The presence of ice at two positions on opposite sides of the Moon suggests that the satellite's orientation was once shifted away from its present spin axis — a finding that has implications for the Moon's volcanic history. See Letter p.480
The Moon's north and south poles are among the coldest regions in the Solar System — some areas are colder than Pluto. Water ice can remain stable here for billions of years, even when exposed to the vacuum of space. Scientists have detected small deposits of ice at the lunar poles; these have told us about how water moves across the lunar surface, but on page 480 of this issue, Siegler et al.1 present measurements of lunar ice that tell us about events deep inside the Moon. These events cast light on the Moon's volcanic evolution and on its orientation in the sky. They also reveal that the face of the Moon seen today is different from the one that looked upon Earth billions of years ago.
Since at least 1961, scientists have speculated that volatile compounds such as water could slowly accumulate in the Moon's cold, shadowed polar craters2. The temperatures at these craters are so low because the Moon's spin axis is nearly perpendicular to the line connecting the Moon and the Sun (Fig. 1). Lunar seasons are therefore nearly non-existent, so the bottoms of some polar craters never see sunlight and hence stay cold enough to trap any water molecules that fall there. Unfortunately, the darkness in these craters also makes it hard to determine how much ice is in them. However, in 1998, scientists used an ingenious method to infer the presence of ice by observing how neutrons produced by space radiation interact with the hydrogen atoms in any near-surface water3.
Using hydrogen data from more than a decade ago, Siegler et al. noticed that each lunar pole has a hydrogen deposit that is slightly displaced from the true north and south poles. Moreover, each of these off-axis deposits are symmetrically arranged, such that if the two poles are viewed on top of one another, a 180 ° rotation of the Moon would bring the deposits into perfect alignment. The authors therefore suggest that the Moon's orientation to its spin axis was once shifted away from the present one. This would have produced darkness, coldness and the accumulation of ice at these ancient 'palaeopoles'. The ice that formed in that era seems to have survived when the Moon eventually shifted to its present orientation, despite now being exposed to some sunlight (Fig. 1).
Siegler and co-workers strengthen their hypothesis by proposing a plausible mechanism for such a shift. Planets can change their orientation if their internal mass distribution changes. Pockets of dense material tend to be close to the equator to minimize the planet's spin energy. If a huge pile of lead weights suddenly appeared in New York, the city's latitude would eventually shift to a position slightly southward because of planetary reorientation. The opposite is also true — if New York suddenly became lower in density, it would shift northward. This process is known as true polar wander.
The authors therefore show that the observed shift in ice poles could be explained if certain regions on the Moon became lower in density. One of these regions, the Procellarum KREEP Terrane (PKT), is the Moon's most radioactive area, and was once the most volcanic — this volcanism was responsible for most of the dark plains of the Moon that can be seen from Earth. The radioactivity and volcanism imply that this region must have once been hot and thus less dense than its surroundings. The idea that true polar wander caused the shifts therefore makes a lot of sense.
Siegler and colleagues' proposal has several implications for the history of lunar volcanism. If the PKT is indeed responsible for polar wander, it would help to constrain the magnitude of the heating events that occurred in that region. The scale of these events tells us how deep into the Moon the PKT reaches, and limits the timescales over which it became hot. This is important for understanding how the PKT became so radioactive in the first place — an enduring mystery in lunar science.
But it also presents a new puzzle. The Moon's volcanism mostly stopped 3 billion years ago, which means that the PKT has probably been getting colder and denser since then, not hotter. The direction of polar wander during this period would therefore have been in the opposite direction to the wander that produced the ice palaeopoles (see Fig. 4 of the paper1). Furthermore, one might expect ice palaeopoles to have formed everywhere along this polar-wander path, raising the question of why they are found only at the locations observed in the current study.
The authors' proposal also has implications for how long ago the ice deposits formed and how they have been preserved. The heating event that drove them to their present position began billions of years ago, implying that the deposits are similarly old. If so, how could they have remained there for so long, given that they are in sunlight? One explanation is that they must be buried below the equivalent of a permafrost level. Something similar happens on Earth — in Alaska, for example, temperatures below freezing occur a metre below the surface, even in summer.
But the Moon's soil has been churned over and over by countless asteroid impacts, which should have liberated some of this water. However, such impacts can also bury and protect ancient ice deposits. The details of the equilibrium between liberation and burial require further study, as does the exact nature of the ice and hydrogen at the poles.
Finally, as Siegler and co-workers acknowledge, their polar-wander hypothesis must be reconciled with other independent estimates of polar wander4. The amount of wander proposed by the authors is relatively modest, about 6 °. Other work suggests that the Moon might have changed its orientation along a similar axis by 35 ° (ref. 5), or that much bigger changes occurred along multiple axes6. A key goal will be to reconcile these many stories of the changing orientation of the Moon, and to determine what density changes drove it to wander.
Siegler, M. A. et al. Nature 531, 480–484 (2016).
Watson, K., Murray, B. & Brown, H. J. Geophys. Res. 66, 1598–1600 (1961).
Feldman, W. C. et al. Science 281, 1496–1500 (1998).
Runcorn, S. K. Nature 304, 589–596 (1983).
Garrick-Bethell, I., Perera, V., Nimmo, F. & Zuber, M. T. Nature 512, 181–184 (2014).
Takahashi, F., Tsunakawa, H., Shimizu, H., Shibuya, H. & Matsushima, M. Nature Geosci. 7, 409–412 (2014).