AirSpaceMag.com

Did two sub-moons collide to form our Moon? From Jutzi and Asphaug, Nature 476, 4 August 2011.

A recent paper suggests that early in the history of the Solar System, two sub-moons collided to create Earth’s present-day Moon.  Several people have asked for my opinion on this new concept, so I will examine how this result was obtained, along with some general remarks on the nature of modern scientific research.

Over 25 years ago, a popular model for the origin of the Moon emerged at a special conference on the Moon held in Kona, Hawaii.  Whenever I mention that we had a conference in Hawaii, snickering about exotic travel boondoggles invariably follows, but you should note that at this particular conference, it was hard to get attendees out of the meeting room – the tension and excitement of a new and revolutionary discovery was that great.  The collective understanding of the then-current models of lunar origin was that they were all inadequate in one way or another.  But at Kona, a “new idea” was advocated – that a giant impact sprayed material into orbit around the Earth and that debris coalesced into the Moon.   This concept was supported by nearly all attendees and affectionately became known as the “Big Whack” model.  It seemed to satisfy most of the important physical and chemical constraints on lunar origin.  Subsequent work elaborated on the details concerning this model, but its salient features were pretty well defined at Kona in 1984.

The Big Whack has subsequently entered the realm of “settled science” in regard to lunar origin, although some dissenters remain.  But a “consensus” of working lunar scientists seemed satisfied that the origin of the Moon had become a “solved problem.”  Much of the detailed information on such a planetary scale collision comes from computer modeling, in which the basic physical parameters such as size of the two bodies, impact speed, angle of encounter, and composition in broad terms are specified as input variables.  The output of the computer model tells us how much material was vaporized, melted and ejected, and how fast the ejecta was squirted out and where it was deposited.  As you might expect, these calculations are extremely involved, requiring advanced supercomputers working day and night for weeks to churn out the results.

Some scientists tend to be skeptical of purely computational results.  In computer modeling, results are only as good as the input values and assumptions, the realism of the model, the inevitable simplification necessary to make the model fit into the computer and how carefully and thoughtfully the results are interpreted.  After the first few Big Whack computer models were run and presented at scientific conferences, various lunar workers would advance questions or problems that weren’t well explained by the existing models.  The models were tweaked to accommodate the difficulties.  In fact, it seemed that the models were amenable to endless tweaking.  If a tweak couldn’t be found, the observation was questioned or deemed irrelevant.  Models should be flexible enough to explain data outliers and the odd inconvenient fact, but they should also make predictions that can be tested by experiment or observation.  A model that is infinitely flexible ultimately is scientifically worthless.

So in regard to the origin of the Moon, we find ourselves with a solved problem for which a strong consensus of the experts exists.  Big Whack skeptics either have poor or irrelevant observations or are right-brained, qualitative geoscientists incapable of understanding complex planetary “physics.”

Which brings us back to Two Moon Junction.  The recent study suggesting that the Moon is the product of the collision of two sub-moons is an outgrowth of the same type of computer modeling done on problems in planetary accretion, including the Big Whack.  What’s unusual in the new scenario is that the two objects are relatively small to begin with (not Earth-sized, but a few hundreds of kilometers across) and collide at relatively low velocities, less than 2 km/sec.  The result of these unusual conditions, it is claimed, is that the impactor “plastered” itself onto the larger object, without forming a crater.  This “spackling” of matter adds an anomalously thick crust to the far side of the Moon and shoves semi-molten, late-stage liquids around to the near side, simultaneously accounting for two major lunar conundrums – the thicker far side crust and the concentration of KREEP (potassium, rare earths, and phosphorus) on the western near side of the Moon.

Sounds pretty good, eh?  Well, there are some issues with it.  The idea that a low velocity impact does not make a crater is counter-indicated by the existence of secondary impact craters on the Moon.  Secondary craters are made when blocks and clouds of debris ejected from an impact crater land on the Moon and dig up new craters, either as isolated single holes or as chains and clusters of multiple craters.  Since these features are formed by material thrown from the Moon’s surface, they cannot have been created at speeds greater than lunar orbital velocity (about 1600 m/sec).  Since the ballistic range for most secondaries is typically less than a few tens of kilometers from the primary, most were formed by impacts at much lower speeds, typically less than 1 km/sec.  Moreover, the addition of the far side crust as a sedimentary layer does not jibe with the observation that the lunar crust is a laterally contiguous global layer, composed everywhere of similar rocks (but varying in proportion).  The authors of the study acknowledge this is an issue, but suggest that the two sub-moons would have already formed their own crusts, probably of the same composition since they come from the same region of the Solar System.  This explanation appears rather ad hoc and elastic to me, an example of the “flexibility” for which computer models are renowned.

The Big Splat has not yet been embraced by most of the lunar science community, but will doubtless be examined and considered by many.  At this stage, it remains a model and not a description of reality, but rather, the description of a possible reality.  The distinction is important.  Neither the “votes” of the lunar science community nor the “elegance” of the model are relevant in terms of its validity.  The authors describe some possible tests of their model in the paper, but these seem to me neither particularly conclusive nor easy to accomplish.

So were there originally two moons over Miami (or rather, where Miami would one day exist)?  Maybe.  But the fact that someone can make a computer model of a complex process is not proof of its reality.  In this and similar cases, the burden is on its proponents to offer experimental tests or observations to prove their case.  In the mean time, nothing is settled and consensus is irrelevant.