Edited by William Burr*
Washington, DC, September 22, 2009 - Sixty years ago this week, on 23 September 1949, President Harry Truman made headlines
when he announced that the Soviet Union had secretly tested a nuclear weapon
several weeks earlier. Truman did not explain how the United States had
detected the test, which had occurred on 29 August 1949 at Semipalatinsk, a
site in northeastern Kazakhstan. Using
declassified material, much of which has never been published, this briefing
book documents how the U.S. Air Force, the Atomic Energy Commission, and U.S.
scientific intelligence worked together to detect a nuclear test that
intelligence analysts, still unaware of the extent to which the Soviets had
penetrated the Manhattan Project, did not expect so soon.
Stalin and the Soviet Politburo were probably stunned by Truman's announcement; they did not know that Washington had a surveillance system for detecting the tell-tale signs of a nuclear test and they wanted secrecy to avoid giving the United States an incentive to accelerate its nuclear weapons activities. (Note 1) Joe-1 (as U.S. intelligence designated it) was also a jolt for U.S. intelligence analysis, which for several years had asserted that the Soviets were unlikely to have the bomb before mid-1953, although mid-1950 was also possible. A few weeks after the test, CIA director Roscoe Hillenkoetter argued that "I don't think we were taken by surprise" because of an error of only a "few months," but not all of his Congressional masters accepted that.
How did the Truman administration discover Moscow’s secret?
Shortly after the Soviet test, on 3 September 1949, a WB-29 ["W" for weather reconnaissance] operated by the Air Force's Weather Service undertook a routine flight from Misawa Air Force Base (Japan) to Eilson Air Force Base (Alaska) on behalf of the secretive Air Force Office of Atomic Energy-1 [AFOAT-1] [later renamed the Air Force Technical Applications Center, or AFTAC]. The plane carried special filters designed to pick up the radiological debris that an atmospheric atomic test would inevitably create. So far none of the flights in the Northern Pacific had picked up a scent, but after this flight returned to Eilson and a huge Geiger counter checked the filters, the technicians detected radioactive traces. This was the 112th alert of the Atomic Energy Detection System (the previous 111 had been caused by natural occurrences, such as earthquakes). After a complex chain of events, involving more flights to collect more air samples, consultations among U.S. government scientists, consultants, and contractors, including radiological analysis by Tracerlab and Los Alamos Laboratory, and secret consultations with the British government, the U.S. intelligence community concluded that Moscow had indeed conducted a nuclear test. On 23 September 1949, President Truman announced that "We have evidence that within recent weeks an atomic explosion occurred in the U.S.S.R." (Note 2)
What made the detection of Joe-1 possible in the first place was a series of decisions that began in 1947. In September of that year, Army Chief of Staff Dwight D. Eisenhower assigned the Army Air Force, not yet an independent service, with responsibility for establishing an Atomic Energy Detection System (AEDS) so that physical manifestations of overseas nuclear development activity could be discovered. Later that year, the Air Force created what would later become known as AFOAT-1. During and after World War II, the possibility of detecting radioactive particles and emissions (as well as seismic and acoustic indicators) became the subject of protracted research and development work, which included the collection of radioactive samples following U.S. atomic tests. During 1947-1949, a complex process of review and decision at the Defense Department led to the creation of an "Interim Surveillance Research Net" that was operating routinely by the spring of 1949. A more comprehensive surveillance system integrating radiochemical, seismic, acoustic, and other methods was not yet in place. (Note 3)
While such senior officials as Atomic Energy Commissioner Lewis Strauss
sought detection capabilities to avoid an "atomic Pearl Harbor,"
intelligence analysts did not see a Soviet test as a near-term likelihood.
Thus, intelligence estimates produced at the CIA's Office of Research and
Estimates (ORE) were significantly off-base, with estimates produced in 1948
and 1949 projecting mid-1953 as "the most probable date," although
mid-1950 was possible. Analysts at ORE were so preoccupied with the "big
picture" of Soviet intentions that work on the nuclear issue became a
marginal part of its effort. Exceptionally tight security measures in the
Soviet Union made it difficult to produce accurate estimates of Soviet atomic
energy work, but CIA clandestine operators picked up highly relevant
information that ORE failed to consider. For example, secretly acquired
information provided significant detail on the production in the Soviet zone of
Germany of distilled metallic calcium, which is integral to the separation of
uranium metal from uranium ore. (Note 4) Nevertheless, ORE
analysts were so disengaged from scientific intelligence activity that three
days before Truman's announcement they produced a paper repeating the Joint
Nuclear Energy Intelligence Committee estimate of mid-1953 "as the most
probable date." (Note 5)
The discovery that Washington had lost its nuclear monopoly would have a decisive impact on U.S. diplomacy and military policy; it was one of the stimuli for an interagency report, NSC 68 (14 April 1950), which called for massive military spending to offset the political and military impact of Stalin's bomb. That is exactly what the Soviets had hoped to avoid by keeping their bomb project secret; even when they responded to Truman's announcement, they did not acknowledge that they had tested a weapon and tried to convey "the impression that the Soviet Union had possessed the atomic bomb since 1947." In any event, the Soviet Union's entrance into the nuclear club soon had a direct impact on its policy; it emboldened Stalin to support Kim Il-Sung's plan for a North Korean invasion of the South. As Evgueni Bajanov put it, when Stalin approved Kim's proposal, he "was now more confident of the Communist bloc's strength." (Note 6)
When Truman made his announcement, he provided no further information about the discovery, even the estimated date of the test, and for many years the U.S. government kept the details secret, although that did not stop informed journalistic speculation about the "mechanical" means for detecting atomic tests. No one without a "need to know" and a high security clearance even knew what AFOAT-1 did. Nevertheless, the Soviets, who had been assiduously gathering intelligence on the U.S. nuclear program, which saved them a year or two in building their own bomb, learned how the test had been detected from spies at the British Embassy. (Note 7) Nevertheless, much of the secret would become declassified. Doyle Northrup, one of the leading officials in AFOAT-1/AFTAC wrote an account of the detection of the Soviet test and two versions were prepared, the shorter one of which appeared in the CIA's classified journal, Studies in Intelligence. Eventually, Northrup's narratives were declassified (with excisions) and received some dissemination. It was not until the 1990s that two anthropologists at Brandeis University, Charles A. Ziegler and David Jacobson, pieced together the declassified archival record so they could produce an authoritative and accomplished account of the early history of AFOAT-1 and the detection of Joe-1: Spying Without Spies: Origin of America's Secret Nuclear Intelligence Surveillance System (Praeger, 1995). (Note 8)
It is difficult to reconstruct decades-old U.S. government declassification decisions, but when agencies declassified the primary sources used in this briefing book, they could only have assumed that significant information about the techniques used to discover Joe-1--the filters on WB-29s that caught nuclear debris or the collection of debris in rain water, the analysis of the debris, and the finding that the Soviets had tested a plutonium weapon--were no longer sensitive. Indeed, important elements of the secrecy of nuclear test detection had eroded; for example, during the late 1980s, NGOs like the Natural Resources Defense Council, working with the Soviet Academy of Sciences, and university consortiums established seismological systems on the territory of the former Soviet Union for monitoring nuclear tests. (Note 9)
Nevertheless, a complete picture of the events of September 1949 is still not possible because government agencies continue to see sensitivities in the nuclear test detection business. Only last year, the CIA denied much of a National Security Archive FOIA request for information on the intelligence work surrounding Joe-1, 59 years after it happened. Moreover, the U.S. Air Force, which manages the nuclear detection system today, played a leading role in the controversial reclassification effort at the National Archives between 1999 and 2006.
Read the Documents
Document 1: Kenneth Condit, History of the Joint Chiefs of Staff: The Joint Chiefs of Staff and National Policy, Volume II, 1947-1949 (Washington, D.C,: Office of Joint History, Office of the Chairman, Joint Chiefs of Staff, 1996), excerpt
This provides a useful, if incomplete, overview of the detection of Joe I and the early stages of U.S. nuclear intelligence.
Document 2: Henry S. Lowenhaupt, "On the Soviet Nuclear Scent," Central Intelligence Agency, Studies in Intelligence 11 (Fall 1967), Secret
Lowenhaupt, who was present at the creation of the CIA's nuclear intelligence work, elucidates the painstaking and complex effort to learn about Soviet progress in the nuclear field during the 1940s. Through such methods as opening mail and acquiring bills of lading, CIA scientific intelligence got some insight into the Soviet nuclear program even if most of that activity remained beyond the Agency's ken.
Document 3: Director of Central Intelligence R.H. Hillenkoetter, memorandum to the President, "Estimate of the Status of the Russian Atomic Energy Project," 6 July 1948, Top Secret
Source: Harry S. Truman Library, Presidents Secretary's File, box 249, Central Intelligence-Memoranda 1945-1948 (copy courtesy of Jeffrey Richelson)
CIA director Hillenkoetter reaffirms a 1947 estimate: While it was "remotely possible" that the Soviets would test a weapon by mid-1950, the "most probable date" was mid-1953.
Document 4: U.S. Air Force, Executive Directorate of Intelligence, "Estimate of Soviet Capabilities in the Field of Atomic Energy," 13 July 1949, Top Secret
Source: National Archives, Record Group 341, Records of Headquarters, United States Air Force (Air Staff), Deputy Chief of Staff for Operations, Directorate of Intelligence Top Secret Control & Cables Section, July 1945-Dec 1954, box 45, folder "2-8100 to 2-8199."
An inter-agency assessment produced a year later, and only six weeks before "Joe I" reached the same conclusion as the Hillenkoetter report to Truman. One modification was that if the Soviets were using only "one method" to produce nuclear weapons fuel, then mid-1951 was the earliest possible date.
Document 5: Memorandum by Chief of Staff of the Air Force Vandenberg to Secretary of Defense on Long Range Detection of Atomic Explosions, 22 September 1949, Top Secret, with enclosures: 1) ) Vannevar Bush et al. to General Vandenberg, 20 September 1949, and 2) Doyle L. Northrup, Technical Director, AFOAT-1 to Major General Nelson, Technical Memo. No. 37, "Atomic Detection System Alert No. 112," 19 September 1949, Top Secret, Excised copy, and 2 Top Secret Source: Harry S. Truman Library, President’s Secretary’s Files, box 199, NSC-Atomic
Documents reclassified at Truman Library.
The information in Vandenberg’s memorandum to the Secretary of Defense and the attached reports shaped President Truman's decision to announce the Soviet test. The Truman Library recently indicated that these previously declassified documents are no longer available in the Library’s open files. Despite their reclassified status, the documents are well described in studies by Jeffrey Richelson, Charles Ziegler, and David Jacobson, who had access to them. Drawing on the finding of Tracerlab and other organizations, AFOAT-1 concluded that an atomic bomb had been detonated, that it was a plutonium bomb with a uranium tamper surrounding the pit, and that the date of the test was between 27 and 30 August. Top U.S. scientific and military experts on nuclear weapons, including J.R. Oppenheimer and Vannevar Bush, supported the AFOAT-1 analysis. On 22 September, General Vandenberg wrote Secretary of Defense Louis Johnson that "I believe an atomic bomb has been detonated over the Asiatic land mass during the period 26 August 1949 to 29 August 1949. I base this on positive information that has been obtained from the system established by the Air Force for the long range detection of foreign atomic activities." Further, "[c]onclusions by our scientists based on physical and radiochemical analyses of collected data have been confirmed by scientists of the AEC, United Kingdom and Office of Naval Research." [Ziegler and Jacobson, Spying Without Spies, 210-211; Richelson, Spying on the Bomb, 90.] (Note 10)
Document 6: U.S. Atomic Energy Commission, "An Interim Report of British Work on Joe," 22 September 1949, Top Secret
Source: Harry S. Truman Library, President's Secretary's Files, box 199, NSC-Atomic
When Washington alerted the British government that an air mass containing radioactive particles was going to pass north of Scotland, London ordered special air sampling flights that picked up more traces of the Soviet test. While the British had their own routine air sampling flight program, the next one was not scheduled until 14 September so important evidence could have been missed had it not been for the U.S. alert.
Document 7: U.S. Naval Research Laboratory, "Collection and Identification of Fission Products of Foreign Origin," 22 September 1949, Top Secret
Source: Harry S. Truman Library, President's Secretary's Files, box 200, NSC-Atomic
Some months before the Soviet test, in April 1949, the U.S. Navy began "Project Rain Barrel" to analyze debris from nuclear weapons tests that might show up in rain water collected secretly at stations in Kodiak, Alaska and Washington, D.C. "Rain Barrel" information described in this report was critically important to forming the scientific consensus about the nature of the Joe-1 test. (Note 11)
Document 8: U.S. Air Force Deputy Chief of Staff for Operations, Atomic Energy Office, Section 1, "U.S. Weather Bureau Report on Alert Number 112 of the Atomic Detection System," 29 September 1949 , Top Secret
Source: Harry S. Truman Library, President's Secretary's Files, box 199, NSC-Atomic
This detailed report shows how U.S. analysts back tracked the radioactive samples collected in early September to a nuclear detonation that occurred sometime between 27 and 29 August 1949.
Document 9: R. W. Spence et al., Los Alamos Scientific Laboratory, "Identification of Radioactivity in Special Samples," 4 October 1949, Top Secret
Source: Harry S. Truman Library, President's Secretary's Files, box 199, NSC-Atomic
Roderick Spence, a scientist at Los Alamos, directed the Lab's Radiochemistry Group. He had played a key role in helping AFOAT-1 contractor Tracerlab to improve its capabilities to analyze atomic bomb debris; therefore, when it was time to work on Joe-1 debris, Tracerlab's capabilities were as good as the AEC's. While Tracerlab was analyzing Joe-1, AFOAT-1 sent Spence a sample of the radioactive material for independent analysis. (Note 12) Confirming the Tracerlab findings, Spence's report concluded that the "samples supplied to us contained radioactive isotopes and that the bulk of the activity was due to fission products of fairly recent origin, their age probably being one month or less."
Document 10: General S. E. Anderson, Director, Plans and Operations, memo to Director of Intelligence, "Implications of Soviet Atomic Explosion," 5 October 1949, attached to memorandum from General C. P. Cabell, U.S. Air Force Director of Intelligence to Director Plans and Operations, "Implication of Soviet Atomic Explosion," 6 October 1949, Top Secret
Source: Record Group 341. Records of Headquarters, United States Air Force (Air Staff), Deputy Chief of Staff for Operations, Directorate of Intelligence, Top Secret Control & Cables Section Jul 1945-Dec 1954, box 46, 9300 to 2-9399
With the U.S. nuclear monopoly, if not U.S. nuclear superiority, ending, U.S. intelligence began to look at the military implications. Air Force intelligence made an early effort with this report, which estimated Soviet capabilities to produce atomic weapons and develop means to deliver them to targets in the United States and the United Kingdom. The Air Force projected that the Soviets already had the capability to deliver atomic weapons on targets in the Northwest United States, using TU-4 bombers on two-way missions. Significant industrial, political, and military installations further east would require one-way missions until the Soviets had an aerial refueling capability. The Air Force estimate had little to do with the actual situation in the Soviet Union, which had only a handful of deliverable weapons by the end of 1951 and did not begin producing them in quantity until 1953. (Note 13)
Document 11: "Estimate of the Effects of Soviet Possession of the Atomic Bomb upon the Security of the United States and Upon the Probabilities of Direct Soviet Military Action," 6 April 1950. Top Secret
Source: CIA FOIA release on Federation of American Scientists Web site.
Several months later the CIA produced a longer analysis of these issues, focusing not only on Soviet nuclear capabilities, but also on Moscow's intentions and the extent to which a nuclear weapons capability increased the risk of U.S.-Soviet conflict.
Document 12: Joint Committee on Atomic Energy, "Report of the Central Intelligence Agency," 17 October 1949, Top Secret, excised copy
Source: National Archives, Record Group 128, Records of Joint Committee on Atomic Energ0y, box 3, JCAE Transcripts
During this hearing of the Joint Committee of Atomic Energy (JCAE), a defensive CIA director Roscoe Hillenkoetter argued that "our estimate was not too far off in the first place" because it was an "error of a few months" (p. 5) and that "I don't think we were taken by surprise." (p. 46). Nevertheless, a fuller picture emerged later in the testimony, when the JCAE Chairman, Sen. Brian McMahon (D-Conn) read from the 1949 estimates of the Joint Nuclear Energy Intelligence Committee (see document 3 above) estimating mid-1953 as the "most probable" date. One of the Republicans on the JCAE, Senator Eugene D. Milliken (CO) observed that it was a "very bad mis-estimate" and that "we have not had an organization adequate to what is going on in the past and he gives me no assurance that we are going to have one in the future." Hillenkoetter had his defenders, for example, Rep. Chester Holifield (D-Ca) argued that "you can't order a piece of intelligence out of Russia like you order groceries in the morning." One Senator, Edwin C. Johnson Johnson (D-CO) questioned why the Soviets would "stumble on to the very best way to do this job without a little assistance from some place or the other." Hillenkoetter cited the high proficiency of Soviet science, but part of the answer to Johnson's question would emerge when Klaus Fuchs was arrested some months later. (Note 14)
Document 13: Doyle Northrup, Air Force Technical Application Center, "Detection of the First Soviet Nuclear Test on August 29, 1949," February 1962, Secret, excised copy
Source: American Institute of Physics, Neils Bohr Library, R.C. Williams Papers, box 3, Letters/Interviews (copy courtesy of Michael Goodman)
Thirteen years after the event, Doyle Northrup, who was a key player at AFOAT/1, wrote what may have been the first detailed account of the detection of Joe I, providing a narrative of the creation of the Atomic Energy Detection System and the analytic effort that followed Alert No. 112. He also showed how U.S. intelligence was able to develop a more accurate estimate of the date and location of the test. While acoustic records at the time of Joe 1 were analyzed, they did not then provide useful information. "On subsequent review these records revealed weak signals at two stations. These acoustic signals were very useful because they helped after the fact to establish the location, time, and size of Joe-1 with greater precision than was possible otherwise." (See page 17). Although this report provides no further information, apparently U.S. intelligence was able to determine the site of Joe-1 to "within 100 miles and the time to within 10 minutes." (Note 15)
Document 14: Doyle Northrup and Donald Rock, "The Detection of Joe 1,"Central Intelligence Agency, Studies in Intelligence, September 1966, Secret, excised copy
This is a somewhat less detailed version, written for a wider audience in the intelligence establishment, of Northrup's 1962 study.
2. For a recent description of the detection of Joe-1, see Jeffrey Richelson, Spying on the Bomb: American Nuclear Intelligence from Nazi Germany to Iran and North Korea (New York: W.W. Norton, 2007), 88-92.
4. See Donald P. Steury, "How the CIA Missed Stalin's Bomb," Studies in Intelligence 49 No. 1 (2005), 19-26, and Richelson, Spying on the Bomb, 92. For more on CIA's early nuclear detection activities, see Henry Lowenhaupt "Chasing Bitterfeld Calcium," Studies in Intelligence.
5. Intelligence Memorandum No. 225, "Estimate of Status of Atomic Warfare in the USSR," 20 September 1949, in Michael Warner, editor, The CIA under Harry Truman (Washington, D.C.: History Staff, Center for the Study of Intelligence, 1994), 319.
6. For background, see Melvyn P. Leffer, A Preponderance of Power: National Security, the Truman Administration, and the Cold War (Stanford: Stanford University Press, 1992), 325-333; Holloway, Stalin and the Bomb, 266-267; Evgueni Bajanov, "Assessing the Politics of the Korean War, 1949-51," Bulletin of the Cold War International History Project 6/7 (1995): 87; Vladislav Zubok, A Failed Empire: The Soviet Union in the Cold War from Stalin to Khrushchev (Chapel Hill: University of North Carolina Press, 2007), 86 (citing Bajanov)
8. More recent studies are Richelson, Spying on the Bomb, especially 62-104, and Michael S. Goodman, Spying on the Nuclear Bear: Anglo-American Intelligence and the Soviet Bomb (Stanford: Stanford University Press, 2007).
11. For useful information on the Navy project, see Herbert Friedman, Luther B. Lockhart, and Irving H. Blifford, "Detecting the Soviet Bomb: Joe-1 in a Rain Barrel," Physics Today 49 (November 1996): 38-41.
14. Johnson's own security consciousness was not of a high order; a few weeks later, during a television interview he let slip that the Soviet bomb contained "plutonium," tacitly showing that the United States had acquired traces of the bomb and could analyze them accordingly. Truman later reprimanded him for the disclosure. "Science: So It Was Plutonium," Time, 5 December 1949