I'm imagining something like this: 1) frozen products of decomposition, including methane, are kept homogenous by temperatures permanently below the freezing point of water 2) Global warming causes a freeze-thaw cycle that briefly melts the permafrost. Methane, no longer trapped by frozen water, rises - but not all the way to the surface because half of a freeze/thaw cycle is freezing. 3) Over time, the freeze-thaw cycle "kneads" methane out of homogeneity and into pockets - just as many small bubbles collapse into few large ones, the "foamyness" of methane in permafrost collapses into cavitation 4) For select large, dramatic regions of deep permafrost and aggressive climactic change, those large pockets eventually reach a critical mass that becomes self-sustaining, ejecting large quantities of gas all at once. See also: Lake Nyos aka "The Lake Nyos Disaster"
Sounds reasonable; basically a methane boil as all these pockets start coalescing into ever bigger pockets in a runaway reaction? Still, since it's always above freezing under the permafrost (according to Wikipedia, a max thickness of 4m, and that crater looks far deeper), perhaps these pockets are common, and there are many places where the permafrost serves as a cap over columns of foamy earth. Maybe the freeze-thaw cycle isn't so important in the formation of these foamy regions, but more so in the degradation of the caps over them?
This is where my knowledge of permafrost methane encapsulation wanes but my interest in permafrost methane encapsulation takes me no further. I remember Nyos as a kid and there was a lot of handwaving as to why, but since then I've read two scientific articles on it that both hedge as far as the reason. This is probably one of those "well, we weren't expecting that but it might be because of this" issues that masters' theses are made of.