Coal is not a "manageable, local risk"; coal causes problems with probability ~1.
Also, the average coal power plant produces more radioactive waste than a nuclear power plant.
The real problem: people fail statistics. People can easily observe a nuclear plant when it shows up on the news, and assign a disproportionately high risk to it. The ongoing damage of coal plants doesn't make the news, because it falls under "day-to-day operation" rather than "disaster".
Nuclear plant failure statistics are meaningless as a measure of safety. You are pointing to the nonoccurrence of a rare event as evidence of its rarity.
Even if coal plants cause more damage to the world, failures are local and manageable. When your nuclear plant fails catastrophically, as they have at Chernobyl and now Fukushima, you can't point to "but this is a six-sigma event!" as an excuse. Given this risk, fission plants have been neither economically nor environmentally preferable enough to displace other ways of generating electricity.
So far nuclear power plants cause a big problem with probability ~1 -- we can't reclaim the sites they're built on and we can't safely dispose of the waste. And these problems are going to last longer than any system of government we've ever created has lasted.
Also, the average coal power plant produces more radioactive waste than a nuclear power plant.
The real problem: people fail statistics. People can easily observe a nuclear plant when it shows up on the news, and assign a disproportionately high risk to it. The ongoing damage of coal plants doesn't make the news, because it falls under "day-to-day operation" rather than "disaster".