How dangerous is nuclear power compared with coal, gas and wind?

Measured per unit of electricity produced, studies find nuclear power among the lowest in deaths, close to wind and solar and far below coal, oil and gas. That is the view of researcher Hannah Ritchie at Our World in Data, who compares energy sources by deaths per terawatt-hour, counting accidents and air pollution. One terawatt-hour is roughly the yearly electricity use of 150,000 people in the European Union.

Her numbers for nuclear count 433 deaths from Chernobyl and 2,314 from Fukushima. Even with those two disasters included, she reports that nuclear causes 99.9 percent fewer deaths per unit of energy than brown coal and 97.6 percent fewer than gas. If an EU-sized town ran entirely on nuclear, someone would die about every 33 years. The fossil fuel figures come from a 2007 study by Anil Markandya and Paul Wilkinson in The Lancet, based on European plants, which Ritchie calls conservative. The wind and solar figures draw on a 2016 database of energy accidents compiled by Benjamin Sovacool and colleagues.

There is a catch to note. The comparison rests on assumptions, and critics raise different concerns. Charles Perrow, in a book inspired by the 1979 Three Mile Island accident, argued that complex high-risk systems are prone to accidents however well they are managed. Supporters reply that modern designs add passive safety systems that need no human action. Even the Fukushima response is argued over: the evacuation was accused of causing more harm than it prevented.

So the argument is rarely about whether nuclear is dangerous at all, but about how to compare rare large accidents with steady, spread-out harm from burning fuel. Both sides point to the same numbers but weigh them differently.