The Moon’s Mantle Muddle
Maybe we’ve been looking for the wrong minerals, or maybe our models are wrong.
Geoscientists want to peer ever deeper into the interior of planets. Most rocky planets have a tripartite configuration, with a dense, metallic core at center, overlain by a rocky, iron- and magnesium-rich mantle, and finally, a low-density crust rich in silicon and aluminum on the outside. A surprise from the exploration of the Moon was that it too is configured like the Earth, with a core, mantle and crust, although of different proportions. All our samples from the Moon come only from the crust, so understanding the composition and nature of its mantle has been a high priority for lunar scientists.
With the Moon, we are fortunate to have some natural “drill holes” with which to probe the subsurface. I refer of course to the multitude of large craters on the Moon, including those largest of impact features, the multi-ring basins. Basins are impact structures hundreds of kilometers across; the largest basin, South Pole-Aitken, is over 2500 km in diameter. Because basins are so large, they dig down many kilometers below the lunar surface. If we can identify material of deep derivation that has been thrown out during basin formation, we can characterize the lower crust and possibly, the upper mantle of the Moon.
Needless to say, an effort like this is fraught with difficulty. Basins are large features (at 930 km in diameter, the lunar Orientale basin is as big as the state of Texas) and we don’t fully understand the mechanics of their formation, including such rudimentary properties as their original size (it is suspected that the final diameters of these features were enlarged by collapse, making their original size uncertain). However, that doesn’t stop us from trying. Using basins as probes of the crust and mantle has been a pastime for lunar scientists for many years. At this year’s Lunar and Planetary Science Conference (LPSC 45), the topic of the mantle was raised once again, but this time with a bit of a twist.
Decades ago, a network of seismometers was emplaced on the Moon during the Apollo missions. This network measured the intensity and duration of “moonquakes” over the course of several years. These measurements allowed us to discover the mantle of the Moon and to estimate its density. The density is proportional to the velocity of seismic waves, which can be measured from the difference in arrival times of seismic waves at different stations for the same moonquake. From these data, we know that the mantle is composed of rock quite different in composition from surface rocks. The inferred relative density of the mantle is considerably higher – about 3.2 g/cm3 (grams per cubic centimeter; water = 1.0) than the crustal rock types (about 2.6 g/cm3). Although a single piece of information, this density constrains the mantle of the Moon to be composed of only one or two possible minerals – olivine and/or pyroxene (only these iron- and magnesium-rich minerals are common enough and match the density estimate). We also suspect such mineralogy because the Earth’s mantle is made up of these minerals.
For many years, lunar scientists have searched diligently for deposits of olivine around lunar basins, material that could plausibly be interpreted as ejected from the lunar mantle. Yet to date, few such deposits have been found, and those that are seen, could just as easily derived from crustal rocks, as olivine is a common mineral in the crust as well.
The recent gravity mapping of the GRAIL mission has added confusion on this score. Assuming reasonable densities for crust and mantle, gravity maps can be interpreted in terms of crustal thickness at any given area. The new GRAIL crustal thickness maps indicate a much thinner crust than previous estimates had shown, with a mean thickness of about 35 km, increasing to almost 45 km in some areas of thick crust. These values are about half the previous numbers and suggest that the largest impact events should have easily excavated the upper parts of the lunar mantle. The problem is that there doesn’t appear to be any mantle material on the lunar surface, even proximate to the biggest basins. This absence is quite puzzling; despite the ubiquity of olivine in many lunar rocks, we do not find vast exposures of it near the rims of any lunar basin.
At the recent LPSC 45, H.J. Melosh of Purdue University and his fellow co-workers suggested that we are looking for the wrong mineral. If the upper mantle were composed not of olivine, but a different mineral, large amounts of olivine would not necessarily be excavated by a basin-forming impact. Their calculus is as follows: the crust is thin (from GRAIL maps), computer models show that Orientale basin must have excavated the lunar mantle (from calculation), and we see pyroxene in mineral spectra of basin deposits but not olivine (from remote sensing data). Therefore, the mantle is made up of pyroxene, not olivine.
One might note that this chain of reasoning is a house of cards. IF the GRAIL measurements are telling us how thick the crust is and IF the computer models accurately reproduce basin-forming mechanics and IF the compositional data are correct, then a pyroxene mantle is required. In fact, of these ‘constraints,” only the compositional data are fact-based. GRAIL did not measure crustal thickness – it measured gravitational accelerations, objective measurements interpreted in terms of crustal thickness. Computer models attempt to simulate reality, but there is no way to directly test their validity. The scale of basin impact is so large – orders of magnitude beyond any impact event within our experience base – that unsuspected physical effects (possibly of critical importance) cannot be accounted for.
An alternative explanation is that as large as they are, basins do not come close to digging into the Moon deeply enough to excavate the mantle. At the same session, I presented the results of new research on the composition of ejecta from the Orientale basin, which show its deposits to be low in iron and probably derived only from upper crustal levels. This is quite surprising; the Orientale basin is one of the youngest basins on the Moon and at nearly 1000 km diameter, also one of the biggest. Virtually all of its ejecta come not only from crustal sources (as measured by remote sensing data) but also from upper crustal materials, suggesting that even the biggest impacts apparently cannot punch through the crust of the Moon.
So what does all this mean? I suggest that one or more of the existing lines of evidence is giving us misleading information: either the remotely sensed compositional data for basin ejecta, the GRAIL crustal thickness estimates, or the results of the computer impact models are wrong. I list the preceding in order of their likely verity, which also happens to be in order of their closeness to objective measurement and observation. There is a recent tendency in science to place the value of studies in an order reverse to what I have done. Models have taken precedence over observation in many fields, a perverse attitude that turns empiricism on its head and where the “elegance” of a mathematical model is preferred over some inconvenient, observable fact.
Science is constantly revising its corpus of knowledge and today’s “rational consensus” can be tomorrow’s superstitious nonsense. This debate about the nature of the lunar mantle clearly shows the cracks and fault lines in our understanding. Work will continue to map basin deposits in search of excavated deep materials – such features are still one of our best bets to sample and understand the deep lunar interior. As you hear about the results of these studies in the coming years, keep a skeptic’s eye on claims of revolutionary, new interpretations. Such ideas come and go, but knowledge should always be testable and constrained by observations and facts.