Abstract
Careful study of the consequences of spill fires can settle terminal-siting questions.
In the July/August 2003 Bulletin (Terrorism: Ready to Blow?”), I wrote about concerns surrounding the shipment of liquefied natural gas (LNG). Of all the potential hazards associated with shipping LNG, I believe that a massive spill from a ship, followed by a fire on the surface of the water, is the most likely worst-case scenario. Given the scope of today's terrorist threat, the resurgent interest in LNG in the United States, and the lack of any statutory requirement for consideration of ship-side releases in siting LNG terminals, the possibility of such an event needs to be reassessed.
There is more interest now in building LNG terminals than there has been in the past 25 years. Currently there are only four LNG import terminals in the United States, all built in the 1970s. This is about to change. Shell, Exxon-Mobil, and ChevronTexaco plan to build import terminals along the Gulf Coast. BHP Billiton hopes to supply 15 percent of California's annual natural gas demands with an LNG terminal off the California coast. San Diego-based Sempra is building the first new U.S. terminal in more than two decades on Louisiana's Calcasieu River. At least five terminals are being considered for the East Coast. And last summer, Mexico approved three new LNG terminals, in large part to supply the needs of southern California. Can there be any question that now is a good time to reconsider where and how LNG import terminals are located?
While my previous article in the Bulletin attempted to characterize the hazards associated with LNG operations, I did not present the type of analysis that would support an extreme position. I wanted to avoid having my views used to support alarmist agendas, but this is exactly what some have done with what I wrote, which I would like to clarify now.
The question in dispute is this: If the contents of a single LNG ship tank were released onto water and ignited, how big would the fire be, and how far from the fire would people need to be to be safe?
In the late 1970s and early 1980s, the coast guard (and later the Transportation Department) sponsored tests conducted at the U.S. Naval Weapons Center in China Lake, California, to provide some data for answering this question. A series of spills of LNG on water were ignited and the fire sizes and heat radiation characteristics were measured. The largest fire in the experiment was from a spill of approximately 10,000 gallons, and the fire exceeded 50 feet in diameter and 250 feet in height.
A typical LNG tanker carries up to five separate, effectively double-walled cargo tanks, each holding about six million gallons of LNG. These ships have enviable safety records, and advocates of LNG shipping point to the difficulty–and low probability–of accidentally or intentionally breaching a tank. While I accept this position, the attacks on the U.S.S. Cole and the French tanker Limburg demonstrate that such assurances may not deter terrorists from trying.
Moreover, as I pointed out in my first Bulletin article, it should not be assumed that the spill would be limited to a single tank that was initially ruptured. The fire from a single tank spill could engulf the tanker, raising serious questions about the vulnerability of the ship to further releases, which would increase the severity.
From a public safety perspective, how close is too close when it comes to siting LNG terminals? Answering this question needs a science-based answer, and I contend that an answer has already been developed sufficiently for the public decision-making process to move forward.
The answer relies on one of the most extensively studied subjects of science–scaling, or scale modeling. In the case of LNG the question becomes, “Based on what we know of a 10,000 gallon spill fire, how would a six million-gallon spill fire behave?” To take the extreme position, one might say that it would be 600 times as great, with a six-mile diameter to use the example cited above–but there is no scientific basis for such an opinion. Some straightforward observations, however, are informative.
The diameter of such a fire, if the fire were a sphere, would be proportional to the one-third power of its volume; hence the larger fire would be approximately 450 feet in diameter. But a fire is not a sphere, but more like a cylinder with its base sitting on the water. Further consideration is not appropriate here, but the principal idea is apparent: The size of a fire, as it relates to spill size, does not increase linearly.
I have reviewed the work performed in the last 25 years on this question. With few exceptions, the scientific consensus on the scope of an LNG-on-water spill fire involving an entire tank of LNG is that it would be at least a half-mile in diameter. Scientists also seem to agree that from the edge of the fire to about another half-mile out, people would receive second-degree burns on unprotected skin within about 30 seconds. Obviously, larger fires would result from larger spills.
Some industry advocates have suggested much shorter distances would be safe, apparently relying on information provided by the government. Press reports on a proposed terminal in Mobile, Alabama, state that ExxonMobil cited the coast guard study mentioned above as the basis for much shorter distances. However, my application of that report's guidelines gives distances that are consistent with the larger estimates.
Both the Energy Department and the Federal Energy Regulatory Commission have cited a study conducted by Quest Consultants, an Oklahoma company, which gives shorter distances, and which has apparently been used to ease public concern about terrorist-caused LNG spill fires. It is my understanding that the study–which was originally prepared for use in considering readmitting LNG tankers into Boston following a stoppage ordered after 9/11, and was subsequently used in support of the reopening of the Cove Point terminal this year in the Chesapeake Bay–has been used to support requests for approval of proposed LNG terminal sites. The study does appear in the environmental impact statement submitted with the application for approval of the Hackberry terminal on the Calcasieu River in Louisiana. However, the methods used in the study do not follow the permit requirements for siting a terminal (49 CFR 193). Even the Quest study's author, John Cornwell, acknowledged in the press last October that his company's study, produced at Energy's request, is being misused by federal officials.
I believe that all that is required here is to submit the question to credible scientific review and to proceed accordingly. I also believe there is a greater issue here. Science-based study offers our only hope of dealing rationally with questions like this, which abound, pertaining to subjects from weapons of mass destruction to global warming. If the public comes to regard cynically the arguments presented by credible scientific experts, we will all be the losers.