But those events were temporal and local. Nuclear failures are catastrophic and never ending events in many generations.
The only reason why nuclear sticks so easily is because of the magic word "radiation". It's easier to be scared of it than sit down land learn that it is all a natural physical phenomenon, even if it originated from a man-made isotope. It's difficulty arises from the stochastic (quantum) nature of it's interactions.
Lastly, reactor designs have significantly improved since 50 years ago. New reactor designs I saw coming out of Westinghouse could lose all power and pumps and still use natural convection and reservoirs (elevated pools) within the containment to prevent any sort of critical event leading to meltdown. Imagine if we were stuck with the coal technology of the 70's. I would prefer modernizing the nuclear fleet if possible, which does include decommissioning old reactors, and closing the fuel cycle loop in a way that is proliferation resistant (some sort of pyroprocessing) unlike UREX.
Isn't simply using a non-PUREX reprocessing method sufficient? Realistically, a "nuclear club" nation like the US/France/GB only needs to ensure that the reprocessed fuel contains enough non-Pu239 isotopes that any attempt at a bomb with stolen Pu would necessarily fizzle. The fact that the nation itself could (theoretically) produce Pu239 via the process and cause "proliferation" seems far fetched in the absence of the Cold War level rivalry that was the original proliferation impetus. Warheads are expensive. Nobody who already has a bunch of them already is really interested in making more.