By contrast, you're stuck with the consequences of a 1-2 degree temperature increase. And they're guaranteed, and will be around for centuries before the carbon cycle could ever even hope to compensate. The loss of low-lying property and other problems tied to rising sea levels would be in the tens of trillions [1].
People tend to bake in all sorts of potential negative externalities for nuclear energy, and much of that's baked into current regulations. Unfortunately, the same isn't done for other energy sources. When you start to try and look at those externalities (and that's a very difficult task, as many are hard to precisely measure), it can drastically change the math involved. And many of those negative consequences aren't a matter of chance. They're 100% guaranteed byproducts, especially with coal for instance.
0. https://en.wikipedia.org/wiki/International_Nuclear_Event_Sc...
1. http://www.un.org/climatechange/blog/2014/06/sea-level-rise-...
Here are the costs of just a single nuclear disaster, one considered minor: http://www.greenpeace.org/international/en/getinvolved/They-...
Please bear in mind that this was the most conservative estimate I found when googling.
According to the below post, even a relatively nuclear heavy country like the US would need 5 times more nuclear power to cover the electricity need: https://www.quora.com/How-many-nuclear-reactors-would-be-req...
So, on a global scale that would be many new reactors, including many in countries with dubious safety records in other areas.
[1] https://en.wikipedia.org/wiki/Radiation_effects_from_the_Fuk... [2] http://www.bbc.com/news/world-asia-35761136 [3] https://en.wikipedia.org/wiki/Environmental_migrant#Enumerat...
[1] https://en.m.wikipedia.org/wiki/Effects_of_global_warming