Flying the meteoric skies

By Phil Plait | June 5, 2009 7:00 am
Asteroid and airplane

Over at Cosmic Variance, John Conway ponders the idea that a meteor might have taken down the Air France flight that crashed over the Atlantic the other day.

John’s a physicist, so he goes through the math. You have to make some assumptions, but most of them look solid to me, or at least not crazy. In the end, the odds of any one flight getting hit by a rock substantial enough to do catastrophic damage to a plane is extremely low. That should be obvious, because if the odds were high then we’d see it happen a lot!

But over time, the odds of one flight getting hit are just high enough to make a single disaster in the past 20 years just within the realm of possibility.

Still, I wonder. When it comes to meteoroids (the solid body that forms the meteor), smaller rocks are more common than larger ones. So for every rock that hits a plane that’s large enough to take it down, we should see lots of damage from smaller ones. Yet I have never heard of such damage being reported (and I think it would be huge news if it did, but to be honest I have not done the research).

But I continue to wonder. There is another complication I think people have forgotten. Airplanes fly at an altitude of roughly 10 km. Small meteoroids move at transsonic speeds when they enter our atmosphere, but slow rapidly, and become subsonic in seconds while they’re still about 100 km up. Once they’ve slowed they are no longer ballistic, and really just fall the rest of the way to the ground. By the time they’re at the same altitude as planes, they may only be moving a few hundred km/hr at most.

That changes the physics a lot. Why? Because of relative areas and speeds.

To calculate the odds of a plane getting hit, you need to know how many meteors burn up over the Earth per day, and what percentage of the Earth’s surface is covered by airplanes. If you have millions of airplanes in the air, the odds of them getting hit are pretty high, but if you only have one, it’s a pretty small target.

But that assumes the plane occupies a couple of hundred square meters of area; that is, that’s the area of an airplane as seen from above.

Imagine you are a rock moving at 25,000 km/hr. To you, the plane is essentially standing still, and the amount of area you see it occupy is really just its area as seen from above. In that case, the assumption is correct.

But now let’s imagine you have zero motion, and so you’re motionless, hovering in the sky. In that case, the plane hits you! You don’t see the plane from above at all, you see it coming head-on, and so it occupies less area, only a few dozen square meters.

That latter case is the one that is more physically realistic, because most meteoroids will be falling relatively slowly when they are in the air lanes. That means that the statistical odds of a plane getting hit by a meteoroid (or really the other way around) are far smaller, maybe by a factor of ten or more.

I suspect that may be why we have never definitively seen an airplane taken down by a meteor. Once you realize the rocks are moving slowly, planes are a lot harder to hit. And that might explain why we don’t see much damage even from smaller rocks. It really is a rare event.

And I should add that it’s incredibly unlikely in any case that this was the fate of Air France 447. We know it was flying into heavy thunderstorms, and in this case when you hear hoofbeats you think horses, not zebras. When you have such an obvious culprit, reaching for an incredibly rare one isn’t terribly parsimonious.

Still, as John points out, it’s really just a matter of time before a plane is hit. But because of relative speeds and cross sections, I suspect that time period is longer than most people think.

Airplane image from Marina Avila’s Flickr stream.

CATEGORIZED UNDER: Astronomy, DeathfromtheSkies!, Science

Comments (62)

Links to this Post

  1. | June 5, 2009
  1. blur

    i guess it would be hard to diagnose a meteor-assisted plane crash too… maybe there were such cases, but it’s hard to come up with this explanation, when you have a pile of plane pieces, scattered across the ocean.. black box ain’t telling you that it was a meteorite..

  2. BigBadSis

    Dr. P: You said “By the time they’re at the same altitude as planets, they may only be moving a few hundred km/hr at most.” I think you meant to spell “planes” not “planets.” Just sayin’…

  3. William

    No BigBadSis, ‘planets’ is correct. He is of course referring to Mars, which as we all know will be well within the Moon’s orbit any time now. :p

  4. As Phil points out, a meteoroid that reaches altitudes of 10-12km has to be much bigger than most space debris that encounters Earth atmosphere. I would assume that a lot of energy (heat) would be released slowing down such an object from thousands of km/hr to the terminal velocity (a few km/hr). Surely, an IR (weather) satellite would be able to notice such blast through the atmosphere?!

  5. Also *I* think you’re misusing the term ballistic whe you say that when meteors are slowed they are no longer ballistic. A object that is simply free-falling is still following a ballistic trajectory. Indeed my copy of Websters uses the term free-falling to define ballistic.

  6. I saw a very cool meteor as I was flying from Australia to Hawaii back in 1999. It was incredibly cool looking, and we think it actually went all the way down to the water. Since it was night, with the speed, light, and lack of references, it sure APPEARED a lot closer than it probably was. We did joke about what would have happened if it had hit us.

    But it’s a big sky, and even a KC-135 is small on the scale of things like planets.

  7. Craig Edgar

    Hi Phil,

    apologies for posting this is here but I wasn;t sure how to email you, I found an article where 2 parents have been found guilty of manslaughter for using homeopathic remedies on thier child instead of convential medicine. thought this may be of interst to you, Here is the link:



  8. It’s also likely that any damage from extra-terrestrial sources is presumed to be something more mundane. Fairly large bits of ice, the size of a baseball and potentially even larger, can be caught in the thermal patterns of clouds. Birds hit planes all the time.

    If no one saw the impact any damage detected later would be assumed to be the product of some more likely agent.

  9. Charles Boyer

    It’s a long journey from “might” to “did.” While the mathematics is an interesting exercise, I would be extremely surprised if the post-crash investigators looked seriously into it.

  10. NoAstronomer: I don’t think an object falling through an atmosphere is free-falling. It is being influenced by friction with the air. So I don’t think its trajectory can be described as ballistic anymore.

  11. Phil: You are correct to say that most meteors don’t make it to 10 km in the atmosphere, however, your argument about the cross section of the airplane is incorrect. If you assume that all planes fly at an altitude of 10 km, then they cover some fraction of the surface area of the sphere at that radius. If the meteor goes through this part of the atmosphere, then the probability of it hitting a plane is simply the fraction of the area that the planes cover. The fact that the plane is moving at some point, actually _increases_ the probability of getting hit, because now you have to include the surface area of the front of the plane as well.. of course this is a small fraction of the surface area and so it is not important.

  12. Once they’ve slowed they are no longer ballistic

    Wait, they’re no longer covered by the laws of classical mechanics as they apply to flying or falling objects? ‘Cause that’s what “ballistic” means.

    In fact, the most common use of the term to refer to a “ballistic” missile is one that’s only powered for a short takeoff, and afterwards is only guided by gravity and classical mechanics, as opposed to a missile that has thrust all the way to the target like a rocket. I think you picked exactly the opposite term from what you meant.

  13. Basically, once in dark flight, the meteoroid indeed slows down until an equilibrium between fall velocity and atmospheric drag is reached. For a fist-sized object, that will be about 200 m/s (that is metre, not mile, per second). It is falling more or less vertical by that time.

    It’s seems to become standard that any ‘mysterious’ airplane downing is now leading to a meteorite strike suggestion. We saw it with flight TWA 800 earlier. Sign that people have become aware of meteorite impact. Sign also that they still prefer the exotic over the mundane.

    Question is whether a hit by an average size meteorite will down a plane. Does a hole being punched in a wing down a plane? Depends on exactly where and how big I guess. It wouldn’t be much different from an airplane being hit by anti-aircraft machinegun fire. Military aircraft can survive quite a beating of those, but for civil aircraft, I don’t know. I guess as long as no vital structural parts are damaged, a hit need not be seriously fatal per se.

    Objects that have been documented to been hit by a meteorite already, include;

    – several houses;
    – several cars;
    – a ship (in harbour);
    – a mailbox;
    – a cow;
    – at least two humans (Syllacauga and Mbale; both survived).

  14. Ryan S.

    Hmm. OK, I could be completely off with this assessment, given that I’m a biologist and not an astronomer or physicist, but I would not think that the odds of the plane being hit by a meteor are reduced by a factor of ten, perhaps more like reduced by 1/2. Here’s my reasoning:

    Assuming a Boeing 747, 70m in length, 10m frontal height, travelling at its usual 913km/hr (253m/s). Also assuming a meteor with a terminal velocity of 100m/s (to make things easy).

    The slope of the “danger zone” where this falling meteor could be within during a snapshot in time and still hit the plane should therefore be 0.395-ish (100m/253m). A little trig tells me that the cross section of this “danger zone” (ie, how wide the danger zone is or the relative probabilty of a collision) is about 35 meters for this Boeing 747 plane (sin68.4*(70m/2.35+10m)). In the hypothetical scenario you posted of an extremely fast meteor relative to the plane, the danger zone has a width of close to 70 meters, or the length of the plane. So the danger zone for the terminal velocity meteor ends up being just over half the width (and therefore half the probability) of the very fast meteor.

    I’m sure there are probably other factors I’m not considering in that simple assessment, however.

  15. Pete


    As that Hawaiian Airlines flight that lost a section of the cabin shows, civilian aircraft can take a pretty good beating.

  16. ellindsey

    A meteor impact would seem to be ruled out by the ACARS messages sent back by the automated diagnostic systems on the airplane before it disappeared. AF447 sent back multiple diagnostic warnings over a period of several minutes. Among the first warnings were messages that the readings from the sensors for airspeed, altitude, angle of attack and such were disagreeing with each other. These were followed by autopilot disconnect and reversion to alternate flight control mode, then multiple electrical faults and cabin pressurization warning. This is not consistent with a meteor impact, which would have caused a loss of pressure first followed by other problems. Theories right now are of either pitot tubes icing up, or some kind of complex electronic fault, leaving the pilots with marginal data about the state of the aircraft and no autopilot, at night in a severe storm. Under those conditions it would be very easy to put the plane into an unrecoverable stall or spin, with the plane eventually breaking up from aerodynamic forces.

  17. Monkey

    Have you folks heard about Mars? It is going to be quite close to us soon, and apparently it will be as big as……

    heh, had to.

  18. Here is a not so quick estimate of the effect of the velocities of the plane and velocity of the meteor. If the plane isn’t moving and the meteor is falling straight down, we said the probability of being hit is then length*width of the plane / 4 *Pi * (R_e)^2 where R_e is the radius above the center of the earth it is flying at (i.e. the surface area of the plane divided by the surface area of a sphere). If the objects are moving, the question then is, what is the probability that the meteor intersects the volume of the plane as it moves in the sky. Suppose the plane is moving horizontally at a velocity v_p and the meteor is falling at a velocity v_fall, and the plane has a height h. Then the time the meteor can spend in the volume where the plane is fllying is just t_fall = h / v_fall. The volume the plane takes up during this time is (l*w*h+w*h*v_p*t_fall) that is the volume of the plane (width w x height hx lenght l), plus the volume swept out by the front surface of the plane (w*h*v_p*t_fall). The probability the plane is hit is just this volume divided by the total volume of the shell (4 * Pi * r_e^2 * h).

    This gives the actual probability of a meteor hitting a plane (in the shape of a box):

    = (l*w+w*h*v_p/v_f)/(4*Pi*r_e^2).

    So the calculation before was the minimum probability. If the plane is moving very quickly and the meteor is falling very slowly (i.e. v_p/v_f >> l/h) then it is more likely the front will be hit. But it will _increase_ the probability of being hit. This phenomenon is often called aberration (at least in the context of relativity).

  19. Sir Eccles

    This sounds very similar to that argument as to whether you get wetter by running home in the rain or just walking.

  20. Re: Sir Eccles

    It is the same argument, however there is one difference. In that case you have to include the distance traveled instead of v_p*h/v_fall because there is a continous amount of rain falling, not just one drop. This leads one to the conclusion it best to run, and sideways at that, to stay the driest.

  21. Guysmiley

    Doh, my post was eated up by the site.

    Quick summary:

    The aircraft was in a known severe weather area.

    Airbuses have already crashed due to structural failure. (AA Flight 587, the FO snapped the entire vertical fin off an A300 by getting too aggressive with rudder inputs while encountering turbulence)

    The odds of an aircraft getting hit by a meteor is almost unimaginably slim.

    Occam’s Razor saaaays: It wasn’t a meteor unless some extraordinary evidence is discovered leading to that conclusion.

  22. Come on People!!

    It was a Gremlin……you know…Diabolicle Sabatoogie

  23. I’ve never heard of a plane being shot down by a zebra before.

  24. Matija

    I completely agree with Ryan. If the plane is moving appreciably during the time it takes the meteorite to cover its height, it will be a larger target than if it was more-or-less standing still. There is no way it can be a smaller target, as it will still have the same area seen from above, plus the chance of hitting the meteorite head-on.

    The change in probability is modest, though. If the meteorite is falling at 100 m/s and the plane is 10m tall, the window for a head-on collision would be only about 0.1 seconds. If the plane is flying at at 250 m/s (about 900 km/h) than it would cover only 25m over that time, which is shorter than most jetliners. So the increase in probability is only about 50%, which is a lot for engineers but negligible for astronomers.

  25. Phil, here’s another reason to question the calculation: no human has ever been killed by a meteor strike on record. In fact, I think that only one case of a person getting hit is known, and that was a ricochet. Human bodies certainly cover a larger fraction of the planet than planes. Even if you restrict yourself to people out of doors, I’d wager that this is true. In fact, the surface area is almost certainly a factor of 10 or more larger. Now, certainly, probably about half of meteors that make it to airplane cruising altitudes don’t make to the ground. (The half is based on the amount of air between us and space and the amount between us and that altitude.) Even with this, you would expect many more people to have been hit than planes. Given the lack of reports of people getting whacked, I’m skeptical of odds even as high as John Conway’s analysis suggest.
    (This isn’t to say he’s clearly wrong, but I *am* skeptical.)

  26. Perhaps an engine aspirated a radiosonde or a migrating bird. Meteor strike statistics are slightly boosted by orbital debris re-entry.

  27. I think there’s only one way to resolve this. Have the Mythbusters test it! Who wouldn’t want to see the Mythbusters remote-rig an airplane and try to smash it at a high altitude with falling rocks? 😉

  28. I first heard this type of thing being brought up as part of the TWA flight 800 investigation.

    Unlikely, but not impossible.

  29. Tim G

    We had at least a few hours notice for the meteoroid named 2008 TC3 before it entered over northern Sudan on October 6th of last year. I wonder if sometime in the near future when we get another advanced warning we will send a military jet or drone into the trail left by the object. Then, the plane may have a relatively high chance of getting struck by space debris.

  30. Jeff

    Odds are based on how many times an event will occur assuming several trials.

    For example, if it’s 1/1,000,000 odds a plane will be hit, then after 1,000,000 flights, one will be hit.

    And for the object hit, it is 100% . So maybe this was the one that scored the odds.

  31. “We know it was flying into heavy thunderstorms, and in this case when you hear hoofbeats you think horses, not zebras. “

    Pretty much exactly what I was going to say. If the plane went down under clear skies, I’d be willing to consider a meteoroid impact. But right now I’m leaning towards counting the great big frakkin’ storm it flew into as at least a contributing factor.

  32. Since the flight was sending out automatic messages indicating electrical problems prior to the catastrophe, it seems unlikely a meteor would have brought it down.

  33. Joe

    There’s maybe a bigger chance it was downed by one of these:

  34. Harry D

    A few things I would like to point out. Often times a meteor sighting or to be more precise a bolide sighting is sometimes precluded by unusually intense rain and or hail. This is caused in part by the ice melting off the meteor and coming down as precipitation. This could have been a massive storm caused caused by multiple objects. I believe that the flash of light seen by the pilots on the Air Comet flight (how ironic is it that the name of the airline is Air Comet?) could have been another meteor descending in the area and not necessarily the one that brought Air France down. The other thing that everyone seems to have overlooked when calculating the odds is the fact that the bolide does not have to actually strike the airplane in order to bring it down, it just has to get close enough so that the electromagnetic disturbance caused by it will affect its’ electronic instruments. It seems that the Aerbus is more susceptible to this than older planes. I am looking into reports of meteor sightings coinciding with the same model plane loosing several hundred feet of altitude over Australia in October of 2008. Coincidentally the first ever asteroid tracked from space that had pieces of it recovered on the ground occurred on the same week. Qantas Flight 72 had problems two months later over Australia and that coincided with an uptick in meteoric activity as well. That’s my two cents.

  35. Steve h

    @ michael l.

    Well, comets have been known to effect the electrical systems of 18 wheelers and gas pumps, so its not unfeasible that a meteor could cause havoc to the electrical systems.

  36. Monkey Says:
    Have you folks heard about Mars?……

    heh, had to.

    No, not ‘had to’, but “couldn’t resist”.



  37. Jon Jonjentry

    If you’re a geek, its probably fun to do the calculations here … but a meteor hitting a plane? … has to be the most ridiculous thing I’ve heard in a while.

    While the speculation is interesting, lets be respectful … a lot of people died in this tragic event!

  38. Dan E

    Your argument is that since the meteor is moving slower, it would likely strike the front of the aircraft instead of the top, thus decreasing the surface area as well as the likelihood of impact. However, the slower the meteor is traveling, the longer it is in the airspace of the aircraft, and that in turn increases the likelihood of an impact. It would be interesting to see the math, but the slower speed may more than compensate for the reduced surface area of the aircraft exposed to the meteor, possibly increasing the likelihood of an impact overall.

  39. Jeff, that’s not how probability works. If the odds of something happening are one in a million, after a million trials you have a 63% chance that it will have happened. (I’m a bit ashamed I had to run the numbers on that before I realized why that particular value appeared…)

  40. Tim G

    Sorry, Jeff. John Weiss is right. The probability would be almost exactly 1-e^-1.

  41. Ah, another math-guy. :-)

    In other news, I just flipped open my copy of Hartmann and went digging for meteor flux information. A 1-m (in space) body hits the Earth of average once an hour. According to Hartmann, when get below 10 m, you’re starting to see things break up in the atmosphere. So 1-m radii bodies are reasonably in the ballpark of the minimum sized bodies you’d need to wipe out a plane. If you accept that, this throws the calculations off and reduces the odds of any plane being taken down in the past twenty years to 0.03% (check my math?).

  42. Steve in MI

    Somewhere in this story, there’s a Slashdot-style “In Soviet Russia” joke waiting to happen.

  43. Richard

    John, (or Tim) can you explain to us Liberal Arts types the explanation for the 63% chance of occurrence after a million trials on one in a million odds?

  44. The Other Ian

    I have to agree with the folks arguing that a slow fall will increase the chance of hitting, not decrease it. If the rock is falling fast enough that we can view the aircraft as a stationary object, then we can draw an overhead silhouette of the plane as the “target” the rock needs to hit in order to collide with the plane. For a moving plane, this target gets stretched out to cover all the area the plane covers over the period of time when the rock is at the right altitude to be struck. In effect, the target gets bigger.

    Taking this to extremes, imagine a plane moving so fast that it traverses the entire surface area of the globe in one second, and a meteoroid falling so slowly that it takes one full second to pass through the plane’s altitude. In this scenario, the chance of collision is 100%.

  45. Bryan Feir


    Given the usual explanation of the Tunguska event as a meteoric air burst, I can see why you’d say that.

    And no, I’m not going to succumb to the temptation myself.

  46. Tim G

    Richard (#45) This may not be the best explanation but here goes:

    Imagine playing Russian roulette once. There is a (5/6) chance (83.33%) of getting spared. If you spin the cylinder and play it a second time, there is also a (5/6) chance of surviving and therefore a (5/6) of (5/6) chance of surviving an attempt at two games. Attempt to play the game three times and your chance of survival is (5/6) x (5/6) x (5/6). Try to play the game six times (you must spin the cylinder each time to keep the probability of an individual hit at (1/6)) and your chance of survival is (5/6)^6, or about 33.5%. So your chance of getting shot is 66.5% (not 100%). As long as you spin the cylinder of bullet chambers each time, your luck doesn’t necessarily have to be “used up” after six attempts.

    Now imagine having a 1000 by 1000 pixel screen (one million pixels) where every pixel is white. Now every second you randomly target a pixel to “hit” to turn it to black (regardless of whether it is already black). At first, the number of pixels that are white will decrease more or less linearly because every second it is almost 100% likely that a new pixel will be “hit” and turned to black. It seems that when we make our one millionth “hit” all one million pixels will be turned black. However, after a little while enough black pixels accumulate so that you begin to “hit” more and more pixels that are already black. The rate of decay of remaining white pixels slows down. So instead of hitting all one million pixels, some are spared and some are hit more than once. After one million seconds, the fraction that are spared is 0.999999 to the power of one million, or 36.79% (very close to e to the power of negative one). So the fraction that is hit is 63.21%.

  47. Exactly. If the odds of something happening on one trial are p, the odds of it NEVER happening in N trials is (1-p)^N. (1-p is just the odds of the event not happening.) So the odds of it happening at least once in N trials is 1 – (1-p)^N.

    The interesting thing about (1-p)^N, where p = 1/N, is that you can define the mathematical constant e by e = limit(N-> infinity) (1+1/N)^N.

    (More generally, e^x is defined by limit(N->inf) (1+x/n)^N.)

    So for this case, N=1,000,000 we’ve pretty much limited on the final value, e^-1, and the value we came up with should have been expected.

    — John, who still thinks of himself as a Liberals Arts Type, too.

  48. Gavin Flower

    If a meteor hits the atmosphere and breaks up, then you have something like a shotgun effect.

    So consider the possibility of a meteor of about 0.1 metre diameter breaking up. This would vastly increase the chance of multiple strikes on the same aircraft from one meteor.

    Also remember that meteors normally come in swarms, rarely just as individual rocks – assuming this is from a comet.

    Do if one meteor is in the way of an aircraft, the chances are high that others will also be.

    So statistically speaking, the chance of an aircraft being hit by one or more meteors are much lower than indicated by a naive assumption that the probability of being hit is a ratio of areas times the number of meteors per unit time… However, when an incident DOES OCCUR then the odds of being hit multiple times are very high.

  49. Gavin Flower

    “Do if one meteor is in the way of an aircraft, the chances are high that others will also be.”

    should be…

    “So if one meteor is in the way of an aircraft, the chances are high that others will also be.”

    Clicking on my original comment, came up with the error “load failed” or some such! Hence, I added a new comment.

  50. eddie

    Bermuda Triangle moved south for the summer.

    Mystery solved.

    Next question?

  51. LKM

    I think your reasoning with regards to “falling on the plane” is wrong. It doesn’t matter how fast the plane or the meteoroid are (as long as the meteoroid is actually falling and not standing still), the probability of the meteoroid falling on the plane should remain equal as long as there is a plane.

    The probability of the plane running into the meteoroid increases as the meteoroid becomes slower, as it will be in the area where collision is possible for a longer time (and a faster plane will cover more horizontal space, increasing the probability of hitting a meteoroid). Likewise, the plane will be exposed longer if it is slower.

    However, the space the plane takes up seen from above remains the same, no matter how fast the plane is and how fast the meteoroid is. In other words, the probability of hitting a plane that is going at a higher speed is now lower; it will simply be a different plane that will be hit by a different meteoroid if, say, all planes go twice as fast (but remain in the sky for the same amount of time). Time is what really matters; if a plane is in the air 10 hours, it’ll be hit from above twice as likely as one that is in the air 5 hours. How fast the planes go doesn’t matter.

    All things considered, a slower comet will be more likely to touch a plane than a faster comet. Hitting from above has the same probability, but hitting horizontally increases in probability.

  52. Charles

    Umm… Not to be picky, but isn’t it hypersonic speed, not transonic? I thought transonic was around Mach 1 (~350m/s), and entry speed in the order of 8+ km/s?

  53. Aviation Buff

    I’m no expert, but a long-time amateur student of aviation safety.

    Commercial jets have a lot of damage tolerance. There are many cases of airliners making safe or at least partly controlled landings with all sorts of damage, a few of the causes being:

    * skin or door failure leading to cabin pressure explosion
    * gross upset leading to airspeeds and/or G loadings far in excess of design limits
    * attack by military weapons
    * bird strike
    * flight through volcanic ash
    * engine explosion
    * collision with ground structures during takeoff

    Based on this history, I would expect that an airliner would have a good chance of surviving collision with a small meteor. My understanding being that the frequency of meteors increases dramatically with diminishing size, I would expect many survivable meteor strikes for each catastrophic outcome.

    It is further my opinion that if a jet had a hole or other damage from meteor impact, analysis would very likely reveal the material, so meteor damage would have a good chance of being identified as such.

    If these inferences are correct, the absence of reports of meteor damage during, say, the past 50 years suggest that the probability of a catastrophic incident is a lot less than once in 50 years.

  54. Torbjörn Larsson, OM

    Did you include possible fragmentation?

    And I noted that you all go ballistic on terms.

  55. Gary Ansorge

    I have to wonder, Phil, if a plane traveling at 800 km/hr hits a ten 5 kg chunk of rock traveling at, say 350 km/hr, would that not be likely to initiate a crash, as in, “OMG! What was that hitting our fuel tank?”. followed by a big “whoosh” and flames,,,

    GAry 7

  56. Gavin Flower


    I wonder if it could have been a meteor strike, followed by hitting severe turbulence and electrical storms – each individually survivable, but not in combination? From the BBC article (, the aircraft had at least one problem that it survived that caused loss of radio contact, and it did not crash until many minutes later.

    As someone else pointed out, several aircraft could have already have been hit by meteors without people ralizing it was meteors causing the damage.

  57. coolstar

    Maybe Phil should hire John Weiss as his consultant for back of the envelope calculations. It could certainly lower his rate of stupid mistakes. The comparison of the surface area of in-flight planes to that of people on the planet is particularly clever.
    Since it’s obvious that the speed of the plane doesn’t change the calculation by enough to worry about, another way to approach the problem is asking how many airplanes ON THE GROUND have ever been hit by meteorites. A decent starting guess is that a typical airliner may spend about 0.25 of it’s lifetime actually in the air. Since we’re doing back of the envelope calculations, being off by a fact of a few here doesn’t matter much. So, if you can’t find any reports of airplanes on the ground being hit, there’s a very good chance that none have been hit (much less DOWNED) while flying.

  58. coolstar

    Back of the envelope calcs are like potatoe chips, once you start, it’s hard to stop: A reasonable guess is that the area of planes on the ground is at least 10 times that of those in the air (including planes out of service). Another reasonable guess is that the kinetic energy of a bird strike is roughly the same as that of a meteorite hit to a plane at cruising altitude (yeah, the meteorite maybe has an average speed (delta V, of course) twice as high or so, but the average bird may mass roughly 4x that of a small meteorite etc. etc.). The FAA, naturally, keeps quite good stats on bird strikes: the chance of any given flight having one is roughly 1 in 10,000. In 2009 we know of exactly one airliner that was downed by a bird strike (multiple engines are a GOOD thing) so we can set that as a reasonable guess as the fraction of meteor strikes that will cause a plane crash. So the chances of an airliner being downed by a meteor strike is roughly 1/100000 of that of a plane on the ground being hit. Yeah, I cheated and looked up the FAA stats, but I was going to use something between 1000 and 10,000 anyway………

  59. The amount of damage a commercial airliner can sustain and survive is both amazingly high, and amazingly low. It just depends where and what kind of damage is sustained.

    The Boeing 737 that landed in Hawaii with most of its roof missing is an extreme example of the former. But note that, while the damage there looked (and was) severe, most of the planes critical systems (and a lot of its structure) is in the lower half of the fuselage.

    On the other hand, a few metal shavings in rudder actuators have apparently been enough to bring down multiple 737s over the years (obviously this isn’t an example of externally caused damage, but it does demonstrate how a small failure at a critical point can doom an aircraft).

    In general, modern aircraft are very robust, and have enough redundancy of critical systems to survive all sort of damage. I suppose you could come up with some sort of “critical damage target area” for any given aircraft, and using this would bring down the likelihood of a meteor-impact loss even farther.

    But even that doesn’t take into account the KIND of impact. Speed, area, mass, and the mechanical characteristics of the rock would have to be taken into consideration. A pebble sized rock traveling at a few hundred miles per hour isn’t likely to do much damage unless it hits a particularlly vulnerable area. It might damage a window in a cockpit. It might even cause an engine failure, though such failures are rarely fatal these days.

    I suppose it might also damage a fuel tank, but that’s unlikely to be as catastrophic a problem as an earlier poster imagines. Burning jet fuel requires an ignition source and a proper fuel-air mixture, and the flame would have to burn close to the structure of the aircraft before being swept away in the slipstream. Most likely, it would simply result in a fuel leak with out ignition, with all the problems that result from that. Assuming the fuel loss didn’t result in an emergency landing, I’d imagine the pilot would notice the fuel loss and probably run (or vent) the tank dry before landing as a precaution.

    A large, slow moving boulder, on the other hand, could do considerable damage no matter where it hit. (So would a dropped piano, for that matter, if Mars were to toss one down as it passes by.)

    But when you think about it, it comes back to the same issues as the Hawaii 737. Unless the meteor is very large, or very fast, it is rather less likely to do critical damage, as the most vulnerable systems are mostly on the bottom (structure, landing gear, hydraulics, electrical system) or on the rear (control surfaces). If you punch a hole in the roof, you might depressurize the cabin

  60. (continued from previous, prematurely posted comment)
    If you punch a hole in the roof, you might depressurize the cabin (and even this takes a surprisingly large hole) and injure or kill one or more passengers. But the plane would likely survive.

    If you hit a leading edge on a wing it is (unlike the space shuttle) unlikely to do critical damage.

    Striking the top of the wing would likely result in a fuel leak, but the critical hydraulic lines and structural wing spar offer much smaller and better protected targets.

    The front of the cockpit might be an especially vulnerable area (assuming the strike takes out both pilots, or happens when only one is in his/her seat), but again, that’s probably just the windows and the area right around them.

    The tail surfaces are probably the most vulnerable, though if the assumption is correct that a strike would be head-on, even those should be pretty robust, and again, small targets.

    What strikes me as interesting is the idea that aircraft may have already struck smaller meteors and nobody noticed. As an earlier poster mentioned, anything that didn’t penetrate the skin would likely be mistaken for a bird strike, hail, or a foreign object kicked up on a runway. The people reporting and investigating these things likely would have no idea what to look for to spot such an impact.

    But I’m pretty sure that such incidents and repairs are maticuliously documented and recorded. and it might be possible for an industrious researcher with better knowledge of meteors to dig through them an spot previously missed impacts (if any).

  61. mars

    To those that think a meteor and/or meteor shower could not have brought down AF 0447, here’s something to think about.
    On June 1 the asteroid KR21 was supposed to be in it’s closest proximity to the earth. Within a 0.7 LD, that’s barely 167,000 miles from the earth, according to website . Could debris from this rather large asteroid, 21 meters in size, filter through the earth’s atmosphere and have a direct strike on the aircraft? This may sound like a highly improbable scenario in regards to aircraft, then again so are all the other theories such as a lightning strike or hail storm bringing down a modern aircraft equipped with all the latest in flight technology. If there were means of tracking meteors and/or meteor showers or space junk debris via orbiting satellite or radar, I wonder if such data could reveal a different story to the demise of AF 0447.

    P.s. Here are some links to the and NASA Jet Propulsion Laboratory websites archives for the June 1 meteoric events. Also, links to articles on space debris.;orb=1;cov=0;log=0;cad=1#cad


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!


See More

Collapse bottom bar