I don't know what the population density is, from your graph, it sure looks like Three Mile Island prevented a bunch of cancer. (Assuming of course you want to infer causality from one picture.) Look at all that dark green - those are areas with 50% less cancer than would be expected.
The wikipedia article suggests the science is very mixed, and there is no conclusive evidence one way or the other.
The red areas almost look like where the wind would carry a radioactive plume. Why do you believe it would go towards the green area and disperse in a uniformed way?
There have been several studies of health effects in the aftermath of the Three Mile Island incident, some (like this one) found effects, some didn’t.
But that’s not even that important to the point I was trying to make. The amount of radioactivity released into the atmosphere was very low, in a real disaster (like in Chernobyl) much, much more radioactivity is released – more than any one coal power plant could ever release.
It looks to me as though one or two Chernobyl sized disasters per century can very much upset that balance.
Deaths obviously still matter but I don't think looking only at radioactivity is good enough. I would very much like to know the deaths per kWh for energy produced from coal and nuclear energy. It is certainly possible that infrequent nuclear disasters are, well, worth it.
(A single airplane crash kills many more people than a single car crash. Airplanes are still much safer than cars since they crash so infrequently compared to cars. Nuclear energy might very well be in a similar position, I honestly don't know.)
Chernobyl killed 50 people immediately, and is estimated to have killed 4000 people worldwide from excess cancers. That's worst case, assuming there's no minimum threshold for radiation causing cancer. Some scientists think there is a threshold, which would drop the number quite a bit. There's too much statistical uncertainty to know for sure either way.
TMI is estimated to have caused only one excess cancer. So we're looking at 4000 people killed over the lifetime of nuclear power world-wide, compared to 24,000 people per year killed by pollution from coal, in the U.S. alone.
Coal accounts for 50% of the U.S. energy supply, compared to 8% for nuclear.
Can't read German. My source is Physics for Future Presidents by Richard Muller, a physics professor at Berkeley. He says that's the United Nations estimate.
Wikipedia also says the estimate is 4000, and references a report by the United Nations Scientific Committee of the Effects of Atomic Radiation. It goes on to say that UNSCEAR later decided that was an over-estimate, and that IAEA data corroborates that. It has citations for everything.
Nuclear energy is actually the safest form of energy generation we have (in terms of "deaths per gigawatt-hour"), although wind and hydro energy is approximately the same. This is as presented by the ExternE project financed by the "EU Commission".
"http://manhaz.cyf.gov.pl/manhaz/strona_konferencja_EAE-2001/... - Polenp~1.pdf"
http://www.southernstudies.org/images/sitepieces/wing_tmi_ca...