Reflection #2

Ratifying the Comprehensive Test Ban Treaty (CTBT)

The Senate’s vote on October 19, 1999, to deny ratification of the CTBT was a national embarrassment. I think the treaty should be ratified as soon as possible and that we should continue to elect Presidents who will support the treaty and abide by it -- even without ratification by the U.S. Senate.

I trust the Ploughshares Fund, a well-regarded nuclear disarmament and non-proliferation nonprofit organization and grantmaker that supports the CTBT. The following is the response by Naila Bolus and Paul Carroll, of Ploughshares, to questions that I raised with them (based on emails I had exchanged with former U.S. Representative Tom Campbell):

We have come up with the following responses to the various points, questions, issues raised. Also I want to note that we refer primarily to those organizations we fund (as well as our scientific advisors) -- they are the experts on the technical aspects of the treaty and the testing issue generally.

I. Safety and Reliability of the U.S. Nuclear Arsenal

Much of the debate and concern about the CTBT centers on the nuclear arsenal "turning to green cheese" over time if we do not constantly update, improve, and test it. These issues are complex, but there are some key points to be made.

1) The definitions of "safety" and "reliability" are very different and distinct, despite the fact that the Department of Energy (DOE) tends to use them together ("safe and reliable"). There is a very technical definition for safety, having to do with the probability of a warhead detonating with a nuclear yield if subjected to certain adverse conditions. Basically, a "safe" warhead is one that will not explode even if subjected to fire, crash, theft, etc. The standard is high, on the order of less than one chance in a million. All current U.S. warheads are certified as "safe" and the passage of time will have little to do with any changes to safety. Safety is a function of the design of the weapon and its integration into the overall weapon system. In fact, DOE’s own data show that "no age-related safety issue has ever occurred in the nuclear component of a nuclear weapon." (This is from a report about the Stockpile Stewardship program by Ploughshares Fund grantee Tri-Valley CAREs of Livermore, California.)

"Reliability" on the other hand, is the assurance that a warhead will explode within a relatively close range to its design yield. In other words, if a warhead is designed and intended to explode with a yield of 300 kilotons (roughly 20 Hiroshimas, a typical U.S. warhead) and it explodes with only 270 kilotons, it is deemed unreliable (only a 90% yield performance). Reliability, it is feared, will deteriorate over time as materials decay and warheads age.

These fears have slightly more merit than those about safety. But "aging" is rarely a significant problem with respect to reliability, either. (A report by another Ploughshares Fund grantee, the Institute for Energy and Environmental Research (IEER) called "Nuclear Safety Smokescreen" found that "Most reliability problems have a minor effect: almost seventy percent [of defects] reduced reliability by only one percent or less. The majority of reliability problems do not need corrective action or can be solved through production changes.")

Thus, the question for reliability is really one of "how reliable is reliable enough?" Congressman Campbell raises this issue in his correspondence of Monday, November 15, when he cites Dr. Robert Barker discussing the issue. Barker talks about those who would argue that "it doesn't matter if the reliability is only 65 percent." Based on the historical record, those few reliability problems that were discovered after the production of a warhead model, it is clear that negative impacts to reliability were far less than this arbitrary number. Furthermore, they had nothing to do with aging.

2) The vast majority of U.S. nuclear tests over the last 50 years were carried out for design proof testing and not for testing safety once a warhead model was in production. Using the automobile analogy that many, including Assistant Secretary for Defense Programs at DOE, Vic Reis, use, the nuclear tests are done on prototypes to test design concepts and performance. The designers use the information to refine the weapons. Once a design is "in production" it has been tested upside down and sideways to ensure that it is both safe and reliable. As thorough study of DOE data has shown in a study by the IEER, those warhead problems over the past decades that were severe enough to require "major modifications" were inherent in the design of the warhead, NOT due to aging. Thus, the "lemon" analogy falls down since it is the design of the car that went into production that was at fault, not time past since the consumer bought the car. The point here is that we need to be clear on what all those tests were actually carried out to do. Very few were for "checking" on the reliability or safety of warheads already produced and deployed. Most were for proof of design concept and actual nuclear weapons effects testing for military planning purposes.

II. Verification

There have been many good points made and questions asked by Bill Krause, Ed McIntosh, and Congressman Campbell about the verification regime in the current CTBT.

It is important to understand some basic facts about developing nuclear weapons and who the CTBT is aimed at. Bill Krause points out that "it is absolutely possible to build nuclear weapons and test them without detection consistently." We would not agree with this completely. There is an important distinction to be made between fledgling nuclear aspirants, and established nuclear powers.

The United States, in 1945, never tested the type of bomb it dropped on Hiroshima. This was a "gun type" atomic weapon that used two parts of uranium-235 and drove them together to produce the blast. The scientists and engineers knew it would work based on solid understanding of the relatively straightforward physics, and the test pile at the University of Chicago. We did test the plutonium bomb, like "Fat Man," that was used on Nagasaki because it was a more complex implosion device. This first-ever nuclear explosion at Trinity in July of 1945 had about a 12 kiloton yield.

We need to be clear about what the CTBT can and cannot do, and what it can achieve even with limitations. There are today at least 9 assured nuclear states. The traditional 5 (U.S., Russia, China, France, and Great Britain) are mature and have long experience with nuclear weapons. We would argue that for these states, the CTBT would act as an arresting hook to stop "vertical proliferation" -- the continued improvement and advancement of nuclear weapons technology.

For the others (India, Pakistan, Israel, South Africa), and nuclear aspirants, their state of maturity is far less. The tests in May of 1998 in India and Pakistan were readily detected and identified by international seismic means not part of any proliferation regime. Like the U.S. experience, the "infancy" stage tests were around the 10 kiloton range.

Back to Bill Krause’s assertion (through Gen. Goodpastor) that it is possible to hide nuclear tests consistently. Yes, a mature nuclear state with the capability to "decouple" a nuclear test and conduct tests in the low-yield range around 1 kiloton could escape detection. But why would they want to? Russia (or the U.S., or China) would have nothing to gain from carrying out clandestine tests in the 1 kiloton yield range.

There is an excellent report containing information regarding seismic testing and the CTBT verification regime written by Gregory van der Vink, an expert on seismic verification issues. He notes that in the 1980s "the range for the verification limit was between 1 and 10 kilotons [Sykes, 1982; Hannon, 1985; Evernden et al., 1986], with a general conservative consensus of around 5 kilotons [U.S. Congress, 1988]." "For an explosion of ten kilotons or more, decoupling would not be a credible evasion scenario." He continues: Without decoupling, a 1-kiloton nuclear explosion creates a seismic signal of M 4.0. There are about 7,500 seismic events worldwide each year with M 4.0 (Figure 1). At this magnitude, all such events in continental regions could be detected and identified with current or planned networks. If, however, a country were able to decouple successfully a 1-kiloton explosion in a large underground cavity, the muffled seismic signal generated by the explosion might be equivalent to 0.015 kilotons and have a seismic magnitude of M 2.5.

A threshold nuclear state would have to make several leaps beyond what every other nuclear weapons state went through to succeed in such evasion. Not a likely scenario. Furthermore, in today's world, it is far more likely that such a state or group would simply seek to buy a weapon rather than undertake the formidable enterprise of producing fissile materials, designing a weapon, excavating a test site to "decouple" the explosion, and test the bomb. It should be noted that these other steps would likely be detected by other means such as satellites, radiological detection, and even human intelligence. There is a good discussion of the range of the verification system under the CTBT in the September 27, 1999 Brief Issue on the Coalition to Reduce Nuclear Dangers site.

It would be remarkable indeed if a brand new or "threshold" nuclear state or actor were to kickoff its nuclear birth with a 1-kiloton range test. These are the levels of tests that might easily "fall below the radar" of seismic verification. But these are only tests that mature nuclear states could carry out. Furthermore, tests in this range are likely to yield only information of use to advanced, thermonuclear capable states. Ray Kidder of Lawrence Livermore National Laboratory is an authority on nuclear testing, safety and reliability, and what various tests and their magnitudes can and cannot do. We defer to him for a fuller discussion on who can gain what knowledge from what kinds of test, let alone actually carry them out. (Also see the article by Hans Bethe in the New York Review of books; it’s on Ploughshares Fund’s web site now.)

III. Other Issues

Ed McIntosh stated that "we may be out to do a lot more nuclear testing in the near future" as part of the potential decision to deploy a National Missile Defense (NMD) system. This is inaccurate. While the far-flung Star Wars program of the 1980s foresaw a number of different kill mechanisms that would destroy incoming missiles in the boost phase and post-boost phase, and re-entry phase, the currently planned NMD is strictly a "kinetic kill" system. Star Wars did have provisions for using nukes to kill nukes. NMD is a more pure "hit a bullet with a bullet" system, and the defending bullet is decidedly non-nuclear.

Bill Krause writes: "another severe limitation has to do with the fact that the current CTBT is forever binding with no outs." This is not accurate. The United States demanded several "safeguards" as part of its treaty ratification position. Most of these have to do with the DOE’s massive Stockpile Stewardship program, another whole topic of conversation. However, these so-called "safeguards" and the treaty itself allow for parties to withdraw under a clause of "supreme national interest." The safeguards can be found on the Coalition to Reduce Nuclear Dangers site.

As to your point about testing "detonators," there are three distinct parts, to oversimplify, of modern thermonuclear weapons:
1) a chemical high explosive that implodes the plutonium "primary," also called a "trigger"
2) the primary that results in a fission explosion and sometimes a "boosted fission" explosion and
3) the fusion "secondary" that gives today’s weapons most of their explosive power.

A note on "triggers" and deterioration. While it is true that we don’t know with certainty what will happen to warheads say, 100 years from now, we do know a lot about the special nuclear materials in them. Plutonium, the key element in the "trigger" or primary, has a half-life of 24,400 years. Doubtful much will happen in the way of deterioration over even 100 years. The secondary, or the "H" part of today’s H-bombs, is mostly uranium-238 with a half life of about 400 million years. Also, as the IEER report points out, the nuclear components are the most solid state and robust parts of today’s weapons. The non-nuclear components are more likely to experience age-related problems. These components can be fully tested in a non-nuclear manner, and in fact are tested under the DOE’s Stockpile Surveillance and Maintenance Program.

A final word. As you (and we have argued), with any arms control treaty the ultimate measure of its worthiness is a cost/benefit one. Will the CTBT’s benefits outweigh the costs from possible cheating? We would argue yes, since the verification regime can continue to improve, since cheaters are extremely likely to be "caught," and since the CTBT does have powerful symbolic impact as well as real world affect in arresting the advancement of nuclear weapons technology and the proliferation of more rudimentary capabilities. No treaty will ever deter a party that is un-deterrable, just as no amount of nuclear weapons will prevent a terrorist attack. This is an unrealistic standard to set. But the international norms that can be achieved by the CTBT can go a great distance toward global stability in the nuclear weapons realm.

Q: Can you test the detonators without doing a nuclear test? e.g., to find out whether the electronics still functions? Campbell’s argument is that if we don’t know if the ICs still work, the weapon may not fire. And we can’t just replace the electronics in 10 years because the components of the original may not be manufactured anymore. So is it possible to test a detonator without a nuclear explosion?

As I understand your question, the answer is yes. Our grantees have pointed out that under the Stockpile Surveillance program warheads of each type are annually withdrawn from the stockpile, disassembled and inspected - visual inspection for signs of corrosion or deterioration of materials, plus electronic and mechanical checking of components. We can use non-nuclear testing methods even for the high explosives next to the pits. DOE documents show that the majority of aging effects occur in the testable and more easily replaced non-nuclear components.

Also, test ban opponents claim that the original materials in warheads, such as some plastics and adhesives, might not be available in the future. While this might be a minor problem in the future, our grantees and experts (such as Ray Kidder) argue that in reality replacement materials need not be exactly the same as the originals; they need only to perform the same functions, within established tolerances for error. That performance can be evaluated under stockpile surveillance and occasionally full-scale dynamic testing short of nuclear detonation.