Meteorite from Planet Earth, or Clever Fake from a Moroccan Suq?

A presentation on 2023-07-11 at the European Association of Geochemistry and Geochemical Society Goldschmidt2023 conference in Lyon, France, “Northwest Africa 13188: a possible meteorite … from Earth!”, argues the object is a genuine rock ejected from the Earth which then returned to fall as a meteorite around 10,000 years later. Here is the abstract:

NWA 13188 was classified as an ungrouped achondrite [1]. It is a vesicular igneous rock with overall basaltic andesite composition (Mg# 58.5) and subophitic texture. It is dominated by plagioclase (49 vol%) and pyroxene (26 vol%), a fine-grained mesostatis and accessory FeTi oxides. Its oxygen isotopic composition is δ18O=8.03±0.08‰, δ17O=4.16±0.12‰ and Δ17O=-0.02±0.03‰ (n=2). The CI-normalized REE pattern display an enrichment in incompatible trace elements, with (La/Sm)N=2 and (La/Lu)N=3.5, and a depletion in Nb-Ta. The μ ^{142}{\rm Nd} is −0.59 ± 3.3. These characteristics are compatible with terrestrial calc-alkaline arc volcanism, raising doubts that this rock is a meteorite.

However, the presence of a well-developed fusion crust (see images) strongly suggests that NWA 13188 is indeed a meteorite. Moreover, the concentrations of cosmogenic ^{10}{\rm Be}, ^3{\rm He} and ^{21}{\rm Ne} point to a very short (~10 kyr) but significant exposure to galactic cosmic rays, and preclude that NWA 13188 is a man-made “fake” meteorite.

Therefore, we consider NWA 13188 to be a meteorite, launched form the Earth and later re-accreted to its surface. This scenario matches the latest definition of meteorites: “Material launched from a celestial body that achieves an independent orbit around the Sun or some other celestial body, and which eventually is re-accreted by the original body, should be considered a meteorite. The difficulty, of course, would be in proving that this had happened, but a terrestrial rock that had been exposed to cosmic rays and had a well-developed fusion crust should be considered a possible terrestrial meteorite” [2]. The launch process (impact or direct ejection during a volcanic eruption) remains to be determined. Finally, we will further constrain the formation processes of NWA 13188 by measuring its crystallization age using the ^{40}{\rm Ar}/^{39}{\rm Ar} technique. Importantly, this approach will enable us to test if it contains trapped atmospheric argon, which should be particularly abundant for a young terrestrial eruption. We will also measure the ^{38}{\rm Ar}_{\rm c} cosmogenic exposure age of the rock.

References: [1] Gattacceca J. et al. 2021. The Meteoritical Bulletin, No. 109. M&PS, doi:10.1111/maps.13714. [2] Rubin A.E. and Grossman J.N. 2010. Meteorite and meteoroid: new comprehensive definitions. M&PS 45:114-122.



OK. You mathematicians here show me why this is not possible.

I am not that versed on all the mathematical mechanics of orbital flight. However, I have to wonder just how this small piece managed to achieve orbital velocity. Even if thrown from an active volcano, I would be surprised to find it could go into orbit, much less a star-centered orbit. And that’s what this seems to imply.

Once we pass that hurdle, then please explain to me how it came back to earth. The obvious implication is that it was thrown into an earth-centric orbit, so it circled the sun at the same distance as earth. ?What math would predict this, considering the large mass discrepancy. ?Why would it “return” to earth. If it were not going at the same velocity, it ought not to be in the earth orbit. If it is going the same velocity, ?how did we “catch” it.

You math geniuses here please explain in terms a dummy like me can understand. (Not being snarky - just not a math genius.)


There are two possible ways such an object might have been launched from Earth into a heliocentric orbit. One is being propelled by an explosive volcanic eruption. Volcanic eruptions release plenty of energy—far more than needed to accelerate an object to escape velocity—Mt St Helens was estimated at around 24 megatons, more than any hydrogen bomb tested by the U.S., but I am dubious of the mechanism that could concentrate enough of that energy in a rock to cause it to escape. The other mechanism, which I believe is much more probable, is secondary ejecta from an asteroid impact. We have found meteors on Earth which are known to have originated on the Moon and Mars (isotope ratios provide indisputable evidence of this), and it’s almost certain they were ejected by impacts on those bodies. It’s harder to escape the Earth due to its stronger gravity and thick atmosphere, but then asteroids have energy in abundance, coupled with hitting faster than escape velocity.

Check out the “manhole cover to the stars” from the Operation Plumbbob Pascal-B nuclear test on 1957-08-27. If it made it through the atmosphere without burning up (probably not), it was estimated as travelling six times Earth’s escape velocity and may have escaped the Sun’s gravity onto an interstellar trajectory.

Any object orbiting the Sun in an orbit similar to the Earth’s has a high probability of eventually hitting the Earth or Moon. Every time it passes close to either, their gravity perturb its orbit and eventually it will be captured into an unstable orbit in the Earth-Moon system and either escape back to a heliocentric orbit or hit the Earth or Moon. A famous case of this is the third stage of the Saturn V rocket that launched Apollo 12 toward the Moon. See “The Strange Voyage of Apollo 12’s S-IVB Stage”, posted here on 2023-01-26. Launched from Earth in November 1969, it eventually went into orbit around the Sun where it remained until September 2002, when an amateur astronomer discovered it orbiting the Earth. It had been perturbed into an unstable orbit while passing the Earth-Moon system. It escaped again in June 2003 and is now back in orbit around the Sun. It may next be captured sometime in the 2040s. It is almost certain to eventually hit the Earth or Moon, but it may take millions of years before this happens.


Assuming that the rock in question is not a Moroccan fake, isn’t the most likely explanation that the rock never left the Earth’s gravitational field?

Maybe a volcano or an asteroid strike launched it into a sub-orbital trajectory. Travelling fast as it re-entered the atmosphere would explain the fusion crust. It would be interesting if there were enough isotopic peculiarities to identify where/when the originally Earth-bound rock started its journey.


The conference presentation abstract says (second paragraph) they found isotopic signatures that indicate the rock was exposed to galactic cosmic rays for around 10,000 years. This would require that it have been in space for that period of time. This would almost certainly require a heliocentric orbit, because a highly elliptical Earth orbit (which it would have due to lack of ability to circularise) would not be stable over such a long time.


Remember the old geological saying – If you want three plausible alternative explanations of some observation, you only have to ask two geologists. :smiley:

Other rock scientists would undoubtedly have plausible alternative explanations for the particular observations that don’t involve 10,000 years exposure to extra-atmospheric cosmic rays, and might allow for a more plausible history of the the rock’s motions.

But (a) this is all above my pay grade anyway, and (b) my confidence in scientific publications has been shattered by the ClimateScam and the CovidScam. It may be unfair to some genuine scientists, but I no longer have automatic confidence in assertions by scientists, even if they are “peer-reviewed”.


Thank you for the explanations.

It seems to me, ignorant as I am, that (a) you do not have a reasonable explanation for how it got off the earth, and (b) I still don’t understand how it lasted 10,000 years in earth’s orbit yet crashed back to earth at the end of that time period.

?Is this not a small particle (ie. a fairly small piece of rock). Even at boulder-size, its energy to hold it at the same orbit as earth is not terribly likely, and especially for some 10,000 years. My limited knowledge of such things make both mass and velocity important to how it gets into orbit and how it stays in orbit, especially the same orbit as earth. I can understand your comment about it falling back to earth, seeing as earth is a peculiarly large mass circulating the sun at a particular distance and speed. ?How do you get those tiny parameters manifested by the rock to match the earth without either it being “hovered” promptly back to earth OR escaping earth altogether - in which case, how did it get back.

?Don’t mass and velocity have important parts in determining orbit. I understand how we put a satellite into “orbit” - but it’s around us - earth - a pretty significant gravitational force to begin with. Earth stays in orbit around the sun only because of the relative differences in mass between earth and sun. ?How does this small chunk of “earth” fit into that. process.


It’s up to each individual to decide which explanations are reasonable to themselves. Most astronomers and geologists who have studied asteroid impacts have concluded that transfer of impact ejecta among the Earth, Moon, and Mars, while rare, occurs as a results of large impacts. We have discovered 277 Martian meteorites and a large number of lunar meteorites (around 1 in 1000 meteorites found) on Earth. Since we have “ground truth” for the isotopic composition of rocks from Mars and the Moon from robot explorers and, for the Moon, returned samples, and these are distinctly different from Earth rocks (and each other), these source identifications are considered highly reliable.

So, we know rocks ejected by impacts on the Moon and Mars make it to Earth. Why shouldn’t we expect rocks to also travel in the other direction? The only reasons that might prevent this are the Earth’s higher gravity and thicker atmosphere. But these simply mean a larger impact is required for rocks to make it through the atmosphere and reach Earth escape velocity (11.186 km/sec). It is estimated that an impactor larger than 10 km should be sufficient to eject rocks at this velocity, and smaller impactors may do the job, depending upon the location of the impact and condition of the atmosphere immediately thereafter.

To put the energy requirement in perspective, accelerating a 1 kilogram rock (around twice the mass of Northwest Africa 13188) to Earth escape velocity requires imparting around 62.6 megajoules of energy to it. This is around the explosive energy of 13.5 kilograms of TNT. A one km impactor striking the Earth at 17 km/sec and an incidence angle of 45° releases energy equivalent to 46.6 billion tons of TNT, so the only question is how efficiently the energy is coupled to the ejecta.

Orbital mechanics seems counter-intuitive, but this only because we have grown up at the bottom of a deep gravity well and thick atmosphere whose friction slows moving objects. In fact, orbital mechanics is entirely based upon Newton’s laws of motion and gravitation, with a little help from kinetic energy as formulated by Leibniz and Johann Bernoulli, all involving no mathematics more complicated than squaring numbers.

When a light object orbits an object much more massive (for example, the Earth and Sun, or an artificial satellite orbiting the Earth), the mass of the light object makes no significant difference in the orbit. Thus, a 1 kilogram rock and the Earth (around 6\times 10^{24} kg) will, at the same distance from the Sun and orbital velocity, follow essentially identical orbits. An object launched or ejected from the Earth at greater than escape velocity will enter an orbit around the Sun (heliocentric) with one apsis around the same as the Earth’s average distance from the Sun (one astronomical unit [AU]) and the other closer or farther from the Sun depending upon whether the object left the Earth in a direction opposed to or in line with Earth’s orbital velocity. For example, if a rocket was launched from Earth with a velocity 1 km/sec greater than escape velocity, aimed along the Earth’s direction of travel in its orbit, it will end up in a heliocentric orbit with a periapsis (perihelion) around the same as Earth’s (1 AU) and an apoapsis (aphelion) between the orbits of Earth and Mars. Had it been launched in the opposite direction, it would orbit with aphelion close to 1 AU and perihelion between the orbits of Venus and Earth.

And, if left alone, that’s where it would continue to orbit forever. But it isn’t left alone. Because its orbital period will be different than that of the Earth, although its orbit takes it near the Earth’s orbit once per circuit of the Sun, usually when it gets there the Earth will be distant, elsewhere in its orbit. But eventually it’s going to arrive at that point when the Earth (and/or Moon) are nearby, and then, depending upon the geometry of the encounter, its orbit will be perturbed by them. Its orbit may grow or shrink; it may enter temporary orbit around the Earth and Moon, or it may hit one or the other. Give it enough time, and there will often be an impact. None of this is pure theory. We’ve seen this behaviour in spacecraft launched into heliocentric orbits from Earth (such as Apollo 12’s third stage, as noted above), as well as in perturbed orbits of Earth-crossing asteroids passing near the Earth.

There is thus no mystery how a rock ejected from the Earth by an asteroid impact found its way back to Earth. Apollo 12’s S-IVB made it back into Earth orbit only 33 years after it left. That time it didn’t hit anything, but give it time, and it almost certainly will eventually.

As late as the 1950s, behaviour of asteroids in near-Earth orbits was largely a matter of speculation, but now, with the discovery of more than 30,000 such objects and regular tracking of their positions and encounters with the Earth and Moon (and in some cases other planets), we know how they behave over time. In addition, computer simulations allow tracking their orbits forward and backward for millions of years to discover how frequently they impact the inner planets.


All you wrote was plausible … until that last sentence. Computer simulations can be useful, and sometimes accurate (provided we know all the physical relationships completely and all the initial conditions exactly), but since those provisos are rarely met in practice, we generally should not overstate their utility. See Catastrophic Anthropogenic Global Warming.

As it happens, I have been banging my head against a wall trying to comprehend a mathematical biography of Henri Poincare (1854 - 1912). Interestingly, one of Poncare’s many achievements was winning a prize in 1889 of 2,500 Swedish crowns awarded by King Oscar II for what was essentially a mathematical analysis of the stability (or lack thereof) of the solar system. Apparently, that was recognized as a matter of concern in those days.

In the immensity of geological time, many strange and highly unlikely things have happened. From the scientific perspective, what matters is recognizing the limits of our understanding and keeping open minds about possible alternative explanations for curious observations.


In the case of the state vectors of solar system bodies, this is manifestly not the case. The recent discovery of long wavelength gravitational radiation required knowing the position of Jupiter with respect to the Sun and Earth to an accuracy of better than 100 metres. Planets are big and heavy, and there’s nothing much to perturb their motion other than gravity, and centuries of observations allow calculating their mass very precisely. Other than the biggest four or so, asteroids don’t have enough mass to make any difference in the calculations.

None of the feats of interplanetary spacecraft navigation since the 1970s: gravity assists, targeting for aerobraking, rendezvous with asteroids and comets, would have been possible with simulations based upon data sets such as the JPL Horizons Ephemeris.


I fear we are talking at cross-purposes here. There is no doubt that calculations of orbital mechanics in the present day are astonishingly accurate. But that is a world away from: “computer simulations allow tracking their orbits forward and backward for millions of years to discover how frequently they impact the inner planets.”

That kind of simulation brings us to the world of Chaos Theory and the Butterfly Effect. An error of 100 meters in the initial position of Jupiter could/likely would result in great differences after millions of years.

My concern is excessive reliance on simulation with mathematical models of physical systems which are necessarily much more complex (see Catastrophic Anthropogenic Global Warming as predicted by models). Simulation has been compared to a chain saw – a very powerful tool in the hands of a skilled user, but with an ever-present risk of causing serious damage to the user and to the world.

1 Like

The people who are doing these solar system dynamics simulations are acutely aware of the accuracy of the state vector information which is the input and of the resulting limits on the accuracy of their predictions. It’s all in the details. For example, the orbit of Pluto is, over a period of several million years, truly chaotic: a change of 1 cm in its initial position can be amplified to a 180° difference in orbital longitude. However, Pluto’s mass (1.32\times 10^{22} kg) is so small compared to the giant outer planets (for example, Neptune, 1.02413\times 10^{26} kg—note how many more significant digits of Neptune’s mass we know compared to Pluto’s) that its chaotic motion has a negligible effect on the other planets. The people that do these calculations are neither ignorant nor charlatans, and if they published something unwarranted by the known precision of the input data, others would be all over them pointing out the errors, especially in the era where thousands of people have access to the same source data and sufficiently powerful computers on their desktop to perform such calculations.

But that’s not what we were talking about. You seemed to imply that because some simulations cannot be trusted (both because we lack accurate input data and the models are untrustworthy and untested against the real world, even looking backward), then all simulations were invalid. Simulations, like anything else, obey the “garbage in, garbage out” rule. When the input data are valid, simulation is often the only way to study the behaviour of systems on which we cannot run experiments, such as the interactions of planets and other solar system objects or the formation of structure in the early universe.


Perhaps you should investigate what engineers call “sensitivity analysis”. Answers precisely that question with a carefully designed set of perturbations applied to multiple simulation runs.


Not at all. I agree that simulations have their place, and can be very powerful in some situations. From bitter expensive experience, I have learned that simulations should never be treated as Gods who always give the right answer. They are models – simplifications of reality, approximations – and as George Box said, All models are wrong, but some models are useful.

Humility! It is the sine qua non for engineers in the real world.


Or, put another way, simulations are only as good as the data inputted. Look at all the “climate change” models put forth as truth. None of them is worth the effort to run them because the “data set” used is untrustworthy.

So, like statistics, you can use simulations to prove pretty much anything - just control the input data. Honest simulation use, like statistics, can be useful, but one always has to be careful. Especially in wholly accepting the results as truth. Truth itself “evolves” as we learn more, so why should sims be statically correct forever.


There are two distinct ways in which a simulation can generate meaningless results, although there’s nothing to prevent a simulation from failing in both ways simultaneously.

The first is where the simulation is a faithful model of the physical system and incorporates all of the relevant effects with resolution adequate for the period of the simulation but the data defining the initial conditions from which the simulation starts are incorrect or insufficiently accurate. An example of this would be trying to simulate the break shot on even an ideal (perfectly flat, no friction, consistent bounce off cushions) billiards table. The simulation would nonetheless never simulate the real system remotely accurately because microscopic differences in the initial position of the balls would alter the myriad mutual bounces in the first moments after the cue ball struck, causing large differences in the ultimate result. This is also the case for simulating systems exhibiting deterministic chaos, where the smallest difference in initial conditions is quickly amplified into utterly different macroscopic states. A particularly simple chaotic system is the double pendulum, simulated here in Second Life by Fourmilab.

The second type of failure is where you know the initial conditions to sufficient accuracy but the simulation fails to take into account all of the relevant physical phenomena that affect the system in the real world. A classic example of this is attempts in the 19th century to calculate the motions of the planets from their observed positions and velocities and Newton’s laws of motion and gravitation. Despite the observational data for the start point being more than sufficiently accurate, the calculations always ended up with a small discrepancy in the position of Mercury. Proposed solutions included the presence of a planet closer to the Sun which was called “Vulcan” or an asteroid belt inside the orbit of Mercury that perturbed the planet, but neither was found despite numerous observing campaigns looking for them. In this case, the problem turned out to be that Newton’s laws were incorrect for very massive objects and high velocities. Mercury’s orbit combined the high mass of the Sun and its high orbital velocity, adding up to visible discrepancies. The mystery was not solved until Einstein discovered laws of motion and gravitation that handled such conditions. This “failure” of simulation actually demonstrates the value of simulations. Had they not been done, there would have been no need to search for explanations that generalised Newton’s laws to handle these extreme conditions. Simulation failure can also occur when the simulation ignores effects which turn out to be significant. When modeling physical systems, one often has to simplify the real-world case to make the simulation calculations tractable (“Assume a cow is a sphere”). If these simplifications are unwarranted, the simulation will produce meaningless results even if the input data are perfectly accurate.

Climate science is an example of simulations where both the input data are insufficiently known and the models used in the simulation ignore many processes that affect the actual flow of mass and energy around the Earth. Plus, the system being simulated may be chaotic.


Thank you for that very clear exposition of the limits on simulation, Mr. W. Let me share an observation about the use of simulation models in the real world – the output from an unreliable simulation tends to look about the same as the output from a reliable simulation.

An example was Ferguson’s UK predictions of huge numbers of CovidScam deaths. The model had complex equations on virus spread and mortality rates and detailed initial descriptions of populations & susceptibilities. The predictions came from a computer, for Goodness Sake! And yet the course of hundreds of millions (arguably, billions) of human lives were permanently altered for the worse by these impressive-looking but simply erroneous simulation predictions.

Testing the reliability of simulation predictions is a very challenging task – one which is often skipped (as in the case of the CovidScam). This is particularly difficult in a world in which under-informed decision-makers tend to place increasingly high confidence in ever more complex & inscrutable simulation models.

The practical answer is that we should maintain a healthy skepticism about the predictions from computer models – guilty until proven innocent. Undoubtedly many simulation models do prove to be reliable, but we should always demand proof – especially when model predictions are used to limit human freedoms.


This is a marvelous tutorial. Thank you! Your ability to explain things in a manner comprehensible to non-experts is a rare gift and an important facet of the value of Scanalyst. If you were a wine, you would be “approachable”.