NASA has sold the Artemis program to Congress, investors, and the public partly on a simple idea: water ice locked in the permanently shadowed craters at the lunar south pole is the key to staying. Drinkable, growable, burnable into rocket fuel. The resource that makes a permanent human presence affordable. A new radar study suggests the geology underneath that assumption is more complicated than anyone advertised.
Lynn Carter at the University of Arizona has been mapping the subsurface of the lunar south pole using two radar instruments in combination: Mini-RF, which operates on the Lunar Reconnaissance Orbiter spacecraft at S-band wavelength, and a P-band radar signal bounced off the Moon from Earth using the Arecibo Observatory's successor facility. The two wavelengths probe to different depths. S-band radar waves penetrate about a meter into lunar material; P-band radar reaches several meters, according to the EPSC abstract. The composite gives a picture of both surface texture and what lies beneath. What Carter's team found, presented at the Europlanet Science Congress and Division for Planetary Sciences joint meeting in 2025, is a subsurface layered with material that did not originate where it now sits.
The lunar south pole is not solid rock down to some convenient depth. It is mantled in melt sheets: layers of rock that were liquefied by the heat of asteroid impacts billions of years ago, thrown outward from the crater, and then re-solidified as flat sheets draped across the landscape. These are not rare or exotic. The data suggests they are widespread across the south pole region, and they are thick. The melt sheets Carter's team modeled run between 10 and 100 meters, buried under 5 to 10 meters of regolith, the loose soil that covers most of the lunar surface. They underlie many of the permanently shadowed regions, the cold traps at the bottoms of craters where water ice has been confirmed to exist.
The water ice is real. The SOFIA airborne telescope detected water at the south pole in 2020, and LCROSS confirmed water ice in Cabeus crater by crashing a spacecraft into it in 2009. What the new radar data adds is a question about what lies beneath that ice. The melt sheets are not ice-bearing rock. They are solidified impact melt, and in some areas they sit directly below the regolith layer in which the ice is found, between it and whatever solid bedrock exists below. That changes the drilling calculus.
Artemis astronauts landing at the south pole will need to understand what they are drilling into before they commit to a site. An ice deposit sitting on top of a 50-meter layer of solidified impact melt is a different engineering problem than an ice deposit rooted in permeable regolith all the way down. The melt does not evaporate. It does not compress. It has to be moved, not extracted. None of this means the ice is inaccessible. It means the geology is not what the ISRU pitch assumed.
In-situ resource utilization, the industry term for living off what you find at the destination rather than carrying everything from Earth, is the financial foundation on which a growing constellation of lunar startups has built its pitch to investors. Water as propellant, water as drinking supply, water as the thing that makes a fuel depot at the south pole economically viable. Those pitches assume the ice is in ground that behaves like ground. A landscape underlain by sheets of ancient impact glass changes the extraction cost model in ways that are not yet quantified.
The Wapowski and Hale Q craters, both south polar features studied in Carter's radar data, show surface boulder fields and landslides consistent with steep slopes and unstable surfaces. Small craters south of Drygalski show ponded melt material on their floors, similar to melt ponds identified elsewhere on the Moon. These are not marginal features. They are the terrain Artemis is designed to operate on, and the new radar data is the first systematic map of what sits below it.
The practical implication is straightforward: the south pole is not a water bottle sitting in a cold trap. It is a layered archaeological site of ancient impacts, with ice near the surface in some areas and buried melt sheets below it in others. Before anyone drills, they need to know which areas have accessible ice and which have ice sitting on top of geology that was never part of the plan. The Artemis landing site selection process has always involved tradeoffs. The radar data adds a new column to that spreadsheet.
This is not a kill shot for lunar ISRU. The ice is there. The capability to extract it exists. But the subsurface map that investors have been working from may need a revision, and the crews planned to land at the south pole later this decade are the ones who will find out whether the geology cooperates.
