Friday, May 2, 2008

Sovacool Strikes Again

Benjamin Sovacool - or as he is known to his friends as Dr Benjamin K. Sovacool - is at it again. This time he is attacking nuclear power by arguing in effect that nuclear power plants are accident prone. I have elsewhere in this blog and in bartoncii noted the accident history of dams and the potential danger of dam accidents. Dr Benjamin K. Sovacool has used his connection with the scitizen web site to launch an anti-nuclear crusade. His latest posting is based on a paper he is publishing this month: “The Costs of Failure: A Preliminary Assessment of Major Energy Accidents, 1907 to 2007.” Sovacool's scitizen essay is called, "The Costs of Major Energy Accidents, 1907 to 2007." I will presently show that both titles are misnomers because Sovocool ignores thousands of energy related accidents and tens of thousands of of energy related deaths.

Sovacool tells us, "From 1907 to 2007, a new study finds that 279 major energy accidents in the coal, oil, natural gas, hydroelectric, and nuclear sectors have been responsible for $41 billion in damages and 182,156 deaths." Of course the study is his own. Notice he refers to energy sectors in his introduction. The term "energy sector" is usually understood to refers to the exploration, production, marketing, refining and/or transportation, and use of energy sources including oil and gas, coal, nuclear energy, renewable energy and alternative fuels.

Sovacool asks, "what counts as an energy “accident,” especially a “major” one?" We will presently see that a great many energy sector accidents do not count in Sovacool's research.

How does sovacool get his information. He tells us, "by searching historical archives, newspaper and magazine articles, and press wire reports from 1907 to 2007." this would in fact be far to great a task for one person to undertake in a lifetime, let lone a person of such great intellectual accomplishments that he holds a PhD from an institution of higher learning found in Blacksburg, Virginia. There are however shortcuts. Sovacool tells us. "The words “energy,” “electricity,” “oil,” “coal,” “natural gas,” “nuclear,” “renewable,” and “hydroelectric” were searched in the same sentence as the words “accident,” “disaster,” “incident,” “failure,” “meltdown,” “explosion,” “spill,” and 'leak.'" well this distinctly sounds like that well known research method called googling.

How did Sovacool decide to pick out accidents to study? He reports his criteria as follows:

# The accident must have involved an energy system at the production/generation, transmission, and distribution phase. This means it must have occurred at an oil, coal, natural gas, nuclear, renewable, or hydroelectric plant, its associated infrastructure, or within its fuel cycle (mine, refinery, pipeline, enrichment facility, etc.);
# It must have resulted in at least one death or property damage above $50,000 (in constant dollars that has not been normalized for growth in capital stock);
# It had to be unintentional and in the civilian sector, meaning that military accidents and events during war and conflict are not covered, nor are intentional attacks. The study only counted documented cases of accident and failure;
# It had to occur between August, 1907 and August, 2007;
# It had to be verified by a published source;

Almost immediately Sovacool's research failure emerges. He writes:
"While responsible for less than 1 percent of total energy accidents, hydroelectric facilities claimed 94 percent of reported fatalities. Looking at the gathered data, the total results on fatalities are highly dominated one accident in which the Shimantan Dam failed in 1975 and 171,000 people perished."

In fact not one but 62 Chinese dams failed in the 1975 dam disaster, and the Shimantan Dam was not the largest.

Sovacool reported finding "279 accidents" which meet his criteria. This is an astonishingly small number, and is a certain clue that something is seriously amiss with Sovacool's study.

How far has Sovacool missed the marek in his study? He tells us "The second largest source of fatalities, nuclear reactors, is also the second most capital intense, supporting the notion that the larger a facility the more grave (albeit rare) the consequences of its failure."

In fact had Sovacool not made such stupendous blunders in his research, he would have known that neither of these assertions were true. Arguably the largest number of fatalities are associated with the hydroelectric sector since around 200,000 people were killed by or died as a result of the collapse of several dozen Chinese dams in 1975. It is quite possible that the cumulative death tolls for the coal mining industry is higher than the death toll for hydro. Sovacool mentions one oil pipeline in Nigeria, actually there were several. A 1998 accident at Jesse, Nigeria killed 1200 people. Two 2006 accidents killed 150 and 500 people. An oil pipeline explosion kills 508 in Cubatão, Brazil, during 1984. Other oil pipeline disasters have occurred.Sovacool also ignores oil welk fires, and both oil refinery fires and explosions. One explosion and fire in Texas City, in March 2005 killed 15 people and did hundreds of millions of dollars worth of damage. The U. S. Occupational Safety and Health Administration levied a $21 million fine against BP after the fire. A Shell Oil refinery fire, loss worth $49 million (2003 dollars), Roxana (IL), 1985. An oil refinery fire, in Norco (LA), 1988 did $513 million in damage. A Union Oil refinery fire kills 17 & loss worth $177 million (2003 dollars) at Romeoville (IL) in 1984. A Shamrock Oil & Gas Corp. refinery fire kills 19 firefighters at Sun Ray (TX) in 1956. A Phillips Petroleum plant fire, loss worth $1,113 million at Pasadena (TX) in 1989. A 1975 fire at the Gulf Oil Refinery Philadelphia killed 8 firefighters. Other large fires occured at the same refinery on May 16, 1975, and on October 20, 1975.

It would also appear that fatalities involving natural gas pipelines also accounted for far more casualties than nuclear power related accident. A single LPG pipeline explosion near Ufa in Russia killed up to 645 people on June 4, 1989.

Thus It would appear that the coal, hydroelectric, oil and natural gas sectors have accounted for a far higher death toll than the nuclear sector has.

What about property loss? In order to assess the cost of nuclear related accidents, relative to costs related to accidents in other sectors, we must have to make a comprehensive list sector accidents. Clearly coal mining and other coal sector related accidents might well involve greater property loss. Our lists of energy sector accidents would be extremely long, and involving literally thousands of accidents. Many accidents would involve the loss of human life, but thousands of accidents would meet Sovacool's property loss criteria. Since it is quite clear that Sovacool has failed to do include thousands of energy related accidents that would have meet his criteria in his research, no value can be ascribed to his work.

Sovocool has produced another typical example of his work. His research is weak, his research methods are suspects, and his conclusions will not withstand critical examination.

Benjamin K. Sovacool, “The Costs of Failure: A Preliminary Assessment of Major Energy Accidents, 1907 to 2007,” Energy Policy 36(5) (May, 2008), pp. 1802-1820.


Anonymous said...

By any rational criteria many coal plants, particularly grandfathered plants and plants in developing countries, are in continous and catastrophic failure mode.

Fine particulates, SO2 and other pollutants cause asthma, COPD and cardiovascular disease. By any estimate the number of premature deaths is horrendous.

If this is not to be regarded as an accident and not to be counted am I to regard it as an intentional sacrifice?

Charles Barton said...

Of course, the many deaths and sicknesses due to toxin pollutants associated with burning coal have long been known. These illnesses and deaths were due to a socio-political decision to let the victims of coal fired electricity subsidize the use of coal via a sacrifice of their health forced on them by society. This was done with the complicity of Ralph Nader, Helen Calticott, and the anti-nuclear Green movement which does not care in the slightest about human life or human health.

DV8 2XL said...

I note that Sovacool mentioned that he consulted for the Virginia Center for Coal and Energy.

Why is it every time I look into the past of some anti-nuke I find a connection to coal somewhere in there?

Anonymous said...

Wow! This "study" is just lazy -- there's no other way to put it.

These results were actually published in a journal?! The quality of scholarship these days has really gone down hill. It's sad.

Charles Barton said...

Sovacool is not much of a scholar.

Energy Policy appears to be a reputable journal. I cannot imagine that a Sovacool article of such poor quality slipped past Energy Policy readers without some awareness of its failings. I posted a negative comment on Scitizen a week ago, and Sovacool has not seen fit to respond yet.

Anonymous said...


In the spirit of open dialogue and debate, I am responding to your comments here (on Scitizen) and on your blog. But I am incredibly wary. While it is obvious you actually never read my study on energy accidents, that uncomfortable fact did not stop you from attacking it. For someone dedicated to “educating people to overcome irrational fears,” this is embarrassing bordering on disingenuous.

I also find it disconcerting that every time I have responded to your comments on my Scitizen column, you pretend that I haven’t and even state so on your blog. You’ve called me incompetent and lazy, but what you’re doing—attacking someone without even reading their work, and then ignoring their responses—is much worse.

Now, to the study at hand and your comments on Scitizen. You claim that the 1975 dam accident involved the failure of 62 dams, not 1. Had you read my study, you would already know that the failure of the Shimantan hydroelectric facility caused all of the other dams to fail. Shimantan was a Soviet-style hydroelectric facility constructed in the early 1950s on the Ru River. Engineers designed it to be part of a flood control and electrification scheme intended to reduce the incidence of severe flooding in the Huai River Basin and provide local villages with energy services. In early August 1975, Typhoon Nina dumped almost 8 inches of rain into the Basin in 24 hours, exceeding the yearly precipitation rate, collapsing buildings and destroying thousands of villages. Sedimentation clogged the sluice gates on many of the adjacent reservoirs and dams, worsening the problem, and telegraphs to open nearby dams failed to reach any of the facilities because of the storm.

While Chinese policymakers were debating whether to open dams upstream and downstream by air strike to relieve the water pressure on Shimantan, shortly after midnight on August 8 the dam failed to handle more than twice its capacity and released 1,670 million tons of water in just five hours, creating a massive tidal wave that cascaded into the failure of 61 other dams. Approximately 16 billion tons of water were released in total, resulting in a flood wave 6 miles wide and 23 feet high that traveled at nearly 30 miles per hour as it destroyed 4,600 square miles of property. Seven county seats were inundated, 6 million buildings collapsed, and 11 million people lost their homes. The Hydrology Department of Henan Province reports that 26,000 people died immediately and another 145,000 succumbed to fatal injuries during subsequent epidemics and famine.

Also, had you read the study, you would know I never claim that nuclear caused more deaths than hydro. I said that it involved more property losses, and for that number, you’ll need to see table A1 of the study, which lists every single accident investigated along with a short description, the amount of damage, and associated mortality/morbidity. It also shows every single one of the 63 nuclear plant accidents. I’d be happy to send you a copy—just email me—but I do find it deeply disturbing that you presume to know what the study says (or doesn’t say) without reading it.

And yes, the study does include coal mines—which you obviously would know if you had, once again, read the study—but it only includes those accidents that could be confirmed. I mention explicitly the Chinese deaths in the study, but I also mention how since they could not be verified by independent sources they had to be discounted. I spent a good part of the first half of the study talking about its methodology, especially how it will be very conservative (better to underestimate than to overestimate, I say).

You also claim I ignore data when it doesn’t fit into my “anti-nuclear agenda.” Had you actually read some of my other works (also published in peer reviewed journals), you would have discovered that I am not anti-nuclear. I have stated numerous times that nuclear plants are a superior alternative to coal and other fossil fueled plants. For instance, I said this in an article published in the Springer journal Policy Sciences:

While substantial parts of the fossil fuel waste stream are released into the environment (in the form of stack gases and particulates), the waste from nuclear power is seen as much easier to manage and control. Newer nuclear technologies appear to be much cheaper and safer than fossil fueled alternatives .. The impact on human health from fossil fuel combustion … is much more immediately drastic than nuclear power since nuclear waste becomes less toxic with time as radioactive materials decay, whereas the chemicals emitted from coal combustion often become quickly ingested by humans and other organisms. One study even found that the waste generated by a large nuclear plant per year was 2 million times smaller by weight and a billion times smaller by volume than wastes from a coal-burning plant.

The place we disagree is between nuclear and renewables/energy efficiency; but we both seem to be on the same side against fossil fuels (another reason I find it so odd that you are so threatened by me).

As for the claims about my study not on Scitizen but on your blog, almost all of the accidents you mention—the one in Nigeria, gas and oil fires—are included.

Ultimately, whenever I publish something in a journal or on a website, it is my responsibility to defend it. While I certainly don’t agree that personally attacking people is the best way to move forward, it is important that we have discussions like this. That said, Charles, you need to hold yourself to a higher standard. Attacking a study that you haven’t read is, simply put, shameful. You’re better than that.

We can certainly continue to have discussions in the future, but only if you actually read my work and if you are open and honest with your blog readers. Otherwise, I won’t bother wasting my time in the future.

Charles Barton said...

Benjamine, First I was responding to the account of your accident study which you posted on Scitizen. If I were going to study energy sector accidents, I would have certain methodological choices. First, I could go for a comprehensive data base. That might take years to compile, and probably would involve learning numerous foreign languages in order to do archival research. Clearly you did not seek to compile a comprehensive database. The second option would be to compile a represenative sample. That would also require considerable archival research. A third approach would be to limit the sample to say the United States. These approaches would yield valid statistical data. A fourth approach is to just find accounts of accidents and recover sufficient details to get a picture of what happened. This then would be a qualitative approach, because you collection method would not lead to valid statistical inferences. Your account of your research methods in Scitizen account would appear to be a of the last method. I must consider that method to be hap hazard at best.

Now you complain that I did not read your Energy Policy paper, and this is true. But you failed to provide a link to it. I looked for your paper n the internet and was unable to access it. Thus your Energy Policy paper is a black box. You read things out from it and tell us that they are true, but I have no way checking on anything. All I have to go on is your Scitizen account. My critique then is of what you say in the Scitizen account, which I tried as best I could to make sense of.

In your Scitizen "Oppenion" peice, you state:
From 1907 to 2007, a new study finds that 279 major energy accidents in the coal, oil, natural gas, hydroelectric, and nuclear sectors have been responsible for $41 billion in damages and 182,156 deaths.

Now exactly what does that mean? Between 1918 and the end of the Soviet Union energy related accidents went not publicly reported in the Soviet Union. Accidents, however were reported in archives. In order to get an accurate picture of energy related accidents you have to go beyond what is published in the newspapers. You have to do archival research. Clearly your method falls short of providing an accurate picture of energy related accidents. You claim to have found accounts of records of 279 major energy related accidents between 1907 and 2007. I'll wager that if you visited the city of Austin, Texas and conducted research in the Libraries of the University of Texas you would find published records for more than 279 major energy related accidents in Texas between 1907 and 2007. I'll wager that if you visited the state Library and Archives and archives of various state agencies, you will find records of dozens and perhaps hundreds of major energy sector accidents that went unreported by the print media.

You state, "The study only counted documented cases of accident and failure; It had to occur between August, 1907 and August, 2007; It had to be verified by a published source; . . ." Well not all of them, probably at best only a small and unrepresenative sample of them. Had you acknowledged that in your Scitizen account, I would have little grounds to question your findings.

Now it is quite improbable that you have even an exhaustive and inclusive list of documented energy related accidents in the United States. Energy related accidents, which would meet your criteria are quite common in Texas. I recall witnessing a pipeline fire in North Dallas some 15 years ago. The fire was quite impressive, and occurred very close to an office building which was undoubtedly damaged by the heat. $50,000 worth of damage was done, I am sure. In 2005 a gas tanker tanker wrecked, exploded and burned in Dallas a freeway was shut down for hours, and the cost of the cleanup would have undoubtedly exceeded $50,000, In 2007 another tanker wrecked, exploded and burned on I-35 north of Dallas. Again the Interstate was shut down for hours, and an expensive clean up occurred. There have been fires at filling stations. Were I to try to make an exhaustive list, I could go on and on for weeks or even months working on it.

There are also accidents due to natural gas leaks, Houses and buildings have been blown up, people killed, The New London School explosion of 1937 was caused by a natural gas leak, and killed at least 300 people, mostly children. Coal train accidents are common, An accident involving a coal train caused $1,000.000 in damages at Cactus, Texas in 2006. Thus I can only conclude that "major" accidents are far more common than the numbers reported in your study.

This video shows a major natural gas related accident west of Fort Worth:

This video is from Houston:

Your account of the 1975 Chinese dam disaster, ignores the fact that the Chines Authorities were warned that the larger Banqiao Dam was about to fail hours before the Shimantan Dam failed. Although the Shimantan Dam failure triggered the collapse of the Banqiao Dam, the real cause of the disaster, was a huge rain event, which the Chinese dams were not constructed to withstand.

Anonymous said...


All you had to do was ask me for a copy of the study, and I would have provided it (and still will if you email me). My contact information is listed quite openly in the very same bio you cut and pasted from the Scitizen website onto your blog. The article, also, is in no way a black box. As I am sure you know, most academic articles are not freely available on the internet (although I wish they were), but they are available in thousands of libraries. Simply travel to any local university and I bet their library subscribes to the journal Energy Policy. Plus, most scholars and analysts I’ve met are very happy to distribute their research. It only takes a request, and perhaps a little time.

Now, almost all of the concerns you mention in your most recent response are discussed, quite transparently, in the article. Heck, the title of the study even has the words “preliminary” in it for a reason. Naturally, I couldn’t just cut and paste the whole study onto Scitizen for length and copyright reasons, so I merely summarized it there.

The problem I have, one that I urge you to take very seriously, is when you attack the study (which you never read) by reading only a 1-page summary of it. I appreciate you making the clarification that you never read it now, but when you say “I cannot imagine that a Sovacool article of such poor quality slipped past Energy Policy readers without some awareness of its failings” it sure sounds like you did—and, being rhetorically gifted, I am sure you knew that then.

I’m always up for fair and transparent discussions of energy policy, but those discussions have to be open, honest, and rigorous. This is why publishing a peer reviewed article is so challenging. If you, Charles, are more careful in the future and tone down the personal attacks, I bet you could present your own ideas and research in the academic literature. Publishing a journal article, however, is much different than scribbling down some sentences for a blog. If you take the time to do it, though, I would be the first to congratulate you—and would also take the time and effort to continue the debate there.


Charles Barton said...

Dr. Sovacool posted an identical comment on Scitizen. I have responded to him there and see no reason for duplicating that comment here. Dr. Sovacool is being educated in the school of hard knocks.

Anonymous said...

Charles, posts on Scitizen, by their very nature, must be short and brief, typically no more than 1,000 words. They can never match the length or detail of other forums, including articles and blogs, and I don’t think it’s necessarily a good thing that they should. I view Scitizen as a place to post online what would have been an op-ed published in print. Op-eds have their own merits, but one of the most important is their brevity: to provide an important but always brief contribution to an ongoing discussion or debate.

In my Scitizen post about the Energy Policy study, I never once claimed (or even slightly gave the impression) that the op-ed was somehow equivalent to the study/article itself. In essence, the op-ed is entirely about the study. It summarizes its conclusions, talks about some of its methodology, and then provides the reference for those that want to learn more. It would be one thing if I wrote an op-ed making the claims made in the Scitizen post without referencing or referring to the study. But this op-ed let everyone know it was just a sample, and then directed them where to go for the real thing.

Moreover, talking specifically about the study’s estimate of 182,156 deaths, if anything this number is grossly underestimated. It’s incomplete by excluding accidents, meaning that an equivalent 9-11 probably happens much more frequently than 1.65 years. (Also, there is a brief literature review in the study as you suspected).

To your final concern, yes, scholars can and should use the internet to discuss the full merits of their research, but Scitizen cannot be that place, since it is limited to op-ed style pieces. In essence, we’ve kind of done that in our additional discussions—but much time and rancour could have been avoided if you read my study first, since it covered almost all of your concerns. Simply put, I ask that you take the Scitizen piece for what it is—and op-ed that talks about a study. There certainly must be a place for such things in both the internet and energy policy debates. If posting such a thing is not playing by your internet rules, I suppose I have to ask what those rules are, who set them, and what their value ultimately is.

Anonymous said...

Charles, no need to post this, just an FYI that I've placed a short reply to your last post on scitizen.

Anonymous said...

This engineer is not sufficiently familiar with Sovacool's work to participate fully, but on the subject of coal-related deaths, I return (in memory only) to an old Scientific American article authored by (former) prof. B. Cohen which relied on published data about coal-related power hazards.
Focusing on sulphuric emissions from US power plants, Cohen estimated back then that a single power station caused some 25 deaths and as many as 60000 cases of lung disease annually. (can't recall property damage) I think the coal plant size was 1000MW. To be sure, there have been pollution improvements since those days.
But if those epidemiological figures applied to the whole world (they probably would have, 30 years ago) then the cost in lives would have been some 60000-70,000 a year simply from sulphuric emissions from smokestacks.
From another angle, I also recall the late Sir Philip Baxter, distinguished Australian nuclear engineer, who once estimated that if the world relied on coal for all its energy, then in the final quarter of the 20th century, it would have cost 20,000 lives per year from mining deaths alone.


Blog Archive

Some neat videos

Nuclear Advocacy Webring
Ring Owner: Nuclear is Our Future Site: Nuclear is Our Future
Free Site Ring from Bravenet Free Site Ring from Bravenet Free Site Ring from Bravenet Free Site Ring from Bravenet Free Site Ring from Bravenet
Get Your Free Web Ring
Dr. Joe Bonometti speaking on thorium/LFTR technology at Georgia Tech David LeBlanc on LFTR/MSR technology Robert Hargraves on AIM High