Re: Noone harmed? Pull the other one! Try 2500 deaths.
So you quote two articles above, one of which predicts 2500 cases of cancer, the other 1300 deaths. That's a little over a 50% mortality rate, which is a little harsh given remission rates for multiple types of cancer in Japan, but lets play along.
According to statistics compiled by the Ministry of Health, Labor and Welfare, 353,000 people died of cancer in Japan in 2010, accounting for one in every three deaths. [source http://www.jcancer.jp/english/cancerinjapan/]
So, assuming their projected figure of 1300 deaths from cancer occur in a single year (it would almost certainly be spread out over a number of years, if not decades, but we'll take this as a worst case scenario seeing as that's what you're apparently concentrating on), then at 2010 cancer death rates in Japan, that would represent an increase of just 0.3%. Taken in another context, the population of Fukushima prefecture in 2010 was just over 2 million [source: http://en.wikipedia.org/wiki/Fukushima_Prefecture]. Based on that figure, approximately 0.125% of residents *may* develop cancer as a result of this event during their lifetime.
The tsunami itself and its associated damage was responsible for (at last count) 15,854 deaths. [source http://www.telegraph.co.uk/news/worldnews/asia/japan/9132634/Japan-earthquake-and-tsunami-478-bodies-remain-unidentified-one-year-on.html]
So a natural disaster of epic proportions caused a meltdown of two reactors in an ill-maintained, generation-one reactor (read as 'outdated as all hell') that wouldn't have passed routine regulatory inspections in either the US or UK, which has no deaths directly attributable to it at this stage, *may* (the term used in the study you quoted) cause an uptick in cancer deaths of less than 1% at 2010 levels in Japan if all the proposed fatalities occur in a single year, therefore less than a 0.1% annual increase (not compounding) if any cancers emerge over the period of a decade, which is a more likely timespan.
I'm already paying out the nose for "environmental technology subsidies" every month to my local power co for wind and the like. I'd be far, far happier for the same amount to go instead to the construction and maintenance of a generation-3 nuclear reactor now, or further research into generation-4 reactor technologies. And yes, I'd be happy to live next door to it. Why? Because I understand that the risk of having something bad happen in a new plant with modern technology, trained and experienced staff, overseen by an anal-retentive regulatory body (probably one of few instances where this is a good thing), built in an area that's not prone to floods, seismic activity or other geological or natural disasters are miniscule.
And even if something were to happen to the plant on a comparable scale, I'll take the 1 in 800 odds that the event would result in cancer (figure based on the possible deaths quoted by the study you reference as measured against the approx population of the area).
In return for this, I get reliable power that is a near-zero carbon emitter fueled by an energy source that will be available for centuries, if not tens of centuries if the technology referenced in this article can be adopted. But the general population hears the word "nuclear" and instantly stops thinking rationally. The NIMBY brigade are no better. However, I'm willing to bet that people will change their tune pretty damn quick when our fossil fuel supplies dry up, the lights (and heat) go out, and it's zero degrees outside (that's Farenheit daytime temperatures). Unfortunately by then, it'll be too late.
Your last sentence however is brilliant, considering the content of your post. Have a petard.