Silencing the Bomb Read online

  A SOVIET ASSESSMENT OF THE THRESHOLD TEST BAN TREATY

  Roland Timerbaev of the Russian Ministry of Foreign Affairs, who attended the negotiations for the Threshold Treaty in 1974, wrote in 2006 about the history of the treaty. In a long footnote, he states that the yield of the U.S. Joint Verification Experiment (JVE) of 1988 in Nevada was 180 kilotons and significantly exceeded the 150-kiloton limit of the treaty. He quotes an interview on December 7, 2005, in which V. N. Mikhailov of the Russian Ministry of Atomic Energy stated that a collapse depression was formed by that JVE. Timerbaev states, “I might add that the Americans were very upset about this and requested that we not speak about the matter publicly; however, the story has since become public knowledge through the media.” Springer and others list the yield of the U.S. JVE, called Kearsarge, as 100–150 kilotons and do not mention a collapse feature.

  Timerbaev also gives information about how the yield threshold for the TTBT was debated in 1974, which I did not know. He states that the Soviet Ministry of Defense and the agency responsible for building and testing Soviet weapons (called at the time the Ministry of Medium Machine Building or Minsredmash) pushed to allow the USSR to undertake one or two tests [presumably per year] with yields over a megaton and three to four of 500 kilotons.

  Timerbaev states that it seemed to him during the [1974] negotiations that the threshold would be set in the range of a few hundred kilotons. He quotes a conversation in 1984 with a former American official reporting that in 1974 the U.S. military wanted to establish a threshold at 600 kilotons but Henry Kissinger, Nixon’s secretary of state, sharply objected. Kissinger and Gromyko agreed upon a 200-kiloton threshold, but Nixon, at Kissinger’s behest at the last minute, wanted a 150-kiloton limit. That number was agreed upon in the last few days of the negotiations in Moscow just before Nixon and Brezhnev completed and signed the treaty in early July 1974.

  11

  RENEWED INTEREST IN A CTBT, THE OTA REPORT, AND THE GROUP OF SCIENTIFIC EXPERTS: 1979–1996

  Much happened about nuclear testing during the fifteen years prior to the signing of the Comprehensive Nuclear Test Ban Treaty (CTBT) in September 1996. U.S. congressional committees expressed considerable interest in it. I wrote and testified extensively about the verification of a CTBT and the determination of yields of Soviet explosions. From 1986 to 1988 Congress’s Office of Technology Assessment (OTA) conducted the first independent review of the determination of Soviet yields and how well a CTBT could be verified.

  During my sabbatical leave from Columbia University from July 1981 to June 1982, I corresponded with Jack Evernden of the U.S. Geological Survey and Dennis Flanagan, the editor of Scientific American, about an article on the verification of a CTBT. Flanagan accepted a proposal by Evernden and me, and our article “The Verification of a Comprehensive Nuclear Test Ban,” appeared as the lead article in their October 1982 issue. At that time Scientific American and its publisher, Gerard Piel, were known for including articles on various aspects of the nuclear arms race. Evernden and I exchanged drafts of materials for the article by airmail in early 1982 when I was at Clare Hall College of Cambridge University and he was working for the U.S. Geological Survey.

  I find in rereading our 1982 article that it is still relevant today for educated, technical, and public policy audiences. It covers most of the issues related to the verification of a CTBT and came very close to forecasting the capabilities that now exist for using internal and external seismic networks to monitor the Soviet Union. Our subheading in 1982 stated, “Networks of seismic instruments could monitor a total test ban with high reliability. Even small clandestine explosions could be identified if extreme measures were taken to evade detection.” We commented that the issues to be resolved were political, which I think still is the case.

  Evernden, Cifuentes, and I followed up our 1982 paper the next year with a longer, more detailed analysis of monitoring the Soviet Union. We showed that its areas of thick salt deposits could be monitored with stations in the USSR down to yields smaller than one kiloton even if small evasive tests were contemplated using underground cavities.

  Our 1982 article helped to stimulate renewed interest in a full test ban throughout the 1980s and 1990s. In 1986 Evernden, Archambeau, and I received the Public Service Award from the Federation of American Scientists for “Leadership in Applying Seismology to the Banning of Nuclear Tests, Creative in Utilizing Their Science, Effective in Educating Their Nation, Fearless and Tenacious in Struggles within the Bureaucracy.”

  NEGATIVE VIEWS ABOUT MONITORING A CTBT BY BACHE AND ALEWINE OF DARPA

  Evernden and I invited Bache and Alewine of the Defense Advanced Research Projects Agency (DARPA) to give talks at a symposium titled “Verification of Nuclear Test Ban Treaties” which we organized for the American Geophysical Union in June 1983. This chapter discusses presentations at that symposium and conclusions in the OTA report of 1988 about verifying a full test ban. Determining yields of underground explosions and accusations of Soviet cheating on the Threshold Treaty were covered in the previous chapter.

  In an unpublished 1983 manuscript “Monitoring a Comprehensive Test Ban Treaty,” Bache and Alewine stated, “The conclusion is that serious technical problems remain to be solved to attain the capability to assert (with any confidence) that there is no clandestine program of a few events/year with yields up to a few tens of kilotons [my italics].” Bache criticized our Scientific American paper, stating, “In that article they state that the reliability of measures for the verification of a treaty banning explosions larger than one kiloton is no longer arguable. I assume that this must be an overstatement of their position, otherwise there wouldn’t be any point to this symposium.” In the 1980s Evernden and I claimed that two kilotons could be detected, much better than a few tens of kilotons.

  To my knowledge, the U.S. government has never issued an unclassified statement about what yields constitute tests of military significance. Its absence has led opponents of a CTBT to change their positions over time about the importance of tests of various yields, no matter how small. What is too small, of course, involves policy as well as technical considerations.

  Bache and Alewine are correct that a reliable regional method of identification is needed if the United States wants to identify events smaller than seismic magnitude mb 4.0. That is, in fact, why Evernden, Cifuentes, and I made computations in 1983 for a network that included fifteen seismic stations at regional distances in the USSR—that is, between 0 and 1250 miles (0–2000 km)—as well as more distant external stations.

  Bache and Alewine presented a table of magnitudes and yields for explosions at the Nevada Test Site in (1) competent water-saturated rocks and (2) dry porous materials like dry alluvium. In 1982 Evernden and I stated that the maximum thickness of dry alluvium in the USSR leads to an upper limit of two kilotons for that evasion scheme, as scientists in the military branch of the U.S. Geological Survey had stated years before. In addition, explosions in dry alluvium are likely to create craters at the earth’s surface, which can be detected by satellites. Bache and Alewine made no mention of these difficulties in associating magnitudes of 4.0 and 4.5 with yields of 15 to 90 kilotons in alluvium.

  Bache and Alewine also cited the difficulty of detecting and identifying seismic events in the few hours after a large earthquake. That scenario, which is called “hide-in-earthquake,” has been resolved for a long time. The two subfigures of figure 11.1 show seismograms made in Norway for the same time. Both show clear signals from a large earthquake in the eastern Soviet Union. The lower subfigure, however, was filtered to better display high-frequency seismic waves and to deemphasize low frequencies. Unlike the standard record at the top, it clearly shows a large arrival from a very small explosion of about 0.5 kiloton in Eastern Kazakhstan following the signal from the large earthquake.

  FIGURE 11.1

  Seismograms from Norway for the same day and time in 1979. Time increases to the right from 0 to 200 seconds. Upper figure is
a conventional recording of low-frequency seismic waves. Both figures show seismic waves starting at about 30 seconds from a large earthquake in the eastern Soviet Union. Seismic waves in the lower figure also include a large, impulsive P wave arriving just after 100 seconds from a very small explosion in Eastern Kazakhstan.

  Source: Office of Technology Assessment, 1988.

  What about testing in a large nearby earthquake? Since earthquakes cannot be predicted except for time periods of a few decades, a potential violator would have to wait years to decades to set off an explosion in a large nearby earthquake. The violator would need to have a nuclear explosion ready to detonate at a moment’s notice and to determine the location and size of the earthquake accurately within a few minutes. “Hide-in-earthquake” is not a realistic evasion scenario even for an explosion of a fraction of a kiloton.

  The remarks by Bache and Alewine in 1983 and previously indicate that they and their office in DARPA did not take problem-solving approaches to test ban verification. Nevertheless, DARPA’s mission going back to 1960 was to improve nuclear verification. DARPA does deserve credit for subsequently funding programs to analyze large volumes of seismic data and for promoting the installation of small seismic arrays.

  SEISMIC MONITORING BY U.S. SCIENTISTS IN THE SOVIET UNION

  In 2013 Frank von Hippel stated that he personally showed Figure 11.1, illustrating the detection of a very small Soviet explosion, to General Secretary Gorbachev in July 1986. Elected to that position in March 1985, Gorbachev and several of his more liberal scientific and technical advisers, such as Evgeny Velikhov, wanted to end the nuclear arms race. Gorbachev’s first arms control initiative, on July 30, 1985, was to declare a unilateral moratorium on nuclear weapons tests for the remainder of that year—to be extended indefinitely if the United States reciprocated. Von Hippel states, “Reagan administration spokesmen argued that the Soviet Union had chosen to stop testing only after deploying a whole new generation of warheads. By contrast, they said, the US needed to test new warheads for the MX and Trident II missiles and for Edward Teller’s antimissile x-ray lasers…. Moreover, they contended, a test moratorium could not be verified.”

  Von Hippel says that he met Velikhov at a conference in October 1985, and “Velikhov suggested that since the U.S. government was not interested in a mutually verified test moratorium, perhaps some nongovernmental organization might be interested in verifying that the Soviets were not testing, even at low yields.” The Natural Resources Defense Council (NRDC), a private environmental group in the United States, was interested.

  In May 1986, NRDC signed an agreement with the Soviet Academy of Sciences to deploy three seismic stations near the nuclear test site in Eastern Kazakhstan. The stations, which were to be operated jointly and the data made available publicly, operated successfully for about a year. I was on NRDC’s advisory committee from 1986 to 1988. Archambeau and several seismologists from UC San Diego set up and maintained those stations. NRDC also helped to set up three similar seismic stations around the Nevada Test Site, which involved U.S. and Soviet scientists. The stations gave important seismic information on tests set off by the other country. Most important, they helped to move the CTBT debate forward.

  CONGRESS’S OFFICE OF TECHNOLOGY ASSESSMENT (OTA) ON VERIFICATION OF A FULL TEST BAN

  The 1988 OTA report Seismic Verification of Nuclear Testing Treaties was the first serious analysis of test ban issues in the United States outside of the departments of Defense and Energy. It stated, “Verification—the process of confirming compliance and detecting violations if they occur—is therefore central to the value of any such treaty. Yet in the arena of arms control, the process of verification is political as well as technical. It is political because the degree of verification needed is based upon one’s perception of the benefits of a treaty compared with one’s perception of its disadvantages and the likelihood of violations.” It involves an assessment of what is an acceptable level of risk and a decision as to what should constitute significant noncompliance. “Consequently, people with differing perspectives on the role of nuclear weapons in national security and on the motivations of Soviet leadership will differ on the level of verification required.”

  About half of the 1988 OTA report was devoted to monitoring the Soviet Union under either a full nuclear test ban or a low-yield threshold treaty. It stated, “A hypothetical seismic network with stations only outside the Soviet Union would be capable of detecting well-coupled explosions with yields below 1 kt anywhere within the Soviet Union and would be able to detect even smaller events in selected regions.” It went on, “From a monitoring standpoint, stations within the Soviet Union are important [more] for improving identification capabilities than for further reduction of the already low detection threshold.” The report cited the recording of high-frequency seismic waves by the three NRDC stations in Kazakhstan, but OTA had few other data on capabilities for monitoring within either Russia or China.

  The OTA report was published just prior to the breakup of the Soviet Union. Many earthquakes once in the former Soviet Union, especially those in areas of thick salt deposits, are now located in independent countries. This considerably reduces the number of small earthquakes and potential nuclear explosions by the Russian Republic that must be monitored.

  In a section titled “How Low Can We Go,” the 1988 OTA report discussed various Soviet yield levels and how well they could be monitored.

  Level 1—yields above 10 kt. Nuclear tests can be monitored with high confidence with external seismic networks and other national technical means [satellite images and other intelligence]. No method of evading a seismic network is credible.

  Level 2—below 10 kt but above 1–2 kt. Demonstrating a capability to defeat credible evasion attempts would require seismic stations throughout the Soviet Union (especially in areas of thick salt deposits) and provisions in the treaty to handle chemical explosions. Expert opinion about the lowest yields that could reliably be monitored ranges from 1 kt to 10 kt.

  Level 3—below 1–2 kt. The burden on the monitoring country would be much greater.

  Level 4—comprehensive test ban. There will always be some threshold below which seismic monitoring cannot be accomplished with high certainty. The OTA report states that a comprehensive test ban treaty could still be considered adequately verifiable if it were determined that the advantages of such a treaty would outweigh the significance of any undetected testing below the monitoring threshold.

  Experts from the weapons labs and the Department of Defense (DoD) were involved in the four panels convened by OTA from 1986 to 1988. It was not clear until the last minute if the DoD would permit Robert Zavadil, chief of the Evaluation Division, Directorate of Geophysics of AFTAC, to brief the OTA panels at the Secret level about classified network capabilities, identification, and yield determination. He briefed the OTA panels in an informative and straightforward way. Nevertheless, some officials of DoD (not Zavadil) criticized the OTA report soon after it was published in 1988.

  OTA went on to conduct a separate study, The Containment of Underground Nuclear Explosions, which was published in 1989. Gregory van der Vink was the project director. I discuss containment, which involves assuring that radioactive products are not released into the atmosphere, in later chapters on international monitoring and evasion.

  LETTER FROM BACHE TO ME ABOUT HIGH-FREQUENCY SEISMIC MONITORING

  The American Geophysical Union asked me to write a six-page review on nuclear testing for the U.S. National Report to the International Union of Geodesy and Geophysics, which is published every four years. My paper “Underground Nuclear Explosions: Verifying Limits on Underground Testing, Yield Estimates, and Public Policy” was printed in Reviews of Geophysics in 1987.

  Bache soon wrote to me (figure 11.2) concerning my statement that the paper “An Evaluation of Seismic Decoupling and Underground Nuclear Test Monitoring Using High-Frequency Seismic Data,” by Evernden, Archambeau, and Edward Cranswick of t
he U.S. Geological Survey, probably was the outstanding article of the last four years published by U.S. scientists on test ban verification. I had made a general statement about the potential use of high-frequency seismic waves at regional distances in the Soviet Union. The paper by Evernden and colleagues represented an effort to move forward on the monitoring of explosions smaller than one to two kilotons, especially ones that might be detonated in an evasive manner. I think it was the start of an important process, not the end. The monitoring of high-frequency waves by seismic stations at distances up to 1250 miles (2000 km) within the Russian Republic, China, and surrounding countries became one of the outstanding contributions to nuclear verification in the following twenty-five years.

  FIGURE 11.2

  Letter from Thomas Bache, 1987.

  In his letter, Bache asked me if I had read their paper. Of course I had; undoubtedly he was being facetious. He went on to say that “this article goes beyond any reasonable bounds—and, of course, many of the seismological ‘conclusions’ are absurd.” Then he stated, “Rather, I assume it reflects your opinion that the article is politically correct. That is a sort of a ‘red guard’ approach to science, isn’t it?…But you can see that it goes beyond bounds I am willing to tolerate quietly.” In a note attached to his letter, Bache wrote, “Some things are beyond endurance….”