I don't think that such scenarios would fall under any sort of long-tail thinking, since the chance of having N waves converge at a point decays exponentially i

I don't think that such scenarios would fall under any sort of long-tail thinking, since the chance of having N waves converge at a point decays exponentially in terms of N, whereas the height/energy/damages/etc. increase polynomially at most. This means we can place finite (and not very large) 50th, 90th, 99th, etc. percentiles on the heights of combined waves, assuming we can accurately predict the height distribution of a single wave. As a related thought experiment, suppose I jump down from a 5-foot ledge and land on the ground, sending a wave of pressure through it. In principle, I could be walking down the street, when a similar wave of pressure comes up through the ground and launches me 5 feet in the air. Yet the vibrations in the ground are normally so dispersed that such a ground wave will never occur in anyone's lifetime (outside of an earthquake, of course).