This discussion had not progressed very far when I inadvertently side-tracked it by introducing the topic of inverse flames.
After thinking about the original problem of whether a collapsing portion of ceiling under a pyrolyzate-rich, hot attic could cause a backdraft explosion, I have come to believe that it may well be reasonably possible.
The pressure of the cool gas at ceiling level below the attic will be higher than the pressure of the hot gas at floor level in the attic because of the effect of temperature on density.
If we assume that the upper attic structure is “leaky,” (as is typically the case) and that the room below has some access to outside air via leaks or an open door or window, the cold air will be driven at least slowly into the attic and progress to some point short of the neutral plane in the attic. Under these condition, the cooler air will tend spill across the attic floor the same way it would if it came in from the usual vertical sources (gravity currents). The rate of this flow will be limited by the rate at which hot air can leak through the upper elevations of the attic or the rate at which fresh air can enter the room below, depending on which process is slower.
If the amount of oxygen in the attic increases sufficiently before the mixture finds an ignition source such as a small flame in a remote area of the attic or a glowing ember which erupts into flame, all the conditions necessary for an for an explosion will be satisfied.
Of course, I may be dead wrong. If so, I’m sure I can rely on the readers to straighten me out.