Which Way Security?
I read your article ”What About the Nukes?” [March] with great interest. I have worked on and/or managed our nation’s nuclear weapons stockpile evaluation program from June 1968 until I retired from the federal government in January 2006. From 1989 to 2002 I was the Department of Energy’s (DOE) program manager for weapons evaluation. From summer 2002 through 2005 I was the director of the Weapons Quality and Surveillance Division, with responsibility for oversight of how the surveillance program was meeting DOE testing requirements. The purpose of my comments is not to take a position on the need for a reliable replacement warhead; it is to point out a number of errors in the article.
Unfortunately, the conclusions drawn in the article are predicated on this incorrect information. Let me start by stating that the Stockpile Stewardship Program doesn’t inspect 20-plus samples of each type of warhead in the active stockpile every year. When I retired, the sampling plan used was between 10 and 11 randomly selected samples of each type of warhead in the active stockpile. Of these 10 to 11 samples, only one sample of the nuclear components was selected for testing. Starting in late 2002, it became evident by the actions of top senior management at the National Nuclear Security Administration (NNSA) that they no longer considered surveillance to be a top priority. Because of this lack of assigned priority and the lack of an adequate budget for the surveillance test program by the highest levels of management in the DOE and the NNSA, starting in late 2002, larger and larger backlogs in sample-system and component testing started to develop. NNSA’s solution to this problem was to administratively eliminate large segments of the backlog and to reduce future testing requirements to avoid being confronted with new backlogs later on.
The bathtub curve pictured in the article illustrates the theory that surveillance managers used to develop their sampling philosophy from the inception of this program, over 50 years ago. We always concentrated very large numbers of samples in the early stages of production to find both birth defects and design problems. This was done so that corrective action could be taken before an entire stockpile of a new system was fabricated. We then had our steady-state sample of 10 to 11 samples for each system for the remainder of the weapon’s life. At that point, we were at the bottom of the bathtub curve, where we looked for failures affecting only a segment of the weapon system population or aging problems occurring during the weapon’s planned original design life. It is only in the more recent past that we have gotten to the point where we might be experiencing ”end-of-life wear out.” One might think that it would be important to increase sampling at this stage, as we do in the early stages of production. But as a result of decisions made by top management in 2007, the sampling rate has been reduced to less than what was used when a weapon was in the field for the 20 years of life referred to in your article.
Now let me go to the chart ”Old but Okay.” This chart is misleading at best. We must remember that each weapon system in the stockpile has components designed by Sandia National Laboratories and either Los Alamos or Lawrence Livermore National Laboratory. The chart addresses only the nuclear components designed by Los Alamos. What about the components designed by Sandia? You must also understand the sampling philosophy used by NNSA. The 11 samples for a specific system selected each year are randomly selected from that system’s stockpile. So, hypothetically, if a system was currently 25 years old and had a 10-year production run, when the 11 samples are selected you might expect to have only one of those samples that is 25 years old. One would be 24 years old, another 23 years old, and so on. Of the many thousands of samples tested over the past 50-plus years, only an infinitesimal number of samples were over 25 years old when tested. The vast majority of the stockpile never reached the age of 25. So the amount of data that exist on older weapons that have seen the effects of 25 years of aging is extremely limited. Also, remember that NNSA has reduced the established sample size from 11 samples per system per year to around 5 or 6 samples per system per year. This also reduces the number of Significant Finding Investigations (SFIs) that would be opened. If in the future NNSA was to reduce the sampling to zero, we could expect to have no SFIs at all. This does not mean that no problems exist in the stockpile; it only means we have not found problems that might exist. From over 37 years of working on and/or managing the surveillance test program, I believe the realistic assumption to be made is that we will not detect existing problems.
We are told by the nuclear laboratories that the modeling used in the Stockpile Stewardship Program can successfully allow for reductions in sampling rates. Unfortunately, history proves that to be wrong. Modeling can be used only to project defects when a specific failure mechanism is understood. During the history of the stockpile evaluation program, most failures, especially the more serious ones, would have never been detected through modeling. There was never the slightest expectation that the specific failure mechanism even existed. It was only through actual testing that a failure was detected and corrective action could be taken. In the absence of adequate testing, a failure would still exist and be undetected and uncorrected.
Now why do the above problems exist? From the inception of the surveillance program in the 1950s until 2002, all levels of management within the Atomic Energy Commission, the Energy Research and Development Agency, the DOE, and the Albuquerque Operations Office of the DOE believed that it was necessary to have an organization responsible for surveillance that would be independent and separate from the organization responsible for the production of warheads. Production and surveillance competed for budgetary resources and priorities at the labs and plants mentioned in the article. With this independence came advocacy for the program. In July 2002, NNSA’s top management decided that the independence was no longer necessary. Within a matter of months, surveillance went from a top priority in NNSA and the DOE to something that had virtually no priority whatsoever—hence the backlogs previously discussed, the reductions in sample size, and no increase in sampling based on an aging stockpile. All of this leads me to ask, What is the true reliability of our nuclear arsenal? Is it as safe as we profess? Do we need a reliable replacement warhead?
History has shown that with the development of each new nuclear weapon system, there is a need for extensive testing to find both design problems that exist into production as well as production defects. Under the current NNSA management philosophy of surveillance, I would be deeply concerned that a new reliable replacement warhead might not be as safe or reliable as reported.
Ralph Levine, IEEE Member
Authors Francis Slakey and Benn Tannenbaum respond: We agree with Levine that the surveillance program should be enhanced—that’s why we recommend the imbedding of microsensors to ease the reliance on dismantlement. Sandia is responsible for nonnuclear components and is therefore not included in the data used in the article.
A statistically significant number of warheads are sampled, as described in detail by the National Academies of Science: ”In order to meet the stated goal of affirming, at 2-year intervals, a 90 percent level of confidence that if 10 percent or more of the warheads of any given type contained a flaw this would be detected, 21 warheads of each type with a population under 500 and 22 warheads of each type with a population of 500 or more are withdrawn randomly from the stockpile every 2 years for examination. These weapons are disassembled and the parts inspected, including the nuclear components (one per weapons type is destructively disassembled each year).”
Levine says, ”We are told…that the modeling used in the Stockpile Stewardship Program can successfully allow for reductions in sampling rates.” Yes, we are. As part of the Annual Certification process, the three weapons lab directors validate that a statistically significant number of warheads are sampled in a formal letter sent to the secretaries of Energy and Defense, who in turn send a memorandum to the president. We do not wish to engage in a debate over whether the lab directors have lied, exaggerated, or ignored findings in the certification process, and we encourage Levine to direct this concern to the Department of Energy, Inspector General.
The United States has the worst record of all countries regarding nuclear weapons. We have enough to eliminate all life on Earth!
Here are my suggestions to reduce the risk of disaster: 1) restrict our president’s authority by changing our published policy to ”no first use”; 2) follow the advice of Reagan and Gorbachev by dismantling most of our nuclear weapons; and 3) use the fissile materials in reactors to generate electric power. This will destroy the dangerous materials, reduce coal burning, and reduce global warming.
Here is a quotation from Albert Einstein: ”A world government with powers adequate to guarantee security is not a remote ideal for the distant future. It is an urgent necessity if our civilization is to survive.”
Robert J. Rorden, IEEE Life Member
Los Altos, Calif.