MYS0CW4WNP04GV8·Mystery · event

The 1979 Vela incident

Los Alamos National Laboratory · Attribution

Detected: 22 September 1979, 00:52:43 UTC, by Vela 6911 (Alert 747)⁠[3]
Sensors: Two bhangmeters (YCA and YVA) recording a double flash⁠[1]
Region: South Atlantic or southern Indian Ocean; Prince Edward Islands area favoured from hydroacoustics⁠[3]
Estimated yield: A few kilotons; about 1–2 kt from the timing of the flash⁠[1]
Official review: OSTP Ad Hoc (Ruina) Panel, 23 May 1980: probably not a nuclear explosion⁠[7]
Leaked: Reported by ABC TV on 25 October 1979⁠[3]
Suspected tester: Israel, possibly with South Africa (not acknowledged)⁠[4]
Status: Partly explained
Questions: 0 of 4 answered

People Christopher M. WrightAustralian scientist; co-author of the 2017 and 2018 analyses Jack RuinaMIT professor, former ARPA director; chaired the White House panel Lars-Erik De GeerSwedish scientist; co-author of the 2017 and 2018 analyses Jimmy CarterUS President, 1977–1981 Lester VanMiddlesworthUniversity of Tennessee (Memphis) scientist who measured iodine-131 in sheep thyroids

All 7 people ›

Richard L. GarwinIBM physicist; Ruina panel member Alan BermanDirector of the Naval Research Laboratory in 1979

Catalog keyMYS0CW4WNP04GV8

From 1963 the United States launched Vela satellites into orbit to watch for nuclear explosions banned by the Partial Test Ban Treaty⁠[1]. Each carried bhangmeters, photodiodes that recognise the unique light of an atmospheric nuclear burst: a millisecond flash, a dip as the fireball's shock front hides the core, and a second, longer peak that carries most of the energy⁠[2]⁠[1]. Just before 1 a.m. UTC on 22 September 1979, both bhangmeters on Vela 6911, a satellite then about ten years old, recorded that double flash from somewhere over the South Atlantic or southern Indian Ocean⁠[3]⁠[4]. A month later the National Security Council recorded the Intelligence Community's "high confidence" that a low-yield atmospheric nuclear explosion had occurred⁠[5].

The finding was politically difficult. The suspects were South Africa, Israel, or both together⁠[6], and an Israeli test would have raised hard questions for the Carter administration soon after the Egypt-Israel peace treaty⁠[4]. A panel of outside scientists convened by the White House and chaired by Jack Ruina reported in May 1980 that the signal was "probably not from a nuclear explosion", because the two bhangmeters disagreed in part of the second peak, and suggested light reflected from debris knocked off the satellite by a meteoroid⁠[7]. The Defense Intelligence Agency, Los Alamos, Sandia and the Naval Research Laboratory disagreed⁠[8]⁠[9]⁠[10]⁠[4].

Declassified documents, released through the National Security Archive from 2006 on, and two papers by Lars-Erik De Geer and Christopher Wright in 2017 and 2018 have since rebuilt the case. They argue that the bhangmeter discrepancy has an ordinary explanation in an ageing satellite, that a meteoroid is very unlikely, that hydrophones near Ascension Island recorded an explosion-like signal from the direction of the Prince Edward Islands at the right time, and that iodine-131 in Australian sheep thyroids fits fallout carried east from there⁠[1]⁠[3]. The official record has not caught up: no government has acknowledged a test, and the NRL hydroacoustic report remains classified⁠[3]⁠[11]. Of the 4 questions on this site, 0 are answered. The most likely answer to “If it was a test, who conducted it?”, not yet confirmed: israel, possibly with South African help or territory.

Key questionsset by the editors

If it was a test, who conducted it?Most likely: israel, possibly with South African help or territory

Was the signal of 22 September 1979 a nuclear explosion?Most likely: a low-yield atmospheric nuclear test of a few kilotons, near the ocean surface

All 4 questions

What happened

At 00:52:43 UTC on 22 September 1979 Vela 6911 recorded the signal designated Alert 747. The Air Force Technical Applications Center, which runs US nuclear-detection systems, began an extensive search for corroboration, including air sampling over the Southern Hemisphere⁠[3]⁠[12]. At 02:43 UTC hydrophones near Ascension Island recorded an impulsive signal that, traced back along the ocean's sound channel, implied an origin close to the time of the flash⁠[3]. On 22 October the NSC reported high confidence in a nuclear explosion⁠[5]; on 25 October ABC Television broke the story⁠[3].

First page of NSC memorandum of 22 October 1979 with the discussion paper on the South Atlantic event — NSC memorandum of 22 October 1979 with the discussion paper on the South Atlantic eventNational Security Council, 1979. Public domainArtifact

The government studies divided. A CIA-sponsored panel in October 1979 found the signals consistent with a nuclear explosion⁠[4]. Richard Garwin, soon to sit on the White House panel, wrote on 19 October that he would bet 2 to 1 on a nuclear explosion; he later changed his view⁠[4]. An SRI evaluation of January 1980 found two-meteoroid scenarios would produce such a signal about once in a billion years⁠[13]. The Ruina panel's report of 23 May 1980 concluded the signal was probably not nuclear⁠[7]. The CIA's Nuclear Intelligence Panel reviewed the data in February 1980; its judgments were released fully redacted⁠[14]. Los Alamos, Sandia, the DIA and an AFTAC contractor concluded that it probably was⁠[9]⁠[10]⁠[8]⁠[15]. A DIA official called the White House report a "whitewash" influenced by "political considerations"⁠[4].

Map of the worldOpen the full map

+−

Syowa Station, AntarcticaResearch station

Within 1 km

Prince Edward Islands (Marion and Prince Edward), South AfricaSub-Antarctic islands

Within 10 km

Melbourne, Victoria (abattoir supplying the thyroid samples)City

Within 10 km

Ascension Island hydrophone arrayHydroacoustic station

Within 10 km

Argentia, Newfoundland (SOSUS station)Hydroacoustic station

Within 10 km

Wilson Center, Washington DCResearch centre

Within 100 m

The White House, Washington DCGovernment office

Within 100 m

MappedWithin 1 kmA regionNatural Earth

Timeline

Timeline

Squeezed where little happened. Each mark is a published claim, by how it stands:AcceptedMost likelyStill possibleDisputedRuled out

The accounts

President Carter wrote in his diary on 22 September 1979 that there was an indication of a nuclear explosion in the region of South Africa, by South Africa, by Israel using a ship at sea, "or nothing"; on 27 February 1980 that "we have a growing belief among our scientists that the Israelis did indeed conduct a nuclear test explosion in the ocean near the southern end of South Africa"⁠[4]⁠[11]. Zbigniew Brzezinski expressed a similar view to a biographer in 2016⁠[11].

In November 2019 the Wilson Center, the Middlebury Institute and the National Security Archive held a critical oral history conference with surviving participants: Ruina panel members Garwin and Richard Muller, NSC staff, State Department, CIA and Energy officials, Senate staffer Leonard Weiss, the CBS correspondent who reported Israeli involvement in 1980, and the analysts De Geer and Wright. Alan Berman, director of the NRL in 1979, sent a statement explaining why the hydroacoustic data persuaded him a nuclear detonation had occurred near the Prince Edward Islands⁠[11]. The organisers note that some participants were constrained by continuing classification⁠[11].

What was recorded

The optical record is two bhangmeter traces. They agree closely through the first pulse and the minimum, and diverge in the second pulse; the Ruina panel took the divergence as evidence against a nuclear source⁠[7]. Wright and De Geer show that Vela 6911's memory fault had cut off the second half of its bhangmeter records for every test it saw from July 1972 to March 1978, so there was no like-for-like comparison, and that a similar second-pulse effect, background modulation, appears in all confirmed tests⁠[1]. The timing of the minimum (5 ms) and second maximum (53 ms) gives yields of about 1 to 2 kilotons by the empirical scaling laws⁠[1]. The traces and the comparison signals remain with AFTAC; the published copies are plots in declassified reports and in the 2017 paper⁠[7]⁠[1].

First page of Vela 6911 bhangmeter records of Alert 747 (YCA and YVA traces) — Vela 6911 bhangmeter records of Alert 747 (YCA and YVA traces)Office of Science and Technology Policy Ad Hoc Panel, 1979. Public domainArtifact

The hydroacoustic record is held by the US Navy. A team of up to 75 NRL staff analysed it in a report of several hundred pages sent to the White House on 30 June 1980. According to that study as summarised in a declassified letter of 11 December 1980 from the NRL's research director, three hydrophones near Ascension Island recorded at 02:43 UTC a signal more than 300 times the noise, on a bearing of 198 ± 10 degrees and with dispersion implying a path of about 10,000 km; a second station at Argentia, Newfoundland, also recorded it after a reflection from Antarctica⁠[3]. The NRL report itself remains classified⁠[3].

The radionuclide record is a set of sheep thyroids. Since 1956, thyroids from sheep slaughtered in Melbourne had been sent monthly to Lester VanMiddlesworth at the University of Tennessee for iodine-131 measurement. Samples slaughtered in late October and early November 1979 showed iodine-131, for example about 1.9 ± 0.6 pCi/g corrected to slaughter for one sample. VanMiddlesworth's notebooks are at the Nuclear Testing Archive in Las Vegas and some documents at the National Security Archive; De Geer and Wright published three of his gamma spectra⁠[3]. Other possible correlates, an electron-precipitation event seen by TIROS-N and a patch of aurora seen from Syowa Station in Antarctica seconds after the flash, are suggestive but not unique⁠[16]. Defense Support Program satellites with overlapping coverage did not trigger⁠[9].

The explanations

A low-yield nuclear test is the explanation most of the technical studies reached. The double flash is the signature for which the system was built, and the bhangmeters' first pulses agree with each other and with earlier tests⁠[1]. Los Alamos and Sandia analysts in 1980 judged the signal consistent with a low-yield atmospheric burst, and a 1982 Los Alamos model fitted it as one while accounting for the absence of collected debris⁠[9]⁠[10]⁠[17]. De Geer and Wright add the hydroacoustic signal at the right time and bearing and the iodine-131 carried to south-eastern Australia by a plume that met rain there on 26 September, and estimate a yield of a few kilotons near the ocean surface, with most debris rained out locally by a passing storm⁠[3].

Who tested is a separate question. The CIA's December 1979 assessment, assuming a test, weighed South Africa, Israel, and both together⁠[6]. The 2016 documents show that US officials had sought and received Israeli denials of nuclear-weapons cooperation with South Africa in 1977, and that a later State Department report recounted the NRL finding of a source near the South African-held Prince Edward Islands⁠[4]. The editors of the 2016 posting conclude that a test, most likely Israeli and possibly with South African help, is the stronger reading⁠[4].

First page of The OSTP Ad Hoc Panel report on the September 22 event (Ruina panel) — The OSTP Ad Hoc Panel report on the September 22 event (Ruina panel)Office of Science and Technology Policy, 1980. Public domainArtifact

First page of "The South Atlantic Mystery Flash: Nuclear or Not?" (DIA paper) — "The South Atlantic Mystery Flash: Nuclear or Not?" (DIA paper)Defense Intelligence Agency, 1980. Public domainArtifact

A non-nuclear event near the satellite is the official alternative. The Ruina panel judged the meteoroid explanation "reasonable" but did not claim it was correct, and did not compute the probability of a nuclear explosion⁠[7]⁠[3]. Against it, the SRI study put two-meteoroid scenarios at about once in a billion years, the DIA put a single-meteoroid cause at under one in a hundred billion, and Wright and De Geer found that most debris from an impact on the satellite's solar panels would travel away from the sensors' field of view⁠[13]⁠[8]⁠[1]. A sensor malfunction remains possible in principle for an aged satellite, and the panel noted the absence of "persuasive corroborative evidence"⁠[3].

Published claims

Published claims and their status

Year	Who	Claimed	Status	Reason
1979	CIA-sponsored scientists' panel⁠[18]	The Vela signals are consistent with a nuclear explosion.	Accepted	Every later re-analysis of the optical data by the weapons laboratories and in the open literature reached the same view⁠[15]⁠[1]
1979	US Intelligence Community, via the National Security Council⁠[5]	There is high confidence that a low-yield atmospheric nuclear explosion took place on 22 September 1979.	Accepted	The double flash, the Ascension hydroacoustic signal and the iodine-131 in Australian sheep thyroids all point to a test, though no fission debris was collected⁠[1]⁠[3]⁠[4]
1979	Director of Central Intelligence⁠[6]	If the event was a nuclear test, South Africa, Israel or the two together are the most likely to have conducted it.	Most likely	Later documents and testimony narrowed the reading towards Israel, possibly with South African help, though neither state has acknowledged a test⁠[4]⁠[11]
1980	George Oetzel and Steven Johnson (SRI International)⁠[13]	A meteoroid strike producing the Vela signal would happen about once in a billion years.	Accepted	The DIA reached a still smaller figure, and the satellite's geometry sends most debris away from its sensors⁠[8]⁠[1]
1980	Office of Science and Technology Policy panel chaired by Jack Ruina⁠[7]	The signal was probably not from a nuclear explosion, and light from meteoroid debris knocked off the satellite is a reasonable alternative.	Disputed	The bhangmeter difference it relied on matches background modulation seen in confirmed tests, and laser calibration found the sensors working⁠[1]⁠[10]
1980	Sandia National Laboratories (G. H. Mauth)⁠[10]	Laser calibration tests show the bhangmeters were working, so their differing readings do not make the signal false.	Accepted	The 2017 re-analysis of the optical record agrees⁠[1]
1980	John Mansfield and Houston Hawkins (Defense Intelligence Agency)⁠[8]	The flash was nuclear; the chance of a single-meteoroid cause is under one in a hundred billion.	Accepted	It agrees with the SRI meteoroid evaluation and with the later optical re-analysis⁠[13]⁠[1]
1980	D. S. Sappenfeld and D. H. Sowle⁠[15]	A nuclear surface burst is much more likely than any non-nuclear trigger of the bhangmeters, and the meteoroid theory has serious flaws.	Accepted	Its conclusion is borne out by the 2017 re-analysis of the double flash⁠[1]
1991	Seymour Hersh⁠[19]	Former Israeli officials say the flash was an Israeli test of a low-yield nuclear artillery shell, one of three such tests.	Most likely	It fits Carter's diary and the 2016 editors' reading, but it rests on unnamed sources and Israel has never acknowledged a test⁠[11]⁠[4]
1997	Aziz Pahad⁠[20]	There was definitely a nuclear test in 1979, and the bomb was South African.	Still possible	He later said he had not made the statement, and the claim has not been substantiated⁠[20]
2015	Leonard Weiss⁠[21]	The flash was most likely an Israeli nuclear test carried out with South African logistical help, and the US government covered it up.	Most likely	It agrees with the declassified record and the 2016 editors' judgement, though key documents remain classified⁠[4]
2017	Christopher Wright and Lars-Erik De Geer⁠[1]	The double flash matches confirmed nuclear tests, and the bhangmeter difference is ordinary background modulation, not a fault.	Accepted	It rests on the full optical record of earlier tests, and no published rebuttal has appeared⁠[1]
2018	Lars-Erik De Geer and Christopher Wright⁠[3]	Iodine-131 in Melbourne sheep thyroids and the hydroacoustic signal at Ascension both trace to a nuclear explosion near the Prince Edward Islands.	Most likely	The trajectory and timing fit, but air sampling in Australia and New Zealand found nothing and the thyroid detector was criticised at the time⁠[3]

What would settle it

Declassification would settle most of it. The NRL hydroacoustic report and its data, the full bhangmeter records with the comparison events used by the Ruina panel, and the AFTAC air-sampling results would allow an independent analysis of each strand; two Carter Library files on the event remained classified in 2016 and had been requested for review⁠[4]. Physically, a search of the seabed and island soils near the Prince Edward Islands for long-lived fission products or plutonium with a weapons isotopic signature would test the location. Archival records of naval movements in the area in September 1979 would test the means of delivery.