The worst case scenario of a nuclear reactor going wrong is much worse than any other alternative, however slim the chances of that happening may be. This fact is enough to be concerned about the safety of nuclear reactors.
Risk of death isn't the only issue. In a car accident you don't have to worry about cesium-137 rendering the area near the accident unusable by humans for 100 years or more.
Quantitatively, cars are extremely safe based on the number of drivers and the frequency with which they drive. If you put a billion people around Chernobyl during the accident the number of people dead would greatly outnumber the number dead from all car accidents throughout history.
Put another way, on average driving 100 million miles is safer than flying 100 million miles (contrary to popular belief that flying is safer than driving).
You can see the fallacy of "cars are dangerous" when you compare cars to motorcycles on that chart, because most people don't ride motorcycles so the probability of death for most people is lower than for a car.
Fukushima is extraordinarily safe for everyone living in Iceland. It is far less safe for everyone living in Japan.
The worst case in Japan would be a containment vessel blowing up from the pressure. That could kill the remaining workes on site and pose quite a messy cleanup problem, but already the 20 km evacuation zone could be seen as a bit of an overreaction.
However, that is not going to happen, because the containment is rather vented than letting it burst.
Anyway, the carnage and tragedy _actually having struck_ in form of the earthquake and tsunami _dwarf_ even the most horrible scenario at the nuclear plants.
