≡ Menu

Hit-and-Run: Earth, Venus and Planet-Shaping Impacts

The gradual accretion of material within a protoplanetary disk should, in conventional models, allow us to go all the way from dust grains to planetesimals to planets. But a new way of examining the latter parts of this process has emerged at the University of Arizona Lunar and Planetary Laboratory in Tucson. There, in a research effort led by Erik Asphaug, a revised model of planetary accretion has been developed that looks at collisions between large objects and distinguishes between ‘hit-and-run’ events and accretionary mergers.

The issue is germane not just for planet formation, but also for the appearance of our Moon, which the researchers treat in a separate paper to extend the model for early Earth and Venus interactions that appears in the first. In the Earth/Venus analysis, an impact might be a glancing blow that, given the gravitational well produced by the Sun, could cause a surviving large part of an Earth-impactor (the authors call this a ‘runner’) to move inward and subsequently collide with Venus. So we’re not talking about impacts alone, but about impact ‘chains.’ The implications of this multi-impact theory on planet composition may be profound.

Alexandre Emsenhuber (now at Ludwig Maximilian University, Munich) is lead author of the paper on Earth/Venus interactions, pointing to the different impact scenarios for Earth and Venus:

“The prevailing idea has been that it doesn’t really matter if planets collide and don’t merge right away, because they are going to run into each other again at some point and merge then. But that is not what we find. We find they end up more frequently becoming part of Venus, instead of returning back to Earth. It’s easier to go from Earth to Venus than the other way around.”

Image: The terrestrial planets of the inner solar system, shown to scale. According to ‘late stage accretion’ theory, Mars and Mercury (front left and right) are what’s left of an original population of colliding embryos, and Venus and Earth grew in a series of giant impacts. New research focuses on the preponderance of hit-and-run collisions in giant impacts, and shows that proto-Earth would have served as a ‘vanguard’, slowing down planet-sized bodies in hit-and-runs. But it is proto-Venus, more often than not, that ultimately accretes them, meaning it was easier for Venus to acquire bodies from the outer solar system. Credit: Lsmpascal – Wikimedia Commons.

This work draws on a 2019 analysis by the same authors that first examined hit-and-run collisions and subsequent mergers of the two bodies. The authors point out that most simulations of this stage of planetary evolution assume perfect mergers for all impacts that are not completely catastrophic. Reflecting on this, they write:

Emsenhuber & Asphaug (2019a, hereafter Paper I) showed that this is not generally the case. They studied the fate of the runner following hit-and-runs into proto-Earths at 1 au, for thousands of geometries, and found that, contrary to expectation, only about half the time (depending on the runner’s egress velocity, which depends on the impact velocity and angle) do they return to collide again with proto-Earth. When they do, the return collision happens on a timescale of thousands to millions of years.

That work — fully treated in the first of the papers cited below — also revealed that the majority of the runners that did not return to the forming Earth would be likely to collide with Venus, given the assumption of their current masses and orbits. Those runners that did return would show an impact velocity in the second collision similar to the egress velocity after the first hit and run, thus slower than the original impact because of momentum loss. Follow-on collisions, then, are likely to be slow.

So we have a scenario in which the Earth takes repeated hits and spins off many impactors toward the inner system as they fall deeper into the Sun’s gravity well rather than eventually assimilating them itself. It’s an interesting notion given that, while Earth and Venus (so-called ‘sister planets’) have similar mass and density, Venus is nonetheless in a distinctly different state, its rotation retrograde compared to other planets, with a single rotation taking 243 days. There are also no moons at Venus. Do impacts during formation account for the differences?

To put the thesis to the test, the scientists built predictive models from 3D simulations of such impacts, drawing on machine learning techniques. They simulated terrestrial planet evolution over the course of 100 million years, calculating both hit-and-run collisions and those in which the impactor merged with the object struck.

The simulations explore the dynamical evolutions of remnants of hit-and-run collisions until the impactor is finally accreted or ejected.The different scenarios, says Asphaug, portray a sharply different formation history for the two worlds:

“In our view, Earth would have accreted most of its material from collisions that were head-on hits, or else slower than those experienced by Venus. Collisions into the Earth that were more oblique and higher velocity would have preferentially ended up on Venus…. We find that most giant impacts, even relatively ‘slow’ ones, are hit-and-runs. This means that for two planets to merge, you usually first have to slow them down in a hit-and-run collision. To think of giant impacts, for instance the formation of the moon, as a singular event is probably wrong. More likely it took two collisions in a row.”

Image: The Moon is thought to be the aftermath of a giant impact. According to a new theory, there were two giant impacts in a row, separated by about 1 million years, involving a Mars-sized ‘Theia’ and proto-Earth. In this image, the proposed hit-and-run collision is simulated in 3D, shown about an hour after impact. A cut-away view shows the iron cores. Theia (or most of it) barely escapes, so a follow-on collision is likely. Credit: A. Emsenhuber/University of Bern/University of Munich.

Earth’s impact history thus has a telling influence on planetary composition. From the paper:

…if the terrestrial planets formed in multiple giant impacts, then Venus is significantly more likely than Earth to have accreted a massive outer solar system body during the late stage of planet formation. Earth, by contrast, has no terrestrial planet beyond its orbit to act as a vanguard. Mars is about the same mass as the late-stage projectiles…, 0.1 M, and thus relatively inconsequential in terms of slowing them down through hit-and-run, so Earth has to do it on its own.

The late stage of terrestrial planet evolution in our own Solar System thus may hinge on how each world dealt with these impact runners. One thing that emphatically emerges from the work is that, according to these simulations, the terrestrial planets were hardly isolated during this period. Hit-and-run objects strike one planet, then the other, the probability of the impacts factored into the simulation via relative velocity and orbital configuration choices in the analysis.

In this study, Earth slows down projectiles, but accretes no more than half of them itself. Venus becomes a sink for these objects, retaining the majority of them in all simulations after their encounter with Earth as the slowed velocity of the runner allows for subsequent accretion. This would naturally lead to differences in composition between Venus and Earth and would account for differences in everything from Venus’ spin state, its formation (or lack of it) of moons, to its core-mantle dynamics. The authors promise a follow-up paper exploring these issues.

The papers are Emsenhuber et al., “Collision Chains among the Terrestrial Planets. II. An Asymmetry between Earth and Venus,” Planetary Science Journal Vol. 2, No. 5 (23 September, 2021), 199 (full text). The second paper is Asphaug et al., “Collision Chains among the Terrestrial Planets. III. Formation of the Moon,” Planetary Science Journal Vol. 2, No. 5 (23 September, 2021), 200 (full text)

tzf_img_post
{ 21 comments… add one }
  • Maurizio Morabito September 28, 2021, 10:11

    Fred Singer decades ago suggested that Venus got resurfaced 500M years ago by the impact of a temporary moon – this would explain almost everything, including the age of the crust, the retrograde and extremely slow rotation, the massive atmosphere and the enormous temperatures at the surface

    • Passing Starship September 29, 2021, 15:36

      Hmm. Such a collision 500 million years ago isn’t really in line with planet formation 4,600 million years ago.

      I also disagree it would explain any of those things Singer claimed. How big an impacter does a planet need to stop or even slightly reverse its rotation, and would the planet really survive such a big impact (bigger than Theia)? How does such a giant impact leave a heavy atmosphere, and not strip it away?

      So I looked up Fred Singer on Wikipedia. An atmospheric physicist, not a planetologist, who was a noted climate change skeptic (he did not agree with the term ‘denier’). I could not find anything about Singer making these claims about Venus, so I do not know how seriously he thought of these things.

      Except, we now know that Venus’s high surface temperature is due to the greenhouse effect of the heavy carbon dioxide atmosphere. The greenhouse effect means the surface itself is heated by the atmosphere, not the other way around.

      Carl Sagan found so, but Singer seemed to disagree with Sagan on a number of things. When you say “decades ago,” could you mean before this greenhouse effect cause was realised, or could it be after but Singer didn’t want to admit to a greenhouse effect on Venus, or subsequently anthropogenic climate change on Earth, and that explains a lot about his ideas?

  • Alex Tolley September 28, 2021, 11:07

    It seems to me there are a lot of questions about this.
    Does Jupiter act as a “vanguard” planwt for the inner system or different?
    Apart from Venus’ rotation, what prediction is there for Venus’ composition?
    Where is moonless Mercury in this model?
    Earth is the only terrestrial planet with a substantial Moon. Mars ( 2 small rocks), Venus and Mercury none.

    If only Velikovsky had heard about this – “Worlds in Collision” might have been written differently!

    • Paul Gilster September 28, 2021, 15:00

      I had the same thought re Velikovsky, Alex!

    • Ron S. September 28, 2021, 17:16

      Velikovsky? I doubt it. He didn’t let science and facts get in the way of his earnest and ill-considered fictionalization of reality in order to justify a literal reading of the old testament.

  • Michael Fidler September 29, 2021, 0:12

    800 million years ago***

    Space: Massive meteor shower hit Earth 800 million years ago.

    https://www.bbc.co.uk/newsround/53496593

    “800 million years old.”
    “Later, material believed to be Copernicus ejecta was sampled by Apollo 12 astronauts, and these samples were radiometrically dated to be close to 800 million years old. 800 million years old is certainly an ancient age by terrestrial standards but is a relatively young age for the Moon.28 Sep 2010.”

    We are like children playing in a giant crater…

    • Michael Fidler September 29, 2021, 9:05

      Moon May Have Fragments of Ancient Venus on Its Surface.

      “Many scientists believe that Venus might have had an Earth-like atmosphere as recently as 700 million years ago.”

      http://www.sci-news.com/space/moon-fragments-ancient-venus-08929.html

      Lunar Exploration as a Probe of Ancient Venus.

      https://arxiv.org/abs/2010.02215

      Rodinia
      Breakup
      In 2009 UNESCO’s IGCP project 440, named ‘Rodinia Assembly and Breakup’, concluded that Rodinia broke up in four stages between 825 and 550 Ma:[21]

      The breakup was initiated by a superplume around 825–800 Ma whose influence—such as crustal arching, intense bimodal magmatism, and accumulation of thick rift-type sedimentary successions—have been recorded in South Australia, South China, Tarim, Kalahari, India, and the Arabian-Nubian Craton.
      Rifting progressed in the same cratons 800–750 Ma and spread into Laurentia and perhaps Siberia. India (including Madagascar) and the Congo-Säo Francisco Craton were either detached from Rodinia during this period or simply never were part of the supercontinent.
      As the central part of Rodinia reached the Equator around 750–700 Ma, a new pulse of magmatism and rifting continued the disassembly in western Kalahari, West Australia, South China, Tarim, and most margins of Laurentia.
      650–550 Ma several events coincided: the opening of the Iapetus Ocean; the closure of the Braziliano, Adamastor, and Mozambique oceans; and the Pan-African orogeny. The result was the formation of Gondwana.

      https://en.wikipedia.org/wiki/Rodinia

      Rodinia was the supercontinent before Gondwana formed. The breakup was initiated by a superplume around 825–800 Ma. A large superplume was probably created by a large impacter. ;-}

      The formation of Gondwana began c. 800 to 650 Ma with the East African Orogeny, the collision of India and Madagascar with East Africa, and was completed c. 600 to 530 Ma with the overlapping Brasiliano and Kuunga orogenies, the collision of South America with Africa, and the addition of Australia and Antarctica, respectively.

      Take a look at Gondwana, all the land masses where pushed together.

      https://upload.wikimedia.org/wikipedia/commons/c/cb/Gondwana_420_Ma.png

      • Passing Starship September 29, 2021, 16:10

        Okay, I’m not sure of the Rodinia relevance, but as this is here, I just wanted to pop a thought on plate tectonics out into the world …

        Why does Earth keep forming supercontinents every 400 million years or so? Continents drift across Earth’s surface at about 5 cm per year. The Earth’s circumference is almost exactly 40,000 km. If a supercontinent breaks up, then the fragments will meet up at the antipodes about 400 million years later.

        I recall a Scientific American article (I think) that argued a different pattern: the fragments travel something like half-way to the antipodes, but then reverse to reform the supercontinent. From my understanding of the history of Earth’s plate tectonics, I think a mixture of both patterns has occured.

        • Alex Tolley September 30, 2021, 8:25

          Do you recall why they would reverse direction? That implies that the ridges separating the fragments close down and new ridges open up elsewhere to reverse the direction, requiring subduction zones at the fragment edges to consume the ocean plate between fragments.

          • Passing Starship October 2, 2021, 9:55

            I actually don’t, and embarrassingly, I can’t find any link or reference to the Scientific American article (about 25 years ago, I think).

            However, I did find this: “Wilson Cycle” (https://en.wikipedia.org/wiki/Wilson_cycle). It gives the Atlantic Ocean as an example of an opening and closing oceanic basin, though since the Atlantic hasn’t started closing yet, I thought the Iapetus Ocean would have been a better example. Note that the Atlantic Ocean has indeed formed next to old sutures though: Appalachian, Hebridian, and Norwegian mountain ranges.

            Also, the article doesn’t actually explain why continent separation reverses, but I recall that a supercontinent could not avoid separation because it insulates the mantle below, which heats up and developes a collection of hot spots underneath it. These cause the rifting that splits the supercontinent. I think that after the continents are well separated, the mantle that was under the prior supercontinent has cooled enough to start to sink, and that drags the continents back in.

            Later, it says this:

            “A Wilson cycle is not the same as a supercontinent cycle, which is the break-up of one supercontinent and the development of another and takes place on a global scale. The Wilson cycle rarely synchronizes with the timing of a supercontinent cycle. However, both supercontinent cycles and Wilson cycles were involved in the formation of Pangaea and of Rodinia.”

            I think the Scientific American article was trying to propose the Wilson Cycle as a global mechanism. It had a graphic that showed the stages: 1) a circular supercontinent split three ways into 120° segments with Red Sea style rifts; 2) well sperated continents, subduction zones on the outside and spreading centres on the interior; 3) subduction zones forming on the interior coasts; and 4) continents rejoining to form the same supercontinent, though with one of the continents rotated by 120° so that a different coastline formed the orogenic margin.

            That article made me wonder whether one needed this pattern to exclusively explain periodic supercontinent formation, as they would meet up at the anipodes anyway.

    • Ashley R Pollard October 2, 2021, 5:58

      Wow! That’s a great image, and it sparked an idea for an SF story about a war billions of years ago.

  • Antonio September 29, 2021, 5:30

    OT: The Icarus Interstellar project is definitely dead now? Their website has been down for at least a week now. It would be very sad that the project will be closed after so many years without presenting a report with their conclusions and improvements over Daedalus. I know that their members are more interested in lightsails and other topics now, but it’s a pity that they don’t present some result at the end.

  • Mike September 29, 2021, 10:51

    Under these formation conditions shouldn’t Venus have become larger then Earth? More material is being directed inward to Venus then is being accreted by Earth.

    • Passing Starship September 29, 2021, 15:51

      Yes, I thought this was an implication of the theory too.

      I think the answer that spares the theory (I think it’s a plausible theory) is this. Proto-Earth (and Theia) were acquiring more mass more quickly through oligarchic accretion than proto-Venus. So, maybe proto-Venus was, say, half the mass of proto-Earth. But then, proto-Venus got more of the quasi-Theias thrown its way and caught up.

      It’s also possible that Jupiter was acting as a vanguard planet as Alex Tolley wonders, and so proto-Earth got first dibs on proto-Venus. Even Jupiter perturbing masses of minor planetesimals inwards would favour proto-Earth.

    • Christian G September 29, 2021, 17:09

      I was wondering the same thing. Does the paper say anything about what percentage of the objects mass would be absorbed by Earth, on average, before the remainder goes on to Venus?

  • Geoffrey Hillend September 29, 2021, 20:41

    From what I recall reading that the glancing blow is what caused Earth to survive the impact. The result was the entire mantle being blasted off into space and the iron core of Theia going to into the Earth. What ever was left did not escape Earth’s orbit and became the Moon. The broken fragments of the Moon recollected into the Moon, but could not escape the gravitational field of the Earth,. Consequently, there can be no continuation of the giant impactor onto Venus. The collision gave Earth the angular momentum for it’s fast rotation.

    Venus has no fast rotation because it did not have a large impact. Maybe many small ones. It also has no magnetic field because it never had a giant impact to give it angular momentum.

    • Alex Tolley September 30, 2021, 8:26

      What accounts for Mars’ fast rotation?

  • Geoffrey Hillend September 30, 2021, 15:33

    There is some evidence Mars had a giant impact, the Borealis basin, the giant northern basin which is even larger than the Hellas basin. https://www.nasa.gov/mission_pages/MRO/news/mro-20080625.html

    Mars is also nearer to the asteroid belt. It has a thin atmosphere and has had a lot of small and medium impacts during the early and late bombardment period.

  • Dave Moore September 30, 2021, 15:43

    This could explain Mercury. Mercury has a grazing collision with Venus leaving behind its outer crust, but before hitting Venus again, it has an orbital perturbation that puts it inside the orbit of Venus. This orbit is slowly turned into its present one via solar tidal interactions which lock it into it’s current 3:2 resonance orbit.

  • Passing Starship October 2, 2021, 10:42

    Is anyone aware of the rough constancy of specific angular momentum per unit mass regarding the rotation periods of the planets in our Solar System?

    I saw this in a very large astronomical tome when I could still access University libraries freely in the 90s. Compute the angular momentum of each planet’s rotation, divide it by the planet’s mass and plot the result against planetary mass. (A log-log scale will be needed.) Mars, Jupiter, Saturn, Uranus and Neptune all sit on a straight line. Earth, Moon, Mercury and Venus don’t. Earth and Moon together do! I think Ceres fitted too.

    I guess it is a natural outcome of accretion physics.

    What I found interesting was that one could ‘deduce’ what Earth’s natural rotation rate should have been, about 12-16 hours. Venus similar.

    So, my accumulated understanding of what happened to the inner planets is this:
    – Mercury. A direct hit giant impact stripped off much if its mantle, leaving it with an oversized core.
    – Venus. Wasn’t there a theory that Venus’s slow retrograde rotation was due to atmospehric solar tides? I thought it features here on Centauri Dreams, but as I can’t find it, another false memory? Sigh. Anyway, no need for a giant impact there (not that such can’t have occured).
    – Earth: Started with a 12-16 hour rotation rate, a giant impact may have sped this up, or not, but the Moon has taken angular momentum away from Earth’s rotation, which is why Earth no long sits on that line anymore.
    – Mars. Rotates at its natural rate. There’s no need to invoke a giant impact for this planet’s rotation either. The ‘Borealis’ basin, isn’t that the whole of the Martian lowlands? Or the Utopia basin? That’s a real, circular impact basin, but is way too small for a giant impact.

    I think a giant impact cannot leave an impact basin, as the basin would have to cover more than the whole surface, so a giant impact strips the surface off instead (and giant impacts might be defined this way).

Leave a Comment