With better tooling this kind of status quo will change. Choosing a single worst case for design is a product of manual design processes. With modern computer technology we can run multiple cases easily, so we can now use probabilistic engineering techniques.

In the future it won’t be “this bridge is good for 1 in 100 year storm” but “this bridge has a 99.99% chance of surviving every storm during its design life.” That will include the very rare events beyond the 100 year horizon.

Mitigation ≠ total prevention

It's really not that hard to see this isn't a binary choice.