Print 33 comment(s) - last by Mumrik.. on Apr 15 at 9:54 AM

Bill Dedman mislead the public in his claims about an NRC report and he told us misleading statements. Thanks to the U.S. government, we caught him in that falsehood.  (Source: MSNBC)

Our nation's nuclear plants are at no significant risk.  (Source: James Marvin Phelps)

The U.S. Nuclear Regulatory Commission helped us expose's misinformation and get the true story to the public.  (Source: Brendan Hoffman/Getty Images North America)

The public is already uninformed and unnecessarily fearful about nuclear power. Mr. Dedman's report spread damaging misinformation and alarmist inaccuracies to thousands in the American public.  (Source: Greenpeace UK)
Nuclear fear-mongering for profit -- government provides our strongest evidence in stunning tale of misinformation

On Wednesday published a story, which claimed to analyze a report [PDF] sponsored by the U.S. Nuclear Regulatory Commission.  That story, written by Investigative Reporter Bill Dedman discussed which commercial U.S. nuclear plant was at "the most risk" of exposing the public to radiation.  

We wrote a piece on Wednesday criticizing numerous factual inaccuracies in Mr. Dedman's piece.  At the same time we contacted the U.S. Nuclear Regulatory Commission (NRC).  On Friday, after two lengthy phone interviews and an email dialogue with the NRC we had the complete story -- the NRC backed nearly every one of our assertions.

All journalists make mistakes.  But Mr. Dedman made nearly every one in the book in this report.

Sensationalism and factual errors -- the report and Mr. Dedman's assertions demonstrate an appalling disregard for the facts and a blatant attempt to alarm the public.

But don't take our word for it, read the facts.

I.  Reporter Refuses to Correct Factual Inaccuracies: Report DID NOT Assess Public Exposure

Mr. Dedman ( writes:

It turns out that the U.S. Nuclear Regulatory Commission has calculated the odds of an earthquake causing catastrophic failure to a nuclear plant here. Each year, at the typical nuclear reactor in the U.S., there's a 1 in 74,176 chance that the core could be damaged by an earthquake, exposing the public to radiation.

But the U.S. Nuclear Regulatory Commission says they calculated no such risk.  

The report itself states:

In contrast [to the seismic core damage frequency], the containment performance analyses conducted under the IPEEE program did not produce sufficient quantitative information to allow the estimation of either LERF or public dose.

That seems pretty clear -- the report does not talk about the risk of public exposure to radiation, so we helpfully suggest to Mr. Dedman:

This seems to be a clear cut factual error that's misleading and disingenous -- be it intentionally or unintentionally so. It seriously discolors the estimates and makes them something they explicitly are said to NOT be by the NRC.

Mr. Dedman writes us back:

No, Jason, the article is about core damage, which the NRC says would release radiation. You've decided that I must have been talking about something else, which I wasn't, and now you're saying, why aren't about that something else...

That is a clear mistake -- intentional, or unintentional.

Engaging in the due diligence that Mr. Dedman neglected to we discussed Mr. Dedman's comments and our analysis with government authorities at the U.S. regulatory commission.  They told me they never told him that.

Their spokesperson, Neil Sheehan writes us:

Seismic CDF is the probability of damage to the core resulting from a seismic initiating event. It does not imply either a meltdown or the loss of containment, which would be required for radiological release to occur. The likelihood of radiation release is far lower.

That was only the first of several falsehoods and factual errors in Mr. Dedman's correspondence and work that we were able to definitively verify.

States a separate NRC team member, "There were numeous inaccuracies in that story."

And this was but the first.

[MSNBC and its employee Mr. Dedman have not corrected this error in their story, despite knowing about it, at the time of this article's publication.]

II. Another Mistake - Was Told to Use the "Weakest Link" Model

In Mr. Dedman's original piece he writes:

The chance of a core damage from a quake at Indian Point 3 is estimated at 1 in 10,000 each year. Under NRC guidelines, that's right on the verge of requiring "immediate concern regarding adequate protection" of the public.

In our report we question this number, pointing out that there's three different models and Mr. Dedman seemingly purposefully picked the most severe one.  The other models take into account other scenarios of vibration damage so they seem equally valuable to "weakest link" scenario, as it's unclear what parts would be damaged by what vibration frequency.

Mr. Dedman writes us stating:

You're cherry picking. You've decided that the weighted average is the right column to use. Based on what? The NRC staff prefers the column that we've used, the "weakest link." That's the number it sent us, when it sent us one number for each plant. And as the report explains, the NRC has no basis on which to weight the averages, so it says a weighted average wouldn't be meaningful.

There are three separate falsehoods in this statement.  As you will see, the NRC told Mr. Dedman nothing of the sort and he's clear mislead me, as he's done to his readers.

And again, Mr. Dedman mistakes the work of the USGS, for the NRC, clearly indicating his lack of understanding of the material he's writing on.

We write him:

Do you have a contact at the NRC who can substantiate your claims? How can you weight data without having a factor to do so? If you get me this information I can [edit my article].

Virtually always weighted data is what you would use in a case like this, as the data is typically weighted by the frequency of occurrence of the event (e.g. a probability of the probability). It's possible your correct, that would just be a bit unusual.

Mr. Dedman refused to provide us the identity of his phantom "contact" at the NRC, so we contacted them ourselves.

We asked them if they told Mr. Dedman to use this figure or told him that the weighted average was non-meaningful.

We inquire:

Did an NRC spokesperson tell MSNBC's Bill Dedman that the weighted risk was invalid and to use the weakest link model?

They respond:


And they add:

The weighted average is not invalid (see Answer 5 below). All of the values in Appendix D were developed by NRC staff. Table D-1 in Appendix D uses the (2008) US Geological Survey (USGS) seismic source model, but the Seismic Core Damage Frequency results were developed by US NRC staff. The USGS seismic source model is the same one used to develop the USGS National Seismic Hazard Maps.

Tables D-1 through D-3 in Appendix D of the US NRC study show the “simple” average of the four spectral frequencies (1, Hz, 5 Hz, 10 Hz, peak ground acceleration (PGA)), the “IPEEE weighted” average and the “weakest link” model. These different averaging approaches are explained in Appendix A.3 (simple average and IPEEE weighted average) and Appendix A.4 (weakest link model). The weighted average uses a combination of the three spectral frequencies (1, 5, and 10 Hz) at which most important structures, systems, and components of nuclear power plants will resonate. The weakest link is the largest SCDF value from among the four spectral frequencies noted above.

Most nuclear power plant structures, systems, and components resonate at frequencies between 1 and 10 Hz, so there are different approaches to averaging the Seismic Core Damage Frequency (SCDF) values. By using multiple approaches, the NRC staff gains a better understanding of the uncertainties involved in the assessments.

In other words, each model is important to gaining a full understanding of various possible scenarios and Mr. Dedman erroneously selected the most sensational model and then falsely claimed the NRC told him to.

The NRC adds:

The weakest link model is a method for evaluating the importance of different frequencies of ground vibration to the overall plant performance. The model and its details are not integral to understanding the fundamental conclusions of the study.

That conclusion?  The nation is quite safe (as we write in our piece).

[MSNBC and its employee Mr. Dedman have not corrected this error in their story, despite knowing about it, at the time of this article's publication]

III.  More Misinformation -- Did the Report Evaluate Risk at All 104 Plants?

We admit, we missed this error in Mr. Dedman's report initially, but the NRC pointed it out for us:

The US Nuclear Regulatory Commission study, released in September, 2010, was prepared as a screening assessment to evaluate if further investigations of seismic safety for operating reactors in the central and eastern US (CEUS) are warranted, consistent with NRC directives. The report clearly states that "work to date supports a decision to continue …; the methodology, input assumptions, and data are not sufficiently developed to support other regulatory actions or decisions." Accordingly, the results were not used to rank or compare plants. The study produced plant-specific results of the estimated change in risk from seismic hazards. The study did not rely on the absolute value of the seismic risk except to assure that all operating plants are safe. The plant-specific results were used in aggregate to determine the need for continued evaluation and were included in the report for openness and transparency.

In other words Mr. Dedman claimed the study looked at all plants and discussed what risk they were at.  It did not.

The NRC adds:

The plant-specific results were used in aggregate to determine the need for continued evaluation and were included in the report for openness and transparency. The use of the absolute value of the seismic hazard-related risk, as done in the MSNBC article, is not the intended use, and the NRC considers it an inappropriate use of the results.

In other words, Mr. Dedman abused the data to support his own fallacious conclusions.

Mr. Dedman accused me:

I don't mind criticism at all, but twisting of facts...

We agree with his assessment.  His statements were unacceptable.

[MSNBC and its employee Mr. Dedman have not corrected these errors in their story, despite knowing about it, at the time of this article's publication.]

IV.  Textbook Alarmism

Mr. Dedman claimed the results of the study showed that the chance of a plant disaster was as likely as winning a $10,000 prize in the national Powerball lottery.  This is factually inaccurate.  The powerball frequency per population is greater than 1 per year.  The chances of a core meltdown per population are less than 1/7,400th per year, according to the report.

Further, Mr. Dedman writes:

How much risk is too much? Is a roller coaster safe only if no one ever dies? If one passenger dies every 100 years? Every year?

This is alarmism at its finest.  People are not dying every year from quake-induced plant damage in the U.S.  And there's little chance a single life will be lost over the course of the next century from quake damage in the U.S.

The report explicitly states:

Plants have seismic margin and the results of the GI-199 Safety/Risk Assessment confirm that overall seismic risk estimates remain small. GI-199 is not an adequate protection issue.

And the NRC tells us:

The study is still under way and it is too early to predict the final outcome. However, the NRC staff has determined there is no immediate safety concern and that overall seismic risk estimates remain small. If at any time the NRC determines that an immediate safety concern exists, action to address the issue will be taken. The NRC is focused on assuring safety during even very rare and extreme events. Therefore, the agency has determined that assessment of updated seismic hazards and plant performance should continue.

[MSNBC and its employee Mr. Dedman have not removed these misleading statements from their story, despite knowing about it, at the time of this article's publication.]

V. The Moral of This Story

The moral of this story is that sensationalism and alarmism may be a ticket to cheap page views, but when you get the facts wrong and then cook up tall tales to cover up your tracks, someone will eventually call you out.

One must wonder what thinks of this performance.

After all, Mr. Dedman's story apparently touched off a U.S. Senate inquiry into plant quake safety.  In other words it not only created mass public panic and misinformation, it triggered a knee-jerk response by the federal government.

Most journalists make mistakes, but most don't reach as many people as Mr. Dedman does or trigger government inquiries.  And few journalists would refuse to fix blatant factual errors like Mr. Dedman did when we clearly discussed these points with him.

Thus far Mr. Dedman has defiantly refused to correct the numerous factual inaccuracies and mistakes in his piece, even when his source, the NRC, clearly stated his story was inaccurate.

We feel that it is critical to's reputation as a legitimate news site that this story be removed and/or corrected immediately.  And they need to take a long hard look at Mr. Dedman's reporting and how he conducts himself.

Otherwise, they are allowing themselves to become a tabloid.

Update 1: Saturday, March 19, 2011 2:20 a.m.

We've temporarily removed a paragraph referencing the ownership of MSNBC.  It appears that Google Finance and Yahoo Finance may have inaccurate information with regards to Microsoft's stake.  We've contacted MSNBC for more information and hope to resolve the question of ownership briefly.

We also amended the text in the opening paragraph to clarify that the NRC was responsible for Appendix D, but was not responsible for any assertion about public radiation exposure.'s Bill Dedman sent us the following statment with some remaining criticism of this post:

Mr. Mink (sic) chastises me for using the weakest-link number in the NRC report, but that's the number the NRC staff provided to me. Several times I sent to Mr. Mick a copy of the spreadsheet the NRC sent to me, containing the weakest-link model as the single number it reported for the risk estimates. (We already had the eastern and central plant data from the report; this spreadsheet, for all plants, gave us the western plant data as well, as we told our readers.) Mr. Mink (sic) fails to mention this in his article. The point, as I told him, was that the NRC staff uses the weakest-link model as the best representation of the risk, and, as I explained to our readers, this is the most conservative estimate. I also pointed out to Mr. Mink that the NRC report describes that its actually has no basis for knowing how to weight a weighted average in this case; he fails to mention this.

The NRC says, "the results were not used to rank or compare plants." Mr. Mink (sic) twists this: "In other words Mr. Dedman claimed the study looked at all plants and discussed what risk they were at. It did not." That's completely nonsensical. Yes, the NRC looked at all the plants, and made an estimate of the risk at each one. Did it rank them? No. We ranked them, from the NRC data, just as we explained in the original article.

If the newspapers starts reporting the American League East standings in alphabetical order, the reader is free to arrange them by winning percentage.

He later adds:

Perhaps you're naive. Perhaps you have a bias for nuclear power. Perhaps -- who knows? But for some reason, you've fallen for the oldest page in the government PR playbook: the non-denial denial.

We will let you know if he shares any further criticism.

Comments     Threshold

This article is over a month old, voting and posting comments is disabled

Jason is right
By Solandri on 3/19/2011 3:25:53 AM , Rating: 5
For those of you having problems picking through what all this means:
The chance of a core damage from a quake at Indian Point 3 is estimated at 1 in 10,000 each year.
Tables D-1 through D-3 in Appendix D of the US NRC study show the “simple” average of the four spectral frequencies (1, Hz, 5 Hz, 10 Hz, peak ground acceleration (PGA)), the “IPEEE weighted” average and the “weakest link” model.

Basically, when an earthquake shakes the ground, they do so at a variety of frequencies. IIRC (it's been 20 years since I studied this) most of the shaking is around 10 Hz for most quakes, but obviously every quake is different and there's a wide variety frequencies.

Every object has a resonance frequency. Think of a child's swing. The number of the times the child swings back and forth in a second is the resonance frequency of the swing. If the child kicks her legs at a different frequency, the height of her swings doesn't change, or may even become lower. But if she kicks her legs at exactly this frequency, she beings to swing higher and higher. This is called resonance. If she kicks at close to this frequency, she'll swing a little higher, but not as much.

When you design a building, it's pretty easy to design it to withstand the accelerations you'd get from an earthquake. If the maximum earthquake is expected to generate 3g acceleration (less than 1g is typical, 1.5g is very large, but let's go with 3g), you just design your building to withstand 4g.

The problem is resonance. If the shaking happens to be at the building's resonance frequency, then with each shake of the ground, the building will shake more than the previous shake, and thus experience bigger accelerations. Like kicking yourself higher on a swing, or sloshing water around in a bathtub, the vibrations build up becoming bigger each time, until you exceed the acceleration the building was designed for and the whole thing comes crashing down.

This is what took down Galloping Gertie. The frequency with which the wind made the bridge sway back and forth exactly matched its resonance frequency. The shaking got bigger and bigger with each sway, until it exceeded the bridge's strength and the whole thing tore itself apart.

So all buildings and structures have a resonance frequency (there are smaller harmonics just like with music, but they're generally less of a problem). The question then is, how much acceleration at a resonance can your structure take? If the accelerations are small enough, mechanical energy losses (friction and something called hysteresis - "stretching" losses - in the deformation of the building material) are enough to dampen the resonance effect so the swaying, while bigger than the shaking which is driving it, stops growing before it gets big enough to destroy the building.

Finding all the resonances for every structure is impractical (unless you're building in earthquake territory, in which case you use all sorts of shakers and accelerometers after the building is made to find out exactly what its resonance response is at all frequencies, not just 4). So what the NRC did was look at the tendency of the nuclear plant buildings/cores to resonate at 1 Hz, 5Hz, 10 Hz, and PGA (peak ground acceleration - the frequency of the strongest single jolt). The NRC then averaged these results to come up with a probability of core failure given an earthquake of indeterminate frequency.

The "weakest link" model is the probability of core failure if the earthquake just happens to shake at the frequency which causes the strongest resonance response in the building. This is the figure MSNBC used in their "Indian Point 3 is estimated at 1 in 10,000 each year" quote.

But that's the chance of the building failing if you assume every earthquake happens to shake at the building's most vulnerable resonance frequency. To properly get a "1 in x each year" probability using the "weakest link" probability, you have to take the weakest link probability and multiply it by the probability that the earthquake will just happen to shake at that specific frequency.

Vastly oversimplifying, they analyzed 4 frequencies, so if you assume earthquake shaking at each of those 4 frequencies is equally probable, then the chance of the "weakest link" figure being relevant is 1 in 4. So the correct value MSNBC should've used (using this vastly simplified assumption) was 1 in 4*10,000, or 1 in 40,000.

But there's really no need to do that since the NRC already did the grunt work of calculating the average failure probability across all frequencies. Just use that and be done with it, which is what Jason is saying.

(I also find the inclusion of PGA a bit suspect, since shaking at the peak acceleration frequency is usually just one jolt, not sustained shaking. But meh, the NRC guys probably know what they're doing.)

RE: Jason is right
By Solandri on 3/19/2011 3:36:47 AM , Rating: 2
Incidentally, the frequency of most quakes happens to match up with the typical peak resonance frequency of a 3-story building. If you look at the buildings that collapsed in the Northridge and Loma Prieta quakes, they were all 3-stories. The one building collapse I've seen from this quake in Japan was a 3-story hospital or factory. So if you're buying in earthquake territory, avoid 3-story buildings.

The typical resonance frequency of a skyscraper is less than 1 Hz, so they're actually one of the safest places you can be during an earthquake. The amplitude of the swaying may be disconcerting (it can be several meters), but if it's up to code it's designed to withstand it.

RE: Jason is right
By Lifted on 3/19/2011 11:08:30 AM , Rating: 2
Jason is right, and the story was pulled from the front page "news" category of AnandTech.

I see.

RE: Jason is right
By Solandri on 3/19/2011 3:47:29 AM , Rating: 3
Oh yeah, to put the worst-case "1 in 10,000" each year in context, your chance of dying in an automobile accident each year is about 1 in 8,000. So even if the "1 in 10,000" chance of core damage (not leakage, not meltdown, not death) figure had been correct, if that's unacceptably high for this country to accept, then we need to ban all cars too.

RE: Jason is right
By Keeir on 3/19/2011 4:57:12 AM , Rating: 4
I agree.

I find the inability to put risk in context to be frustrating.

Dead is Dead. Enviromental Damage is Enviromental Damage.

Over the last 50 years, Nuclear Power has shown that it has the lowest risks of Death and Enivormental damage of any large scale deployed power source.

Even if the worse case situation happens in Japan, the number of dead/injuried by Nuclear raditation and the enviromental damage will be -less- than caused by the earthquake. Yet millions of people -choose- to live in active fault zones each year. (100s of million live in active fault zones, but I admit not everyone has a reasonable choice)

In fact, your much more likely to die/be injured from the -earthquake- then the resulting Nuclear Core Damage... even if you live right next to the Nuclear Plant. 800,000 have died the past 10 years from earthquakes.

RE: Jason is right
By Solandri on 3/19/2011 6:07:08 AM , Rating: 3
Ran another calc on the numbers. I took all the IPEEE weighted probabilities for core damage using 2008 USGS seismic damage data, tossed them in a spreadsheet (accuracy raised to 10^-12), and multiplied them all together. There were 108 reactors (I thought the country only had 104 commercial reactors, so maybe it's including military reactors?).

The data only covers reactors in the East and Central U.S., but there aren't that many reactors in the Western U.S.
Due to the higher prevalence of earthquakes in the West, the reactors there probably have their own individual earthquake damage probability assessment study done. Like I said in the big writeup, big buildings built in earthquake-prone areas typically get a thorough test done with shakers so you can record their full frequency response curve (how much they resonate across all frequencies, not just the 4 picked in this study).

The chance for all 108 to operate safely in a year is 99.8761%

The chance for a reactor to experience core damage from an earthquake in a year is 0.1239%

In other words, the NRC's estimate is that on average, once every 807 years one of our reactors in the East or Central US will experience core damage from an earthquake.

RE: Jason is right
By nafhan on 3/21/2011 9:40:57 AM , Rating: 2
Excellent analysis!
Also, do you think the "extra" reactors might be research reactors?

RE: Jason is right
By superPC on 3/19/2011 6:46:51 AM , Rating: 3
you know sometimes i think people with power and influence deliberately slow nuclear development and deployment. let's face it: our whole economic model and theory are ALL based on scarcity. everything is scarce, nothing is abundant. but with nuclear nothing would be scarce anymore.

all materials can be constantly recycled because we would have more than enough energy for it for the next millennia even if we continue to use fission nuclear reactor. power can be use to desalinate sea water and use for everything that we will ever need. power can even be use to convert coal to gasoline and allow us to use it for another century or two. power can also be use to light up green houses all around the world 24/7 and climate control all of them thereby multiplying all our plant productivity by a factor of two or more. hell with enough power we can use underground cable to inductively charge electric cars that run above them thereby making electric cars with no range limitation. basically with nuclear we will have enough power for everyone to create everything they ever need: food, raw material, water, transportation. no more scarcity. and we already know how to do all this TODAY, not 10 years from now, or a century from now but right now. but if that actually happened a lot of people would loose a lot of power and influence and they won't give up all those power and influence willingly.

RE: Jason is right
By nafhan on 3/21/2011 12:31:14 PM , Rating: 3
You might be onto something, but I think it's mostly (even among the rich/powerful) a combination of ignorance and apathy. You don't have to know much about nuclear tech to know more than most people do, and politicians and reporters are taking advantage of that by spreading fear and uncertainty to their advantage.

RE: Jason is right
By Dorkyman on 3/22/2011 4:42:57 PM , Rating: 4
I'd go one step further. For many people, it's not so much "ignorance and apathy," it's that nukes don't conform to their worldview, their ideology.

With significant numbers of the U.S. population, "green" is a religion, just as much a religion as Islam or Catholicism. As such, science and logic take a back seat to faith. So arguing logic about nukes just falls on deaf ears for some people.

All we can hope is that in the future the politicians guiding our path are grounded in science and logic. Frankly, I wouldn't count on it. I suspect the future belongs to other countries, not the USA.

"Well, there may be a reason why they call them 'Mac' trucks! Windows machines will not be trucks." -- Microsoft CEO Steve Ballmer

Most Popular Articles5 Cases for iPhone 7 and 7 iPhone Plus
September 18, 2016, 10:08 AM
Laptop or Tablet - Which Do You Prefer?
September 20, 2016, 6:32 AM
Update: Samsung Exchange Program Now in Progress
September 20, 2016, 5:30 AM
Smartphone Screen Protectors – What To Look For
September 21, 2016, 9:33 AM
Walmart may get "Robot Shopping Carts?"
September 17, 2016, 6:01 AM

Copyright 2016 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki