[koheripbal]

OP's analysis demonstrates that it's very possible to detect theses events 1000s of years later, if you examine areas near the explosion.

[gota]

Yes, but they are testing the area because we know a settlement there was destroyed. We'd have to test random uninhabited (or at least currently uninhabited) areas for which there are no indications of anything special to get a better picture

And not even that guarantees a complete picture - what if a tsunami event erases or conceals the record of something like this? Volcanic activity? Desertification?

My point is that we can't reason about the rarity or uniqueness of these events as a main factor for accepting or refuting the hypothesis