[saulrh]

Then you need to do more research. Fukushima was not part of the modern nuclear industry. Its technology and hardware and culture were fifty years old, every single part of them. If you go back fifty years in literally any other major technology it's all mind-blowingly unsafe. Fifty-year-old cars are deathtraps. Fifty-year-old houses are deathtraps. Fifty-year-old elevators are deathtraps. Fifty-year-old chemistry sets are deathtraps. Fifty-year-old airplanes are barely not deathtraps. And even though fifty-year-old technologies are pants-shittingly unsafe, you're perfectly fine using modern cars, and modern chemistry sets, and modern elevators, and modern airplanes. Because they're all safe. And you should be perfectly fine using modern nuclear power, because it's perfectly safe.

Besides, you know what? Even if nuclear power plants were as dangerous as the ones at Chernobyl and Fukushima I'd still be agitating for their adoption. Do you have any idea how minor and infrequent nuclear incidents really are compared to the alternatives? Chernobyl is so far out into the worst case column that it's not even reasonable. It's like badmouthing cars today because a tanker full of hydrofluoric acid crashed in downtown Seattle circa 1934. There are two major nuclear incidents in history. Two. Neither of them was as bad as an average dam burst. One of them wasn't even as bad as an average coal mine collapse. And the worst dam-bursts in history outweigh the worst nuclear disasters in history by four orders of magnitude. Go look up Banqiao Dam. Even the "green" alternatives are pretty awful; more people have died falling off wind turbines than will ever die to nuclear power. Compared to the alternatives, nuke plants are rainbows and sunshine.

[13years]

Most other accidents have casualties which are easily traceable. Comparing deaths to those that fall from wind turbines is not a fair comparison.

Nuclear accidents happen in slow motion and the effects are not as easily traceable. It will be decades before we know the impact and it will only be due to statistics. Just because you don't die today, doesn't mean you will not die earlier than you normally would have.

Also, to say they are infrequent is not necessarily clear description considering the Fukishima accident is still not over by any means. It is an accident that continues everyday. http://www.counterpunch.org/2015/06/29/is-fukushima-getting-... and http://enenews.com/officials-trillions-becquerels-radioactiv...

[saulrh]

OK, then, let's go tease out every iota of causality we possibly can from the statistics for various energy technologies. Let's start at the top. Hmm. "Global climate change resulting primarily from fossil fuel use has already caused tens of millions of deaths via famine, flooding, and extreme weather and stands a reasonable chance of killing every single human on earth over the course of the next hundred years". Well, that's not going to work.

[Synaesthesia]

Burning Coal releases more radiation into the atmosphere than all human nuclear activity. The particulates are the major killer though. I believe like 100 000 people die from the effects every year. And that's discounting the atrocious toll on the environment.

[13years]

This is true compared to normal nuclear facilities operating normally. However, it is not a comparison to nuclear activity as due to nuclear accidents. http://www.cejournal.net/?p=410

[saulrh]

And what he said above is also true of fossil fuel industry facilities operating normally. He doesn't consider pollution due to spills, oil well fires, or refinery explosions. If wells or refineries failed at the rate you're implying nuclear plants do, our entire planet would be an inch deep in crude. And, yes, I understand that there physically isn't enough oil on earth to do that. That's part of my point. The alternatives to nuclear are strictly worse. There are no tradeoffs. They are just worse in every way, whether they're operating as designed or not.

[13years]

My point is about the disregard of the significance of the Fukishima accident. You must include that in the analysis.

This was written in 2011, and Fukishima has been continuously releasing radiation ever since.

"...every 600 years worldwide coal combustion releases as much radiation as was released from the nuclear accident at Fukushima..." http://nuclearaustralia.blogspot.com/2011/12/coal-1-fukushim...

I'm not arguing in favor of any energy source that damages the environment; nonetheless, I would prefer we acknowledge risks where they exist. To not do so, would be to send us down the path of possibly less safe nuclear facilities.

[pdonis]

> it is not a comparison to nuclear activity as due to nuclear accidents.

But nuclear accidents are rare, and are not a necessary consequence of nuclear power. Coal ash is not rare, and it is a necessary consequence of burning coal.

[pdonis]

> "Global climate change resulting primarily from fossil fuel use has already caused tens of millions of deaths via famine, flooding, and extreme weather and stands a reasonable chance of killing every single human on earth over the course of the next hundred years"

You said statistics, not unproven speculation. (Which is not to say that there aren't statistics showing harms from fossil fuel use. Just that you should focus on actual data.)

[saulrh]

I agree. I'm just pointing out that, since we're talking about unproven speculation, any statement that fits "Chernobyl killed >3e4 people" is also not statistics.

[13years]

Global climate change doesn't have a definable and agreeable source and/or event that can be confined to definite set of parameters. The cause and effect chain is nebulous in comparison.

But damage caused by radiation is well known, definable criteria which is well studied as part of health studies. What would be more comparable would be studies of heart disease for example based on consumption of certain foods. We readily accept such studies as advice on what is health, what is not, and also what drugs work and which do not. The principle is the same.

[mikeash]

Even if you ignore climate change (which is, frankly, irresponsible at this point), coal pollution alone kills something like a million people per year worldwide.

[dredmorbius]

I'm quite familiar with the Banqiao Dam disaster, and posted a longish comment regarding it on HN recently: https://news.ycombinator.com/item?id=9927596

Borrowing from that:

The worst power plant accident of time, not a nuclear power plant failure, but a hydro station in China, Banqiao Dam. It's instructive several ways:

Any number of fairly simple methods would have hugely alleviated the impact of the disaster. Much as with major nuclear disasters, it was a cascade of failures, starting with poor management and a dysfunctional culture, amplified through poor design, adverse conditions, poor communications, delayed or absent warnings, and little or no disaster response. Many of the deaths were attributed to starvation and disease, not drowning or other physical impacts.

A useful thing to keep in mind, though, is that after a dam break is done being a a massive disaster area, which typically resolves in a few hours to a few weeks, the land is no longer a glowing radioactive mess. It can be re-settled and populated as structures and infrasctructure are rebuilt. Zhumadian City, the region surrounding Banqiao, has a present population of over 7 million.

The disaster area is also contained. Chernobyl threatened vast regions of Asia and Europe, putting a population of over 400 million at risk, and certainly ill at ease.

Or look up the story of the Johnstown Flood, worst dam break in US history (by deaths), which saw the emergence of the Red Cross, of national response to disasters, and changes in liability laws.

(Excepting Johnstown and Banqiao, dam failure mortality falls off rapidly, with another 8 disasters of 1,000+ lives. Wikipedia gives some 908 notable dams, and 137 hydroelectric facilities of 1GW+ net capcity.)

There are other questions, notably whether or not "deaths per GWh generation" is the most appropriate measure of risk. Particularly for a technology whose risk tail spans not years, decades, or even centuries, but millennia. Or longer.

Last I checked, there were few human institutions with lifespans of similar scale. Technical or otherwise.

https://en.wikipedia.org/wiki/Banqiao_Dam

https://en.wikipedia.org/wiki/Zhumadian

https://en.wikipedia.org/wiki/Johnstown_Flood

https://en.wikipedia.org/wiki/Dam_failure

https://www.reddit.com/r/dredmorbius/comments/2awjj2/thought...

https://www.reddit.com/r/dredmorbius/comments/283sz1/key_fac...