Resume Reading — When the Earth Had Two Moons

Close

When the Earth Had Two Moons

A new model—“The Big Splat”—explains the strange asymmetry of the moon.

For more than half a century, the moon had been mocking the best minds in science, and for Erik Asphaug enough was enough.

The taunting began three years before Asphaug was born. On Oct. 7, 1959, the Soviet Luna 3 spacecraft looped behind the moon, snapping off a series of grainy but distinct photos and then radioing them home. Because the moon’s rotation is perfectly synchronized with its revolution, one hemisphere always points toward Earth while the other always points away, unseen. Luna 3’s first-ever images of the lunar far side revealed an expanse of rugged, blandly gray highlands—a vista utterly unlike the near side’s charismatic, Man-in-the-Moon markings. It didn’t take a planetary scientist to recognize the weirdness of that split personality. “I remember as a boy seeing one of the news programs showing the far side of the moon, and thinking it was incredible that a planet could be so different on each side,” Asphaug says.

Now it was 2010 and here Asphaug was, a professor of earth and planetary sciences at the University of California, Santa Cruz, attending a colloquium, still waiting for an explanation for the moon’s aggressive asymmetry. He listened, increasingly antsy, as his colleague Ian Garrick-Bethell sketched out his proposed answer. In this latest theory, Earth’s gravity raised powerful tides on the moon billions of years ago, while it was young and molten. The bulges then froze in place, giving rise to the thicker crust and distinctive geology of the far side. The concept made no sense to Asphaug. “You’d get a bulge on both the far side and the near side, just like when you have high tide on Earth,” he says. But the whole point of the theory was to put a bulge on the far side only. “So the answer has to be that some miracle happens to erase the other half. It makes the problem even worse than before.”

Asphaug was not only annoyed; he was inspired. For years he had been working to develop models of low-velocity impacts in the early solar system. “People have been biased, looking at impacts and thinking only about hypervelocity events,” he says. “People forgot that things can hit at lower velocities.” These kinds of events are constructive rather than destructive: If two objects collide slowly enough they bump and stick together, “like throwing mud at the wall of a house or throwing snowballs at each other.” Asphaug had been thinking that low-velocity impacts, what he liked to call  “splats,” could explain how comets formed. Suddenly he realized he might have the solution to the moon problem sitting right in front of him. He grabbed one of his post-docs, Martin Jutzi (now at the University of Bern), and spelled out his idea. What if Earth originally had two moons, which only later merged into the one we know?

“We went to the lab right after that seminar and Martin coded up the moon being hit by a companion moon,” Asphaug says. The result of those computations was a novel interpretation of lunar asymmetry. In Asphaug’s view, the jumbled lunar highlands are the wreckage of a second moon that once orbited the Earth, pasted onto the surface of the moon. Small wonder that the far side looks like a different world; it is a different world. The new model provides an integrated description of the moon’s ancient origin and its modern appearance, but to Asphaug the concept goes deeper than that. It showcases a broader, and largely overlooked, process in planetary formation: the gentle collision, in which two bodies come together in a kiss.


Like most theories in science, Asphaug’s Big Splat model is sculpted from the marble of previous research. In fact, the first truly scientific description of the origin of the moon is also centered on the interaction between two worlds, but it envisioned a separation rather than a joining. In 1878 George H. Darwin, son of Charles, proposed that the moon was flung off from the fast-spinning newborn Earth, like a wayward child falling off a merry-go-round. The missing chunk of our planet is still evident, he speculated, as the basin of the Pacific Ocean.

That “fission hypothesis” stuck around a long time. I remember reading about it as a child, in the dreamy pre-Apollo science books I inherited from my older brothers—books that predated Luna 3 by a few years. But from a dynamics perspective, it doesn’t work. There is no plausible way to get the Earth spinning quickly enough to cast away part of its surface, and even if there were there is then no way to siphon off enough angular momentum to match the current physical state of the Earth and moon. As for the Pacific Ocean, it is a transient feature associated only with the modern arrangement of the continents. When the moon formed, about 4.4 billion years ago, the Pacific did not exist. (Go easy on Darwin; his theory predated the concept of plate tectonics by nearly 90 years.)

What it does not do is answer the big, shining question hanging in front of us: Why the hell does the moon look like that?

Other lunar origin stories followed, each positing a different relationship between Earth and moon, and each suffering from its own fatal flaw. American astronomer T.J.J. See proposed that the moon formed as an independent planet and was then captured into orbital servitude by the Earth. Flaw: Our planet’s gravity is not strong enough to have set the snare. A number of other researchers, most volubly French mathematician Edouard Roche, argued that the moon and Earth coalesced side by side in the early solar system. Flaw: When geochemists started digging into the 842 pounds of rocks returned by the Apollo astronauts, they found that the moon’s overall composition is distinct from Earth’s, deficient in easily vaporized compounds known as volatiles. A parallel world should be a much closer match.

Yet in many other ways the moon’s chemistry is very similar to Earth’s—strangely similar, in fact. The ratio of two different types of oxygen in those lunar rocks matches terrestrial rocks almost perfectly. The oxygen ratio is like an ID tag that tells you where an object formed. Meteorites have their own ratio. Mars has its own ratio. The moon just looks like a twin of the Earth. There you have it: The moon is nothing like the Earth, except that it is also exactly like the Earth. Another, more subtle asymmetry needing to be explained.

Splat! How the Moon Was Formed: This simulation of a collision between the proto-moon and a companion moon shows the lunar masses becoming lodged and producing asymmetrical hemispheres. According to planetary scientist Erik Asphaug, this process could explain how Earth’s moon became lopsided. The colors indicate a crust (light blue), mantle (dark blue), and a layer of upper mantle material (yellow) representing a magma ocean. Most of the companion moon is accreted as a pancake-shaped layer, forming a mountainous region comparable to the farside highlands on the Earth’s moon.M. Jutzi (U. Bern), E. Asphaug (ASU, UCSC)

Planetary scientists digested all the convoluted evidence and settled on a new creation narrative, the Giant Impact theory, inspired by two groundbreaking 1975 papers. In this model, Earth was born moonless. Then shortly after its formation it collided violently with a Mars-size body, commonly referred to as Theia (the mother of the moon goddess Selene in Greek mythology). The resulting inferno vaporized Theia along with a substantial part of Earth’s outer layers. Some of the material was blasted to kingdom come, but much of it settled into a disk around the bruised Earth. Within a very short period of time—perhaps as little as a decade!—the disk had condensed to form the moon.

The Giant Impact theory handily accounts for most of the broad chemical and dynamical attributes of the Earth-moon system. Despite lingering uncertainties, it is almost universally regarded as the best explanation for why the moon exists at all. What it does not do is answer the big, shining question hanging in front of us: Why the hell does the moon look like that?

It is not just an issue of cosmetics. The Man-in-the-Moon markings are enormous plains of dark, frozen lava, known as maria. For some reason, almost all of the maria are on the near side. More generally, the early moon was much more volcanically active on the Earth-facing hemisphere. NASA’s moon-orbiting GRAIL (Gravity Recovery and Interior Laboratory) mission measured the lunar crust and confirmed that it’s thicker on the far side; the spacecraft also found that the near side (and the near side only) has a network of long, linear, buried features, which planetary geologists interpret as volcanic dikes. The obvious-seeming answer is that Earth’s gravity is to blame for all this imbalance, but there is no physically plausible way to make the connection.

“People have been trying to explain the asymmetry for a long time. Ideas keep getting put forward, they get kicked around for a while, and they often don’t fare very well,” Asphaug says. More often, the issue of the moon’s double identity gets swept aside entirely. Last fall, David Stevenson of Caltech co-hosted a major conference on lunar formation at the British Royal Society. He chuckles at the way many of his colleagues talked around the problem: “There’s this elephant in the room, but people are sort of ignoring it and they’re looking at the cat over in the corner.”


In many ways, Asphaug’s model is an organic extension of the prevailing ideas of lunar formation—in truth, of the whole current thinking about how planets form. The infant solar system gradually built up from dust to rocks to asteroids to planets, all within a swirling disk around the newborn sun, starting a little over 4.5 billion years ago. That story is well established by a library of evidence, from the analysis of meteorites to observations of similar disks around other infant stars. As the planets came together, increasingly large objects had to crash into each other. Some of those crashes had to happen gently enough that the two objects got pasted together into a single, larger one, Asphaug notes. Otherwise the planets would have stopped growing and our solar system would be a swarm of fractured rubble.

Asphaug has spent a career exploring the subtleties of the collision process, first at Santa Cruz and, since 2012, at his current home, Arizona State University. Over the years, he contributed significantly to the mainstream Giant Impact model, helping to refine the mainstream model of Theia as a Mars-size interloper. In the process, he says, “I woke up to the fact that accretion is imperfect and complicated.” He became interested in what he calls “almost-accretion” events, ones in which two planets collide incompletely; he came to regard the Giant Impact as just one case in a larger set of possible encounters, ranging from sloppy mergers to hit-and-run impacts in which both planets survive, mostly intact.

“My big learning experience in grad school was realizing how little you had to do before you get to the limits of knowledge in planetary science. It’s not like physics where there’s 400 years of stuff you have to get through before you can find a problem that’s novel and unsolved,” Asphaug says. “In planetary science, there are a lot of fundamental questions with no good answers. You start to say, ‘That’s what you think? Really? OK, well I’ve got an idea, too.’ ”

His colloquium-room epiphany got Asphaug thinking about what happened right after the Giant Impact. Planetary scientists typically assumed that a single moon emerged from the conflagration, but that was not the only possibility. Asphaug’s idea that two moons emerged dovetailed perfectly with his computer modeling for comets. “We had splats on the mind. It was kind of a coincidence,” he says.

Not one moon but two? A celestial impact that produces a landslide instead of a catastrophe? Yeah, let us get back to you on that.

Matija Cuk at Harvard University had already surmised where a second moon might have come from, filling in a second part of the picture. When the giant impact created a huge, Earth-circling ring of vaporized rock, Cuk noted, there are two points of stability, located 60 degrees ahead of the moon and 60 degrees behind in the same orbit. Those spots are like gravitational tidal pools, calm locations where a second, smaller object could have formed. Cuk did not give this companion moon a name, but for the sake of clarity I’ll call it Endymion, after Selene’s lover in Greek mythology.

At first, the moon and Endymion coexisted peacefully, but (as often happens in romantic relationships) things changed over time. Gravitational interactions caused the main moon to spiral away from Earth, shifting the dynamics of the whole system. Endymion lost its steady perch and fell into its lover’s arms. Cuk theorized that Endymion eventually broke up and rained down on the moon like a hailstorm of asteroids. Asphaug saw a quite different possibility.

“When things co-orbit they hit each other pretty slowly, at about the speed things would fall, so it doesn’t create a shock or a lot of melt,” Asphaug says. “Instead of thinking of this as a gargantuan impact, it’s more like a cosmic landslide. It would be one of the largest landslides you can imagine.” In the later stages, the remains of Endymion would be falling at a rate of few hundred meters per second, just 1/100th the speed of the giant impact that got the process rolling in the first place.

“If you do the calculation you get a bulge where the impactor hits; it’s a thickening crust made of slightly lower density material,” Asphaug says. The shape of the bulge in his model matches the form of the actual crust on the lunar far side. He estimates that Endymion needed to be about 621 miles (1,000 kilometers) wide to account for the mass of that additional crust; that is a plausible size, about one-quarter the diameter of the modern moon. Such a small body would have cooled to solid rock within 100 million years, the approximate age at which Endymion would have become unstable. As it crumbled down onto the lunar surface, it would have created a hemisphere-wide rubble field, nicely accounting for thicker crust on the moon’s far side.

The Big Splat has additional intriguing consequences. The weight of all that rock landing on one hemisphere of the moon would compress molten material, rich in radioactive elements, and force it toward the other hemisphere, Asphaug notes, “so it creates not only a crustal asymmetry but also a thermal asymmetry.” A concentration of heat on one side of the moon could also cause it to expand, creating the dikes seen by the GRAIL spacecraft. Voila: The thick crust of the far side, the volcanic activity of the near side, and the enigmatic lunar dikes, all explained in a single stroke.

Asphaug is less confident that the Big Splat can fully account for the egregious asymmetry that started this whole conversation, the concentration of dark markings solely on the moon’s near side. “It’s a bit of a stretch,” he concedes, because the lava flows that created the lunar maria erupted much later, about a billion years after the formation of the moon. But by explaining the thicker crust on the far side and the greater internal heat on the near side, his model at least sets up the right conditions for the lava to prefer one hemisphere over the other.


The more I listen to Asphaug, the more convincing he sounds. Although the Big Splat can’t account for all of the moon’s off-kilter anomalies, it does cover an awful lot of them. But in the glacier-speed world of lunar science, Asphaug’s ideas have received a lukewarm reaction so far. Not one moon but two? A celestial impact that produces a landslide instead of a catastrophe? Yeah, let us get back to you on that.

Jay Melosh, a veteran planetary scientist at Purdue University and a member of the GRAIL team, raises a concern that the density of the lunar crust is the same on the near side as the far side. It’s possible that Endymion simply had the same density as the surface of the moon, but the lack of diversity “at least makes it more difficult to maintain Asphaug’s Big Splat model,” he says. Stevenson is similarly cautious. “I don’t find it convincing but I would not exclude it,” he says. “It doesn’t emerge naturally from what we think happens after a giant impact.” Dynamics experts like Cuk are much more supportive of the model. In the end, though, the primary critique is not that there’s any evidence disproving the Big Splat—there isn’t, really—but that there isn’t enough uniquely supporting it.

Asphaug is sympathetic. “We may have a hypothesis that explains a lot of things but is relatively untestable,” he says. “We’re starved for data right now.” One way forward would be to place a network of seismic stations on the moon, which would make it possible to read the complete history recorded by its internal structure. Whatever happened during the giant impact (if there was one) and the subsequent Big Splat (ditto) must have left its mark deep inside. In two years NASA will launch InSight, a high-precision seismic station for Mars, but every proposal to send an equivalent mission to the moon has been shot down. “It’s a sad commentary about how America just put an end to landed lunar missions in the 1970s,” Asphaug says. “I think it’s going to be the Chinese who will finally get answers.”

The ultimate Man-in-the-Moon message, according to Asphaug, is that planet formation is sloppy and creative and wildly varied.

Fortunately, that glum note is not where the story ends, because there is another way—a whole set of ways—to test Asphaug’s model and to learn how planets and moons formed and evolved. Evidence of “almost-accretion” events, he notes, may be all around us. Comets may preserve ancient layered splat structures; the European Space Agency’s Rosetta mission, now on its way to Comet Churymov-Gerasimenko, will be able to take a look. Pluto and its large satellite Charon may have formed in a giant impact analogous to the one that formed Earth’s moon. The New Horizons spacecraft will make a close flyby in 2015. Mars has its own peculiar north-south asymmetry, and the InSight probe will help investigate whether an impact might be to blame there, too.

As Asphaug continues to run his models, he keeps coming up with more places to investigate. In a paper published last year in Icarus he argues that Saturn’s complex satellite system might be the result of multiple impacts and mergers, a more elaborate version of the process that made our own moon. His latest idea, now in press, is that the planet Mercury was the victim of one or more hit-and-run impacts as it formed, which would account for both its dense, iron-rich structure and for the improbable presence of water in its searing-hot crust.

The ultimate Man-in-the-Moon message, according to Asphaug, is that planet formation is sloppy and creative and wildly varied, in ways most of his colleagues have overlooked. He wants to make sure he is not missing any part of it. “With modern computer codes, we can explore a very large parameter space,” he says with the same geek fervor that propelled him into the lab in 2010. “The fun never stops.”


Corey S. Powell is editor at large at Discover magazine, where he writes the Out There column and blog, and acting editor of American Scientist. He is also a late convert to Twitter: @coreyspowell

25 Comments - Join the Discussion