Michael KreponBroken Arrows

Consider this long post a crude proselytizing effort in this holiday season for those who doubt the existence of God, angels, guardians, guides, benevolent spirits, or deities of any kind or persuasion. One way to get religion is to have two near-death experiences and three surgeries in a year. I do not recommend this. Another is to read about US nuclear weapon-related aircraft accidents that could have turned very ugly. Thomas D. Reed and Danny B. Stillman list no less than fifteen serious accidents during the peak periods of 1950 and 1956-1958 in appendices at the back of The Nuclear Express (2010).

These lists may not be exhaustive. For example, Sam Black, who updated these tables for me when he was working at the Stimson Center, found a reference to an accident on January 9, 1956 involving a B-36 bomber in a February 1991 report by the U.S. Environmental Protection Agency. If readers can confirm or know more about this event, please send word. Here are Sam’s other additions:

January 18, 1959, Unspecified Pacific Base. A grounded F-100 interceptor carrying a nuclear weapon without its fissile core burst into flames when its external fuel tanks were inadvertently jettisoned during a practice alert. The fire was doused in less than ten minutes and there were no reported contamination or cleanup problems.

July 6, 1959, Barksdale Air Force Base, Bossier City, Louisiana. A C-124 aircraft transporting a nuclear weapon without its fissile core crashed during takeoff, completely destroying the aircraft and nuclear weapon.

September 25, 1959, Off Whidbey Island, Washington. A U.S. Navy P-5M aircraft carrying a nuclear depth charge without its fissile core crashed into Puget Sound near Whidbey Island, Washington. The weapon was never recovered.

October 15, 1959, Hardinsberg, Kentucky. A B-52 bomber carrying two atomic bombs collided at 32,000 feet with a KC-135 refueling aircraft shortly after initiating refueling procedures near Hardinsberg, Kentucky. The ensuing crash killed eight crew members and partially burned one of the weapons. No nuclear material was reportedly released, and the unarmed weapons were recovered intact. Both planes had departed from Columbus Air Force Base in Mississippi.

January 24, 1961, Goldsboro, North Carolina. A B-52 bomber on airborne alert carrying two nuclear weapons with their fissile cores broke apart in midair. The B-52 experienced structural failure in its right wing. The aircraft’s breakup released the two weapons from a height of 2,000-10,000 feet. One of the bomb’s parachutes deployed properly and that weapon’s damage was minimal. However, the second bomb’s parachute malfunctioned and the weapon broke apart upon impact, scattering its components over a wide area. Published reports indicated that five of the six safety devices on this weapon failed.
January 16, 1961, Undisclosed U.S. Air Force Base, Great Britain. A nuclear bomber on airborne alert crashed on takeoff causing spilled fuel to erupt into flames which engulfed the aircraft at an undisclosed USAF base in Great Britain. A nuclear weapon mounted on the aircraft’s centerline pylon was badly damaged before the fire could be extinguished. The U.S. Government has not acknowledged the accident and it is not included on the Pentagon’s list of broken arrows.

January 19, 1961, Monticello, Utah. A B-52 bomber carrying one or more nuclear weapons was reported to have exploded in midair north of Monticello, Utah. The bomber had left Biggs AFB near El Paso, Texas, bound for Bismarck, North Dakota, on a routine training mission. Near Monticello the aircraft began climbing from 36,000 to 40,000 feet and soon experienced severe difficulties. The aircraft descended rapidly and at an elevation of 7,000 feet broke into several pieces that landed within an area two miles wide by 11 miles long. Observers on the ground said the plane’s left-wing engine caught fire, after which there was a midair explosion. Five crewmen were killed in the accident.

March 14, 1961, Yuba City, California. A B-52 bomber carrying two nuclear weapons crashed. The weapons’ high explosive did not detonate and their safety devices worked properly. The aircraft had departed from Mather Air Force Base near Sacramento and was forced to descend to 10,000 feet after the crew compartment pressurization system failed.

January 13, 1964, Cumberland, Maryland. A B-52D bomber carrying two nuclear weapons crashed approximately 17 miles southwest of Cumberland, Maryland. The nuclear weapons were being ferried from Westover Air Force Base in Chicopee Falls, Massachusetts, to its home base at Turner Air Force Base in Albany, Georgia, when it encountered violent turbulence and suffered structural failure. Both weapons were recovered.

December 8, 1964, Bunker Hill (now Grissom) Air Force Base, Peru, Indiana. A B-58 bomber lost control and slid off a runway during taxi, causing portions of the five nuclear weapons onboard to burn in an ensuing fire. There were reportedly no detonations and contamination was limited to the immediate area of the crash.

October 11, 1965, Wright-Patterson Air Force Base, near Dayton, Ohio. A C-124 transport aircraft containing nuclear weapon components and a dummy training device caught fire while being refueled. The fire started at the aft end of the refueling trailer and destroyed the aircraft’s fuselage. There were no casualties and the resultant radiation hazard was reported to be minimal.

December 5, 1965, Aboard the USS Ticonderoga (CVA-14) in the Pacific Ocean. An A-4E Skyhawk aircraft carrying a B-43 H-bomb rolled off an elevator on the U.S. aircraft carrier Ticonderoga and fell into the sea. Officials feared that intense water pressure could have caused the B-43 hydrogen bomb to explode.

January 17, 1966, Palomares, Spain. A B-52 bomber carrying four hydrogen bombs collided in midair with a KC-135 tanker near Palomares, Spain. Of the four H-bombs aboard, two weapons’ high explosive material exploded on ground impact, releasing radioactive materials, including plutonium, around Palomares. Approximately 1,400 tons of slightly contaminated soil and vegetation were later taken to the United States for storage at an approved site. A third nuclear weapon fell to earth but remained relatively intact; the last one fell into the ocean.

January 21, 1968, Thule, Greenland. Four nuclear bombs were destroyed in a fire after the B-52 bomber carrying them crashed approximately seven miles southwest of the runway at Thule Air Force Base in Greenland. The B-52, from Plattsburgh Air Force Base in New York, crashed after a fire broke out in the navigator’s compartment. The pilot attempted an emergency landing. Upon impact with the ground, the plane burst into flames, igniting the high explosive outer coverings of at least one of the bombs. The explosive then detonated, scattering plutonium and other radioactive materials over an area about 300 yards on either side of the plane’s path.

February 14, 1974, Plattsburgh AFB, New York. The nose landing gear of a USAF FB-111 carrying two short range attack air-to-surface missiles and two nuclear bombs collapsed as the aircraft was commencing an engine run-up during an alert exercise. There was no damage to the weapons and they were unloaded without incident.

September 15, 1980, Grand Forks AFB, North Dakota. A B-52H bomber carrying nuclear-armed AGM-69 short range attack missiles caught fire while on the ground during an alert exercise. Wind conditions and the efforts of firefighters permitted the recovery of the missiles.


  1. krepon (History)

    This from Stan Norris, the Master of Crucial Minutia:


    Attached is a Defense Monitor that I wrote while working at CDI a few years ago. It is the official list (and language) of the 32 that the Pentagon acknowledges as “Broken Arrows.” I have improved (and corrected) the annotations over the years and would write something more definitive today. There are other suspicious events that might qualify. The Pentagon has been clever in seizing control of the language and defining an “accident” in a certain way beyond what you and I might consider some mishap that we might call an accident.

    I suppose that Stillman and Reed have used the official list but only chose to list 10 of the 32. For each there is usually a great story behind it. For example take the August 5, 1950 accident. The Pentagon and Stillman/Reed have provided a dry account when it was the first time that nuclear bombs were being deployed abroad (to Guam) in response to the outbreak of the Korean War. It must have created panic in the Pentagon and elsewhere. Likewise the November 10, 1950 accident involved a B-50 returning from Goose Bay after earlier deploying weapons (minus cores) there in response to the Korean War. They dropped the sucker in the St. Lawrence River and it detonated.


    Dr. Robert Standish Norris
    Senior Fellow
    Federation of American Scientists

    The Defense Monitor Stan is referring to is “U.S. Nuclear Weapon Accidents: Dangerin Our Midst,” Vol. X, no. 5, 1981.

  2. George William Herbert (History)

    Personally –

    My criterion for very worrying is an accident which involved a live pit. The insertable pit weapons that had accidents without a pit involved were for the most part just a high-security-required conventional bomb, not putting the public in additional danger of nuclear explosion.

    There were some fusion weapon fissile materials involved in some of those accidents, which adds a radiological risk aspect to those models, even if there was no primary pit to risk a nuclear detonation.

    Fortunately the accidents rate has dropped dramatically as planes essentially stopped routinely flying around with real weapons.

  3. John Hallam (History)

    The two most terrifying nuke weapons incidents to my knowledge are the Sept 26 1983 incident involving Col Stan Petrov at Serpukhov-15, in which equipment indicated that the apocalypse was 20 minutes away and aproaching at appx mach-3 (turned out to be vertical cloud formations directly over the launch silos in North Dakota), forcing Col Stan to take literally umm, apocalyptic decisions (fortunately he took the right ones which is why we are still here), and the 1995 incident involving the Norwegian weather research rocket.

    Not sure if those exactly qualify as ‘broken arrows’ as nothing was actually broken, but in each case we came as damn near as we’d want to, to accidentally destroying civilisation and most complex land-based life forms.


    • shaheen (History)

      Who seriously believes that Boris Yeltsin would have launched nuclear weapons without the near-certainty that an incoming nuclear missile was going to detonate over Russia? I remain to be convinced about the importance of this episode – at least insofar as it is used by abolitionists and others to demonstrate the dangers of nuclear weapons. Sure, it was a grave incident that showed how derelict the post-Soviet system was, and how much the “oops” factor is important in crisis management. (If I remember correctly, the Norwegian notification got lost in the meanders of a still-Soviet-style bureaucracy, somewhere between the Foreign Ministry and the Joint Staff.) Yeltsin was shown the Cheget, right. But what makes one think that he ever came close to ordering a nuclear strike?

  4. krepon (History)

    This from John Lopresti:
    Oskins and Maggelet wrote a book ~3-5 years ago called Broken Arrow, which was based on declassified documents plus the authors’ experience working with nuclear bombs, if I recall the lengthy radio interview accurately. I believe Oskins was interviewed by Terry Gross on NPR’s
    Fresh Air program and/or possibly on one of the Pacifica network radio stations.

    One library card style summary of the Oskins work is at:
    That webpage at Alsos says Oskins’ book describes 36 incidents.

    From the interview of Oskins on the radio, I believe I remember one event not included in your web post, a case of a pilot jetisoning a nuclear bomb offshore near an island about 10 miles from the coast of Georgia or Alabama, because of some airplane malfunction. I believe
    the bomb’s approximate whereabouts were depicted as known, according to the interviewee, but that no effort to locate the device precisely and extract it from the ocean was pursued. I believe the interview on the radio was in 2007-2010.

    A Google search appears to show the two authors have produced a second book on the nuclear bomb accident theme within the past 1-2 years.

    The entire Oskins study’s database is bracketed by the thirty years 1950-1980, not addressing the subsequent thirty year span 1980-2010. I do not know how similar nuclear bomb incident reports are generated (and if they exist in declassified state) in the most recent 30-year
    period, for airplane, missle, or other contexts of mishap/malfeasance.

    I wonder if some modern information, i.e. 1980-2010, was produced by government entities or even CRS to help inform Congress in its attempt during Bush’s 2nd term to update the nuclear bomb stockpile, an effort which eventually did not become part of the final budget on the scale which Bush’s allies sought, as I recall.

    John Lopresti

  5. Steeljaw Scribe (History)

    What about Tybee Island near Savannah, GA where the nuke has never been found?

    w/r, SJS

  6. Devabhaktuni Srikrishna (History)

    Thanks for posting this event log. It’s reassuring that at least we have some catalog of these events in the US.

    Would it be reasonable to estimate 5-10 times as many unreported “broken arrow” events among the world’s other nuclear powers? Is there similar the accounting for other nuclear weapons states in the rest of the world?

    p.s. Also agree with Mr. Herbert’s comment that the most worrisome events are the ones where a real nuclear-armed weapon was lost during transport, although the ones without nuclear material are still concerning.

    • bradley laing (History)

      The Bulletin of the Atomic Scientist ran a “Nuclear Notebook” column about the U.K. nuclear arsenal, where they mentioned in passing that the U.K. had issued a report on “incidents” invovling nukees, with the low end being someone dropping a nuclear weapom a few feet to the deck of a ship. In other words, a very precise accounting of *anything* that did not happen as planned, involving U.K. nukes.

    • John Schilling (History)

      Reliable accounting from most of the world’s nuclear powers would be a most unreasonable expectation. However, I would suspect the total figure might be twice the US-only count, with most of those being in the USSR/Russia. The rest of the world, A: simply doesn’t have that many nuclear weapons to make mistakes with, and B: mostly keeps them as a purely strategic deterrent in the safest possible havens. In particular, I don’t believe anyone outside the US/USSR ever maintained a standing airborne-alert force, though the UK at least experimented with the idea.

      And I third the suggestion that we really want a separate accounting of the accidents with live pits. Fortunately, at least in the US case, this is not hard to come by.

    • krepon (History)

      My question, as well:
      Have we learned anything about Soviet nuclear accidents since the Cold War ended and some archival material has become available?

  7. Dwayne Day (History)

    I realize that this list is not supposed to be exhaustive, but there are a lot of things missing and data that’s left out.

    For instance, there is no mention of the Titan II ICBM explosion in Arkansas in 1980.

    And the Palomares incident entry doesn’t mention that the weapon that came down in the ocean was recovered.

    Apropos of nothing, the National Museum of Nuclear Science & History, located in Albuquerque, New Mexico, has two of the recovered Palomares weapon casings. You can see a photo on their website in the exhibits section:


    The DVD “Nuclear Rescue 911” has some interesting material, but it is nowhere near as good as the other DVDs in the series:


    • krepon (History)

      Yes, ICBM-related accidents are missing. The Titan mishap was a humdinger.

  8. Stephen Schwartz (History)

    “…found a reference to an accident on January 9, 1956 involving a B-36 bomber in a February 1991 report by the U.S. Environmental Protection Agency. If readers can confirm or know more about this event, please send word.”

    I’ve got an AP article in my files about that accident. I’ll look for it next week. The accident occurred at Kirtland Air Force Base, New Mexico. John Fleck may know something.

    Chuck Hansen’s CD-ROM “Swords of Armageddon” also makes use of a lot of declassified documents and extends Stan’s 1980s “Defense Monitor” work. That 1981 list produced by the DOD is far from comprehensive, though it’s often treated that way.

    • bradley laing (History)

      This from an Air Force fansite. This seems to invalidate my idea that the May, 1957 incident is the same as at the January 19, 1956 incident.


      ** ” My Name is Nunzio J Castellani, Cmsgt. Ret. I was station at Biggs in 1953 when it opened the 95 Bomb Wing until July 1963 when I was transferred over seas. I was the crew chief of B-36 44-92041 ,and on board , when it crashed on January 19, 1956. We lost number 5 engine on takeoff. Blew and exhaust stack and had flames shooting out of the cowling. We were fully loaded, both bomb bays full, and extra crew on board. Capt Bartel’s , the aircraft commander, and one hell of a good pilot, wanted to continue on the our destination. But the wheel ordered us to return to Biggs. Gross weight, plus a cross wind shift at touchdown, caused us to break in half. With the rear half dragging on the runway. All pax escaped unharmed. The aircraft was totaled. “

  9. bradley laing (History)


    –this may be what you are looking for about the January 9, 1956 incident. The account is about something that happened in May of 1957, and is about a “Broken Arrow” incident. But the dates are different.

    —But think carefully about it. Anyone can put up a website.

  10. bradley laing (History)

    MOSCOW (Reuters) – Russia used helicopters and tugboats to battle a fire that engulfed a nuclear submarine docked at a shipyard on Thursday but emergency officials said radiation levels were normal.

    A local television channel showed a giant plume of smoke above the yard in the Murmansk region of northern Russia and said helicopters were trying to douse flames which witnesses said rose 10 meters (30 feet) above the stricken vessel.

    Russia said all weapons had been removed from the 167-metre (550 feet) Yekaterinburg, which launched a ballistic missile from the Barents Sea as recently as July, and that the nuclear reactor had been shut down.

  11. FSB (History)

    Seems like all the events are related to Air Force nukes(?)

    Seems like we could gain much safety without that leg of the triad.

    Is there a list of non-Air Force accidents?

    • John Schilling (History)

      The 25 September 1959 and 5 December 1965 incidents involved U.S. Navy nuclear weapons. Also, while the USN for diplomatic reasons refuses to officially confirm or deny the presence of nuclear weapons on board its ships, it is likely that the USS Scorpion was carrying nuclear torpedoes at the time of her loss – the Mark 45 ASTOR with W-34 warhead was the primary antisubmarine armament of USN attack submarines at the time.

      Beyond that, the Air Force carries its nuclear weapons on board airplanes, which occasionally crash. The Navy carries its nuclear weapons on board ships, which almost never sink – I believe Scorpion was the last such. The Army, back when it had nukes, stored them in bunkers and carried them on trucks. A sufficiently severe traffic accident might rate a Broken Arrow, but that seems unlikely in practice and I don’t think a list of flat tires and fender-benders would be useful (though I note that someone considered a collapsed nose gear on a parked FB-111 to constitute a Broken Arrow).

      As for increasing safety by eliminating that leg of the triad, basically no. The Air Force stopped carrying nuclear weapons on aircraft in peacetime in 1968, and you will note that the list of Broken Arrows basically ends in 1968. Since then, we’ve had one serious Broken Arrow on the ICBM side (the 1980 Titan II incident Dwayne Day notes), one probable on the submarine side (Scorpion), and one major oopsie with a nuclear-armed B-52. So, one black mark for each leg of the triad (sort of; the Scorpion was “only” an SSN), and no clear winner in the safety game.

  12. jeannick (History)

    Those are mere technical glitches
    a more worrying set is presented by incident like
    the able archer 83 exercice when thwe Nato country went in protective measure , head of state evacuated , lock-down of facilities , full readiness posture ….
    the soviets though it well could be the initial precautions for a massive first strike attack .
    the then international situation was quite tense with Ronald Reagan in full flight with his comments
    “We began bombing in five minutes ”
    how funny !
    Moscow went on stand by and sweated it out ,

    a young analyst called Robert Gates made a report stating
    that the Soviets really felt under constant threats by an aggressive U.S. mindset , pronouncements and actions ,

    Washington was amazed that anyone would take their blustering at face value.

  13. Stephen Schwartz (History)

    Alas, I have nothing more on that 1956 Kirtland AFB accident. The base itself or the EPA may have more information. If not, it sounds like a great topic for a FOIA request.

    For information on naval nuclear accident, the definitive (though now dated) source is William Arkin, et al, “Naval Accidents, 1945-1988,” which is volume 3 of the Greenpeace Neptune Papers series. It does not seem to be available anywhere online except, partially, here – http://books.google.com.bz/books/about/Naval_accidents_1945_1988.html?id=U4EZAQAAIAAJ. Someone also excerpted the information about Soviet accidents here – http://www.skeptictank.org/treasure/GP5/SOVNUK.TXT.

    Here are a couple of New York Times articles dating back to the release of the report in 1989:

    “U.S. Confirms It Lost an H-Bomb Off Japan in ’65”

    “Dozens of Atomic Warheads Lost In Sea by Superpowers, Study Says”

  14. Sam Black (History)

    Michael – good to see the list again!

    FSB – I wasn’t looking only for Air Force incidents, but the search certainly yielded more of these. I don’t know if if the sample of publicly-reported incidents is reflective of the population of all incidents.

    On another note, there was also this (apparently not too serious) incident just last week with a Russian Delta IV:


  15. Alex W. (History)

    I happened to be at the Library of Congress briefly this morning, downloading some of their images (most of which you can’t get much access to unless you are actually physically in the LOC). There was one that reminded me of this post, relating to the Palomares incident in 1966:


    “Don’t you folks worry about us none, we’re just looking for a hydrogen bomb.”

    I’ve uploaded a somewhat larger version here for your viewing pleasure:

    …it’s kind of problematic.

  16. Tangurena (History)

    I recently finished the book “15 Minutes” and it explained a number of the things that went fooey in a number of these mishaps. The “cores” were carried on-board in contraptions called “bird cages” (there is a photo in the book which makes obvious where the nickname came from) that carried the cores.

    >Published reports indicated that five of the six safety devices on this weapon failed.

    As for the Greensboro incident, I’d phrase that the opposite way: 5 of the 6 links in the permissive action link system worked exactly as intended when the bomb fell out of the aircraft. Several of the environmental sensors acted exactly as designed by comparing what happened to the bomb with how a bomb is intended to be “delivered” when “servicing the target”. For example, it would need to go into free-fall from above a certain altitude. Airplane explodes, bomb falls out: sensor (number whatever) says “no parachute, I’m being dropped from an acceptable mission altitude – it is time to make the donuts” and arms. There are a few other sensors and they all did what they were supposed to. But the important part – command authority to set the PAL to “you are authorized to explode when ready” was not set, so the bomb just made a mess instead of making a nuclear explosion.

    I guess that because I’m an engineer by training, most of the accidents that happened didn’t frighten me much (when reading about them). The one that terrified the pants off me was the 2007 incident where a B52 flew from Minot to Louisiana with 6 armed/live nuclear-tipped cruise missiles on one of the pylons – and it sat unguarded overnight. Maybe I’ve seen Thunderball too many times?