orbit [13]. Today the swing is less than ±0.5%.
The team’s 2024 abstract pointed to a past resonance with the moon Umbriel, as earlier dynamical studies had proposed, without saying which one [12]. A 2026 study looked at the 2:1 resonance with Umbriel, which older work ruled out but which, it argues, Ariel almost certainly crossed if Uranus’s tides are as strong as recent estimates suggest. It found that this resonance lifts Ariel’s eccentricity only to about 0.02, and called reaching 0.04 “very difficult” [5]. How Ariel got to 0.04 is still a hypothesis.
4 · Less than half a moon, seen once
Voyager 2 imaged only about 40% of Ariel’s surface, at a resolution relatively low compared with missions to Europa or Enceladus [3 ¶7][17]. The northern hemispheres of the Uranian moons were largely unobservable during the flyby and were not imaged [18 Abstract].
The team says its results predict where fractures and ridges should appear in the unseen north, a test a future spacecraft could run [1 ¶9].
5 · What would settle it
The study argues an ocean was needed to explain the cracks. Testing that takes new measurements, and some of them could only find an ocean that survives today.
Magnetometry. A salty ocean would answer Uranus’s changing magnetic field with an induced field of its own. Lead author Caleb Strom says magnetometry could provide definitive evidence of an ocean that still exists [3 ¶10], and modelling suggests even a single close flyby could detect one [9 Abstract].
Gravity and shape. How much Ariel deforms under Uranus’s pull would help show whether a liquid layer lies beneath the ice today; models of the moons’ interiors predict this tidal response [8 Abstract].
The northern hemisphere. Imaging it, ideally from an orbiter arriving near equinox, would test the predicted fracture pattern and so the account of the past ocean [3 ¶9][1 ¶9].
Models suggest that any ocean left today would be a residual layer less than 30 km thick in Ariel, and an ocean rich in ammonia might conduct too poorly to detect [8 Abstract].
6 · Our assessment
This section is our assessment, from the release, the team’s abstract and the coverage; we could not read the paper.
The release says “evidence of a past, deep ocean”; the paper’s title says “constraining ocean and ice shell thickness” [1][2]. The ocean is what a model needs to explain the cracks.
Every version of the model needs an eccentricity of at least 0.04, and how Ariel got there is unsettled: a 2026 study finds the 2:1 resonance with Umbriel reaches only about 0.02 [5]. If Ariel never got there, the stresses that crack the ice need another source.
The release opens by saying “growing evidence suggests” an ocean “lurks beneath the icy surface” today [1 ¶1], but this study is about the past: the team’s 2024 abstract put the thin shell “in the geologically recent past” [12], and the researchers aren’t sure exactly how long ago the deep ocean might have existed [1 ¶10].
By our estimate, Ariel’s whole water-ice layer is about 150–240 km thick, about 200 km if its rock is near 3 g/cm³. In that case a 170 km ocean would be most of it, leaving roughly 25–35 km of ice; if the rock is lighter than about 2.7 g/cm³, there is little or no room for it. A big claim, but the one the team’s model points to, not an extreme reading of it.
Everything rests on 1986 images of about 40% of the surface, all in the south, at Voyager-era resolution, which the team acknowledges [1 ¶9][3 ¶7].
The release came out on September 29, 2025. ScienceDaily ran a rewritten version, with new subheads but the same quotations, on September 24, 2026, under a headline ending “Uranus’ moon Ariel may have hidden a 100-mile-deep ocean,” with the same artist’s impression [6][1 ¶0].
Bottom line: a solid addition to the case that Uranus’s major moons were once ocean worlds, alongside the same team’s similar result for Miranda [1 ¶8][15 Abstract]. Read “100-mile-deep ocean” as the thin-ice solution of a model that needs Ariel’s orbit to have been at least 30 times more stretched than today’s, not as a measurement.
Bracketed numbers refer to this list; ¶ gives a paragraph of a web page, counted as its entry explains, and Abstract names the part of a work we used. Links checked October 13, 2026.
Kelley, M. M. (2025). Evidence of a past, deep ocean on Uranian moon, Ariel. Planetary Science Institute, September 29, 2025. ¶ counts the paragraphs of its text; ¶0 is the photo caption.
Strom, C., Nordheim, T. A., Patthoff, D. A., and Fieber-Beyer, S. K. (2026). Constraining ocean and ice shell thickness on Ariel from surface geologic structures and stress mapping. Icarus 444, 116822. We could not read this paper or its abstract.
SETI Institute (2025). The Hidden Ocean of Ariel: Tidal Forces and the Case for a Watery Past. November 7, 2025. A write-up of a SETI Live conversation with Caleb Strom and Alex Patthoff. ¶ counts the paragraphs of its text after the photo caption, not counting the “At a Glance” list.
NASA Science. Uranus: Exploration. Accessed October 13, 2026.
Lari, G., and Rossi, M. (2026). Tidal heating of the Uranian moons during the 2:1 resonance between Ariel and Umbriel. Europlanet Science Congress 2026, EPSC2026-164. Conference abstract.
“We just need to go back” – Uranus’ moon Ariel may have hidden a 100-mile-deep ocean. ScienceDaily, September 24, 2026. Source: Planetary Science Institute.
Tomaswick, A. (2025). Ariel Had A 170km Deep Sub-Surface Water Ocean. Universe Today, October 2, 2025. CC BY 4.0.
Castillo-Rogez, J., Weiss, B., Beddingfield, C., Biersteker, J., Cartwright, R., Goode, A., Melwani Daswani, M., and Neveu, M. (2023). Compositions and Interior Structures of the Large Moons of Uranus and Implications for Future Spacecraft Observations. Journal of Geophysical Research: Planets 128(1), e2022JE007432. CC BY-NC 4.0.
Cochrane, C. J., Vance, S. D., Nordheim, T. A., Styczinski, M. J., Masters, A., and Regoli, L. H. (2021). In Search of Subsurface Oceans Within the Uranian Moons. Journal of Geophysical Research: Planets 126(12), e2021JE006956. CC BY-NC 4.0.
NOAA National Centers for Environmental Information. World Ocean Volumes (ETOPO1). Accessed October 13, 2026.
JPL Solar System Dynamics. Planetary Satellite Physical Parameters. Accessed October 13, 2026.
Strom, C., Nordheim, T., Patthoff, A., and Fieber-Beyer, S. K. (2024). Constraining ocean and ice shell thickness on Ariel using geologic structure mapping and stress modeling. GSA Connects 2024, Anaheim, paper 91-1. Geological Society of America Abstracts with Programs 56(5). The team’s conference abstract for the same study.
JPL Solar System Dynamics. Planetary Satellite Mean Elements. Accessed October 13, 2026.
Williams, D. R. Uranian Satellite Fact Sheet. NASA Space Science Data Coordinated Archive, updated March 11, 2023.
Strom, C., Nordheim, T. A., Patthoff, D. A., and Fieber-Beyer, S. K. (2024). Constraining Ocean and Ice Shell Thickness on Miranda from Surface Geological Structures and Stress Modeling. The Planetary Science Journal 5(10), 226. CC BY 4.0.
JPL Solar System Dynamics. Approximate Positions of the Planets. Accessed October 13, 2026.
Tonoian, S., Lucchetti, A., Massironi, M., Penasa, L., Beddingfield, C., Pajola, M., Rossi, C., and Pozzobon, R. (2026). Geometry of chasmata on Ariel’s surface and its implications. Europlanet Science Congress 2026, EPSC2026-490. Conference abstract.
Cartwright, R. J., Beddingfield, C. B., Nordheim, T. A., Elder, C. M., Castillo-Rogez, J. C., Neveu, M., et al. (2021). The Science Case for Spacecraft Exploration of the Uranian Satellites: Candidate Ocean Worlds in an Ice Giant System. The Planetary Science Journal 2(3), 120. CC BY 4.0.
Method
On October 13, 2026, we checked every statement, number and source on the original infographic, 73 items, against the Planetary Science Institute’s release, the SETI Institute’s write-up, ScienceDaily’s version, the team’s 2024 conference abstract, NASA and JPL data and the other numbered sources. Like the infographic’s author, we couldn’t read the paper: the publisher’s site, which has listed the accepted manuscript as free to read since September 28, 2026, showed us a bot check, and we found no other copy or abstract. We measured the original drawings against their labels and re-ran the estimates. The check corrected several statements; the change log lists them.
After the page was built, a second reviewer who hadn’t seen the drafting checked it against fresh copies of the sources and re-computed the interior, tide, orbit, coverage and drawing numbers; all reproduced. The changes that followed: the team’s 2024 abstract doesn’t say which resonance with Umbriel stretched Ariel’s orbit, so the page no longer calls the 2:1 resonance studied in 2026 the team’s or the likeliest explanation, only one candidate; the release’s hedges (“growing evidence suggests”, “might have existed”) are kept; the claim that the fractures need an ocean now gives the model’s two cases and its orbit premise; the 25–35 km of ice holds for rock near 3 g/cm³, while lighter rock leaves little room for a 170 km ocean; and a few smaller wording fixes.
The verdicts and the items tagged “Our analysis” are our own assessment, built on the cited sources: Ariel’s interior and the ice left above a 170 km ocean, the tidal swing, and our reading of the model’s trade-off. The drawings are our own, to scale where they say so.
This page was researched and written with Claude, an AI assistant made by Anthropic, and then re-checked against its sources in a separate verification pass.
Corrections
If we get something wrong, we fix it and record the change below with the date. If a correction changes a verdict, the log says what the old verdict was.