Recently, a report appeared on BBC News, titled “Nuclear power: Are we too anxious about the risks of radiation?” It has been re-posted or restated at a number of news sites since. It says that 28 people died as a direct result of the Fukushima explosion and exposure to the immediate radiation. It goes on to say that there were 15 deaths in the region from thyroid cancer. And the conclusion that it bases on these figures is specifically stated: “[A]ccording to the UN in 2005, just 43 deaths could be directly attributed to the worst nuclear disaster the world has ever seen.”
Really? Just 43 deaths? And this is according to the UN?
There is a lot of well documented material on the Chernobyl Disaster available. One article at Wikipedia is, “Deaths due to the Chernobyl disaster.” It states, “[T]here is considerable debate concerning the accurate number of projected deaths due to the disaster’s long-term health effects; long-term death estimates range from up to 4,000 (per the 2005 and 2006 conclusions of a joint consortium of the United Nations) for the most exposed people of Ukraine, Belarus, and Russia, to 16,000 in total for all those exposed on the entire continent of Europe, with figures as high as 60,000 when including the relatively minor effects around the globe.”
The other problem is that the nuclear industry has based much of its analysis on what I believe is demonstrably flawed thinking.
I would classify numbers from the nuclear industry in three types. Some are predictably spot-on. Some are predictably off, with the true value turning out to be greater by a factor of two or more. And some are predictably, and sometimes demonstrably, off by at least an order of magnitude. Which number falls into which category seems to depend entirely on what it is to be used for.
The numbers that are spot-on are scientifically calculable. One example is the number of electron-volts associated with a nuclear particle emitted from a specific atomic fission event. Another is the number of megawatts a certain nuclear reactor will deliver at full capacity.
The ones that seem usually off by a factor of two or more are those associated with such things as projections of costs and times it will take to build nuclear power plants. If a plant is expected to cost $6 billion and take six years to build, its actual cost is likely to exceed $12 billion, and it might take twelve years or more to build.
But it is the third type of number associated with the nuclear industry that the BBC article is dealing with. It asserts that the UN said 43 people had died as a result of the Chernobyl Disaster, and the total will be slightly higher. But what you would not know from reading it is that the UN report actually gives a different number that is close to two orders of magnitude greater for the overall death count. That number could be 4,000.
As you might expect, the UN has been scolded for providing numbers that were too low by some anti-nuclear activists. They prefer to say 60,000 people might have died, or even more. And even though activists’ numbers might be the highest ones around, I think it is more likely that they are right than that the total count is 43, a number that seems disingenuous, at the least.
Some people might ask who we can trust on such issues, when there are no demonstrable facts. There may not be a perfect answer for that. Fortunately there are some related numbers on which there is no dispute, because the figures are demonstrable facts. They do have to do with safety analysis. They are explained in a Wikipedia article, “Core damage frequency.” I will try for a shorter explanation.
Back in the old days, when GE was building its early Generation-II nuclear reactors, safety analysis showed that the likelihood of a reactor having a “core damage event” in the course of a year was one in ten thousand. With a thousand such units online, we could reasonably expect about one meltdown per decade.
Not many of those reactors went online before safety upgrades brought improved figures for safety analysis. The new figures showed a likelihood of a meltdown being about one in twenty thousand. And of course, this means we could reasonably expect a meltdown about every other decade, if we had 1,000 reactors going.
Newer reactors yet have higher figures for safety analysis. One event in 50,000 reactor-years is one promise. For some new designs, including the upcoming small modular reactors (small meaning about the size of a small house), promises include the impossibility of any meltdown at all.
We do not have anything close 1,000 reactors in service worldwide, however. We have about 440 operating, and additionally a bit over 180 that have been shut down. Together, they have operated through roughly 20,000 reactor-years. If they had all been older reactors, we might have expected two core damage events. But the great majority of them are of newer types. Reasonably, we should have expected only one meltdown, worldwide, since the first reactor went online.
But we have not had one meltdown. We have had eleven. This counts only reactors that went online to produce electricity for sale to customers. Three of these were in the United States: Three Mile Island, Fermi 1, and the Sodium Reactor Experiment, which, despite its name, was selling power commercially. Three reactors melted down in the Fukushima Disaster in Japan. Two more melted down, on separate occasions, in France. One each happened in the Soviet Union, Scotland, and Czechoslovakia. These are the ones we know of.
Predicting one, but getting eleven, is not a good score.
Trying to find out what went wrong, I have come to understand that because nuclear safety analysis sees human failure as an incalculable possibility, it is simply not included in the calculations. And of the eleven meltdowns, I would count all has having resulted in some important way from human failure.
In other words, safety analysis simply ignores the greatest cause of nuclear meltdowns. And this has made it utterly unreliable.
Nuclear advocates might cry foul at this. After all, what connection is there between human failure and the 14-meter tsunami wave that resulted in three reactors melting down at Fukushima? What analysis could have seen that possibility?
Actually, that event should have been predicted. And please note that I said “predicted,” not “hypothesized.” The highest tsunami waves from the Tohoku earthquake of 2011 were 37.5 meters. The Sanriku earthquake of 1933 produced a tsunami with the highest waves of 27.5 meters. And the Sanriku earthquake of 1896 was followed by a tsunami with the highest waves of 37.5 meters. All of these hit the northeast coast of Japan, where the Fukushima plants sit. The first two should have informed the design of the Fukushima Daiichi plant.
That plant should have been prepared, at the least, for the largest tsunami that hit in the hundred years before it was built. But the seawall at the plant was 5.7 meters high.
My guess is that it was budget, not safety, that dictated the height of the seawall. Regardless of cause, however, that design surely resulted from a human failure with tragic consequences. The seawall simply should have been much taller. The government seems to agree, as the plant is now required to build a 16-meter seawall to protect what remains of the six reactors, three of which were wrecked and three now permanently closed, according to an article in NHK World.
The point of this is that there are numbers brought forward by the nuclear industry and its apologists that are so grossly far from the truth – I am tempted to say “unrelated to the truth” – that they can only be regarded with the greatest suspicion. Unfortunately, the BBC article seems to push these bad numbers without even mentioning that there is a serious question about them. The BBC should do better.
I don’t often feel this way about BBC articles, but my response to this one is, “Shame on you!”
Photo: Wearing safety gear for Fukushima Daiichi. Voice of America image – public domain.