In the early days of the Manhattan project, there was a brief scare, based on some calculations by Edward Teller, that an atomic bomb would ignite the atmosphere. They later redid the calculations and concluded that the probability was very small (something like one in three million).

What if we lived in the universe where that improbable event happened, and the Trinity test ignited the atmosphere in the way that Teller predicted?

  • Could somebody in America get word out to the rest of the world, just before being incinerated?
  • Would the wavefront of burning atmosphere move slowly enough that people in, say, India had enough time to pull together a few people, and put them somewhere where they might have a shot at survival (at least for more than a few days)? And what place could they use?
  • Is there a reasonable scenario where they could survive for say a few months or a year, and use that time to figure out a way survive in the now absolutely barren Earth (would the atmosphere be gone, or just changed into something entirely different).
  • If they could manage cling on to their miserable existence and generate offspring, is there some way they could restore the atmosphere, over many generations?
  • If not, and humanity is doomed, would life survive on earth?

Since the poor souls in this universe were very unlucky, we can allow them some luck in the little time they have to save themselves: being in the right time at the right place, and so on. The odds are obviously against humanity, so they might catch a little break here and there, but we assume only one spectacular coincidence.

EDIT: I'm going to slightly move the goalposts here (with apologies to the authors of the existing answers), because we're getting into discussions about the nuclear physics, which I had hoped to avoid.

So let's add the following: the test causes a very, very slightly self sustaining nuclear reaction. Technically, this would increase exponentially, but the rate is so slow that the result looks more like the world being on fire than the world blowing up. Can some quick-thinking American get word to some quick thinking Indian, ahead of the wavefront of the fire?

Is there somewhere people with advanced warning can seal themselves off with a supply of oxygen and some plants? There's no need to make the point that survival is very unlikely, but given a little luck, is there a way that we could just pull through, or at least cling on for a few more weeks.

EDIT2: To be even more blunt: this question is not about physics, it is about logistics. The Trinity test just gives us a convenient point in time (ie. level of technology), and a basic mechanism to work off to set the parameters. The question is, if the atmosphere were on fire, would there be some way of surviving. How would we get the message out of America? How much time would people in Europe and Asia have? What structures existing in 1945 would allow one to survive until the atmosphere had burnt up. If you're in India, and you get word that the world will be in flames in an hour, what actions should you take?

  • 1
    $\begingroup$ Related: worldbuilding.stackexchange.com/questions/47424/… The actual answer seems to be it just wouldn't happen, but they at least had a mechanism in mind for how it might happen if it would happen, so we don't need to quarrel about low probability vs. physically impossible. $\endgroup$ Commented Jan 17, 2017 at 22:09
  • 1
    $\begingroup$ I appreciate that we need to somehow get a bound on the properties of a chain reaction that basically can't happen. If it helps, imagine that in this world, there is some undiscovered physics principle and the trinity test is basically like lighting a match in a gas-filled room, so that the effect would be that of a gas combusting, relatively slowly (compared to a nuclear explosion). It is however, a chain reaction, and any atmosphere humanity would like to preserve needs to be sealed off. $\endgroup$ Commented Jan 18, 2017 at 11:32
  • $\begingroup$ I like the answer you selected @BenRW so just a comment - "If not, and humanity is doomed, would life survive on earth?" most of the nature of the answers and questions deal with pure physics. I am not sure I saw anything take the Oceans of the Earth into consideration. My stance would be yes life in the Oceans and large lakes would survive, things would reboot and evolution is reset (On land). I recon that if there had been any "HUmans" with enough DAN variation and resources, living under water, then a reboot of Humans could occur. $\endgroup$ Commented Apr 12, 2017 at 13:48
  • $\begingroup$ @EnigmaMaitreya Would the oceans survive? How much of the oceans would boil away from the heat of the atmosphere burning, and how much would boil away once the firestorm has burned itself out and there is no atmosphere left (or would there be an atmosphere left)? $\endgroup$ Commented Apr 13, 2017 at 12:52
  • $\begingroup$ Well there is Water as steam or if you want a heat exchange going. So you have an active heat exchange going on vs the the heat required to sustain the "fire". The thermal exchange will undoubtedly cause ... air movement in all 3 axis. Water contains the elements for our atmosphere. Probably what is required is to know the volume of Water on Earth vs the Volume of participating components of the Atmosphere. I think that in the end the question might be Can the event sustain itself long enough to 100% vaporize the Earth's water. $\endgroup$ Commented Apr 13, 2017 at 13:05

4 Answers 4


If it had ignited, it would have been a nuclear explosion that was, at the absolute minimum, self-sustaining.

If it was even slightly more than self-sustaining, it would be exponentially growing (in both the rhetorical sense of "fast" and the mathematical sense of repeatedly doubling in size in constant time) until it ran out of fuel.

In either case, there would be no survival of any life at all. In the bare minimum case — where it's only just self sustaining — it would be more like a permanent fire than an explosion, but even that would eventually convert all the nitrogen in the air into oxygen, leaving none left for plant life.

In the more extreme case, and given how quickly nuclear chain reactions tend to double in size, all the oceans would first boil, then ionise, then all the hydrogen (or at the very least the trace levels of deuterium) would fuse, and there wouldn't be any water left anywhere on the planet.

If it was sufficiently rapid, and if one of Stephen Hawking's comments in A Brief History of Time is correct, then the planet might even collapse into a black hole from the force of the explosion:

Such conditions could occur in a very big hydrogen bomb: the physicist John Wheeler once calculated that if one took all the heavy water in all the oceans of the world, one could build a hydrogen bomb that would compress matter at the center so much that a black hole would be created. (Of course, there would be no one left to observe it!)

For extra fun, the discovery that even poorly-enriched Lithium enhanced fusion bombs was well after the Trinity test (Castle Bravo and Castle Romeo were both more powerful than calculations beforehand predicted). They didn't know it, but if there had been a self-sustaining chain reaction in the atmosphere as extreme as feared, the trace levels of Lithium on their own would have had an effect similar to "the entire crust of the Earth is now made out of an explosive several times more powerful than the same mass of dynamite". That's, what, 5–70 kilometres deep?

Edit because question was edited:

Because it's a counterfactual scenario anyway, you can make the slow-burn scenario be as fast or as slow as you want it to be. You could even have people survive in a sealed greenhouse (Biosphere 2 style) built in the local state capital of Albuquerque, after the detonation, by making the nuclear "fire" "burn" so slowly the recirculation of nitrogen fuel is more relevant to the dying world than the advance of the flame.

Then you can have people stockpiling all the nitrogen-rich solids and liquids (e.g. ammonia-based fertilisers), and then, once the fire has burned out its own fuel, the protagonists can slowly start re-building the world that's now only slightly more habitable than Mars, one artificial biosphere at a time.

Or, if this is just an excuse for an apocalypse, you can make it exactly fast enough that only one remote science base in Kerala that just happened to be a Biosphere 2 clone, has time to seal itself off from the world.

Edit 2 because of second question edit:


  • Getting word out would be easy (radio)
  • Survival with purely things that had been invented in 1945 (rather than things that could be built if people had the imagination) was not possible. Long-term survival after any event that could be described as "the atmosphere ignited" would need either a Biosphere 2 (completed September 1991) or a nuclear submarine (first one launched for sea trials in January 1955)
  • $\begingroup$ Seawater cannot sustain a fusion chain reaction. While the total amount of heavy water in the oceans is significant, it is can only fuse if concentrated. See my answer here for the gory math details. In short, the energies needed to sustain a fusion reaction in seawater are higher than the output energy of that seawater, so no chain reaction is possible. $\endgroup$
    – kingledion
    Commented Jan 18, 2017 at 1:57
  • $\begingroup$ If the chain reaction is possible in the first place this is the only place I can see it going. That and the sun dying tens of billions of years early and being a lot hotter. $\endgroup$
    – Joe Bloggs
    Commented Jan 18, 2017 at 9:00
  • $\begingroup$ So, in the bare minimum case, how fast would the wave front propagate? Could someone get word to India, tell people to lock themselves in deep mineshafts? $\endgroup$ Commented Jan 18, 2017 at 10:01
  • 1
    $\begingroup$ @kingledion If we were talking about reality, then agreed, in reality it can't — but in the hypothetical and counterfactual world where an atomic bomb was enough to trigger the runaway fusion scenario that Teller warned about, seawater was a necessary component. So in the scenario the question was about, the H₁ and H₂ in the seawater would fuse. $\endgroup$
    – BenRW
    Commented Jan 18, 2017 at 12:41
  • 1
    $\begingroup$ @BenRW thanks for the addition. However, the biosphere would surely perish quickly when enveloped in flames. Moreover, I don't expect any such projects were running in 1945. Sorry if that seems pedantic, but this is really the sort of thing I wanted to get at with this question: the logistics of survival rather than the specifics of how the atmosphere gets burned. $\endgroup$ Commented Jan 18, 2017 at 22:26

"Ignite the atmosphere" worries concerned a chemical reaction not a nuclear one. The reaction of Nitrogen with Oxygen is exothermic, so in theory could be self-sustaining. However it has a very high activation energy, so it only happens at high temperatures (like in the cylinders of a car engine to a small extent). If the heat escapes faster than the reaction generates more heat, then this sort of fire dies out rather than propagating itself.

We've had some pretty large forest fires, and several non-nuclear explosions in the kilotonne range. No atmospheric ignition. There is also geological evidence of high-megatonne meteor impacts. Again no atmospheric ignition catastrophe. So they tested the bombs. It was about as risky (on that particular issue), as the more recent worries about the LHC. If it could happen it would already have happened and we would not be here.


  • 1
    $\begingroup$ Can you link to that? Other answers cited N-N nuclear reactions. $\endgroup$
    – JDługosz
    Commented Jan 18, 2017 at 22:34

The paper in the linked post by @Peter is relevant. It shows that nitrogen-nitrogen reactions cannot happen mostly by virtue of high Coulomb barrier: a 8.6 MeV thermal energy to overcome the barrier is equivalent to an absurd 100 billion K. Its just hard to get and keep particles at that temperature. Even stars undergoing silicon fusion the day before they supernova are only doing about 3 billion Kelvin.

At such high temperatures, temperature loss by any mechanism will be so extreme and so fast that there is a statistically zero chance for a chain reaction to occur.

Given this bit of nuclear science, your question doesn't make much sense. In order for a chain reaction to occur, one of three things must be true.

  • I calculate that Little Boy (15 kT TNT, about the energy at Trinity) has enough energy to raise only 1 kg of water to 100 billion kelvin. So somehow you must contain the entire energy of an atomic bomb into 1 kg of water. That isn't really possible.

  • The laws of physics are different, an the Coulomb barrier for N-N fusion, or the cross section for N-N fusion are different. This will certainly be more important in stellar physics than it will be in nuclear bombs, with unknown effects on the rest of the Universe. In which case, knowing if the world would survive an N-N chain reaction is less interesting than knowing if stars and planets could even form at all.

  • There is no chain reaction.

Either way, there is no real way to answer your question. Just suffice to say that said chain reaction could not happen.

  • 1
    $\begingroup$ This doesn't seem to answer the question. The question is "if it did happen". $\endgroup$
    – Tim B
    Commented Jan 18, 2017 at 11:12
  • 1
    $\begingroup$ @TimB This is a 'bad assumptions' answer. If 'it did happen' stellar nucleosynthesis works differently in your world also, so then who knows what would happen. In conclusion, it can't happen in an 'alternate-history' as goes the tag. $\endgroup$
    – kingledion
    Commented Jan 18, 2017 at 12:48

There is a way this could happen: if the Earth had a small amount of hydrogen >4% in the upper atmosphere then the top of a sufficiently large nuclear explosion (eg Tsar Bomba 57MT) would have ignited it.

The effects would be very bad, global firestorm rivaling that from the K-T event, heating of the surface by over 100C in a few seconds, etc. The good news is that the flash boiling of the oceans would help a bit here as the cooling would be quite rapid and thus some simple life might survive. The loss of oxygen here would be about 10-20% for a minimal event which would be severe in the extreme and suffocate many lifeforms in a matter of hours.

In this case the alternate history would be that during the Earth's initial formation complex metal-hydrogen compounds got trapped in deep geological formations and were later heated up and the H2 vented eg during a series of massive underwater volcanic eruptions. In fact we can see methane (CH4) clathrates so this isn't totally implausible as the volume needed could be quite low in real terms. Multiplication effects ie nitrogen and oxygen combining into nitrites would also add to the damage but ironically this is just a scaled up version of what happens during lightning strikes anyway.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .