It was taken seriously enough that a secret theoretical treatment was made to prove it couldn't happen. Here is how the Manhattan District History, the secret internal history of the bomb project, discusses it (I quote it at length, because it is rather interesting):

It should be mentioned at this point that in the early period of the project the most careful attention was given to the possibility that a thermonuclear reaction might be initiated in light elements of the Earth's atmosphere or crust. The easiest reaction to ignite, if any, was found to be a reaction between nitrogen nuclei in the atmosphere. It was assumed only the most energetic of several possible reactions would occur, and that the reaction cross-sections were at the maximum values theoretically possible. Calculation led to the result that no matter how high the temperature, energy loss would exceed energy production by a reasonable factor. At an assumed temperature of three million electron volts the reaction failed to be self-propagating by a factor of sixty. This temperature exceeded the calculated initial temperature of the deuterium reaction by a factor of one hundred, and that of the fission bomb by a larger factor.

The impossibility of igniting the atmosphere was thus assured by science and common sense. The essential factors in these calculations, the Coulomb forces of the nucleus, are among the best understood phenomena of modern physics. The philosophic possibility of destroying the earth, associated with the thermonuclear reaction, which is the only method now known by which such a catastrophe might occur, is evidently ruled out. The general stability of matter in the observable universe argues against it. Further knowledge of the nature of the great stellar explosions, novae and supernovae, will throw light on these questions. In the almost complete absence of real knowledge, it is generally believed that the tremendous energy of these explosions is of gravitational rather than nuclear origin.

More immediate and less spectacular global dangers to humanity arise from the use of thermonuclear bombs, or even fission bombs, in war: principally from the possible magnitude of destruction and from radioactive poisoning of the atmosphere.

This was written probably in late 1946 or early 1947 (it was "published" in April 1947 but it is not clear how long it was in development). It was not intended for public consumption, and was written before thermonuclear weapons had been developed (but after, obviously, they had been theorized).

So in some sense, the scientists knew that it couldn't happen. But of course this was an argument based on theory. And in your gut, theory doesn't always win. And so no less a rational person than James Conant, the chemist and President of Harvard University, recorded his reaction to the Trinity test in July 1945:

Then came a burst of white light that seemed to fill the sky and seemed to last for seconds. I had expected a relatively quick and bright flash. The enormity of the light and its length quite stunned me. My instantaneous reaction was that something had gone wrong and that the thermal nuclear transformational of the atmosphere, once discussed as a possibility and jokingly referred to a few minutes earlier, had actually occurred.

Which is just a nice indication that there was still some gut appreciation of the fact that they were treading into unknown with this kind of terrestrial energy release.