This is only a problem if it's possible for the transformation to be self-sustaining. That is, the act of transforming atmosphere causes yet more atmosphere to transform. Only in that situation would the whole planet's atmosphere burn.

Is that possible?

Let's make an analogy with trees. You can't burn a living tree outright. Or so you'd think. But once there is enough energy around a forest fire can jump from crown to crown faster than you'd imagine possible outrunning the ground fire. This will continue until the fire runs out of fuel or the conditions change enough to put it out.

http://www.thefreedictionary.com/crown+fire

There is so much energy in the system that the burning crown of one tree dries out the matter in the crown of the next to the point where it can be combusted in an extremely short time, contrary to ordinary conditions when lighting 'wet wood' appears not to work.

The whole point was that Teller thought that the reaction might be self sustaining once the activation energy was reached.

Think about it: someone made the call to ignore that possibility and pull the trigger on deployment even if they weren't sure about the outcome. Possibly the single most irresponsible deed in the history of man.

> someone made the call to ignore that possibility

Actually, they did the calculations and verified that it wouldn't happen. It was Richard Hamming who did the calculations. https://en.wikipedia.org/wiki/Richard_Hamming#Manhattan_Proj...

I said to myself, "What have you done, Hamming, you are involved in risking all of life that is known in the Universe, and you do not know much of an essential part? [uncertain variable]" I was pacing up and down the corridor when a friend asked me what was bothering me. I told him. His reply was, "Never mind, Hamming, no one will ever blame you."