Michael KreponOrganizations, Accidents, and Nuclear Weapons

Readers may recall an earlier post about Vasili Alexandrovich Arkipov and the Cuban missile crisis. Arkipov was one of three officers on board the Soviet diesel sub B-59 who voted on whether to fire their nuclear-tipped torpedo while being depth charged to the surface by the U.S. Navy. Arkipov voted nay, vetoing the idea. This wasn’t the only near catastrophe during the Cuban missile crisis.

Scott Sagan recounts another hair-raising event that occurred at Vandenberg Air Force Base in his book, The Limits of Safety:

When DEFCON 3 was declared at the beginning of the crisis, one Atlas ICBM was ready for a test launch that had been previously scheduled for later that week. Although nuclear weapons were being placed on the ICBMs in surrounding launch facilities [adding nine nuclear warheads to the U.S. order of battle – MK], the test reentry vehicle was maintained on this single missile. Despite the severity of the crisis and the emergency alert operations taking place, this Atlas ICBM was actually launched, without further orders from Washington political authorities, at 4:00 A.M. on October 26th. Ironically, while senior leaders in Washington were carefully monitoring the operational alert status of the Soviet missiles in Cuba, a U.S. missile was actually being launched, without their knowledge, in California.

How could this happen? Crisis management techniques and situational awareness were still at rudimentary stages in 1962. President Kennedy and his team were focusing as hard as they could on Cuba and its environs, not on Vandenberg Air Force Base. The missile flight test was pre-planned and pre-approved. Perhaps the base commander didn’t connect the dots, or perhaps he wanted to send the Kremlin a message. If anyone out there can shed further light on this episode, please do. In any event, the Atlas missile flight test went arcing over the Pacific Ocean during the worst crisis in our nuclear history.

Washington and Moscow have come to rely heavily on satellites to provide early warning of missile attacks. Fortunately, Moscow didn’t have one in orbit back in 1962.

The subtitle of Scott’s book was Organizations, Accidents, and Nuclear Weapons. It might just as well have been subtitled Don’t Presume You Have a Handle on this Problem. The Limits of Safety was published in 1993, the same year the paperback version of Bruce Blair’s The Logic of Accidental Nuclear War appeared. The powerful intellects who ordered the field of nuclear deterrence dealt mostly with rational, unitary actors. Scott and Bruce dwelled on accidents, bureaucratic screw ups and potential breakdowns of command and control.


  1. Andrew Tubbiolo (History)

    By the time alert levels are high, you’ve taken most of the steps to fire. Everybody is on automatic pilot just waiting for chance to go. One wonders what the future will be like with 20+ nuclear powers, having small conventional buffers and the nature of local brushfire wars still holding sway. It’s one thing to be a Soviet sub capt knowing there’s 4+ more subs in the region, and hundreds more in the North Atlantic in and out of port. Can a sub commander exercise the same restraint if he knows he’s the only combat capable sub for the next 3 months? It’s one thing to risk your own destruction in the name of crisis containment knowing the price to the nation is high, but not toxic. However if you’re the last man standing in your nations particular arm of warfare, the temptation will be overwhelming to to use it before you lose it. I wonder if we will find that small armed forces might be far less stable than large armed forces?

  2. bradley laing (History)

    MOSCOW (AFP) – Russia fought a deadly battle Tuesday to prevent wildfires from engulfing key nuclear sites as Prime Minister Vladimir Putin took to the air in a water-bombing plane to join the firefighting effort

    [deleted material was here]

    Two members of the Russian armed forces were killed Monday fighting wildfires around Russia’s main nuclear research centre in Sarov, a town in the Nizhny Novgorod region still closed to foreigners as in Soviet times

  3. Mark Gubrud (History)

    Stanislav Petrov, the Soviet officer who sat on a false attack warning in 1983, reportedly said that he did so because “God told him it wasn’t time for the world to end.” Col. Petrov may well have been a believer, but I suspect the deeper explanation is a principle that I believe makes nuclear deterrence stronger and more resilient to the kinds of upsetting events we’re discussing than many people may realize. This is the principle that it’s never a good day for a nuclear holocaust, it’s never time for the world to end.

    Of course, people may sometimes say “It’s a good time to die,” but usually either because a possibly worse death is definitely imminent anyway, or because they will sacrifice themselves to fight for those who will survive. If nobody will survive, or almost nobody, and nothing else you value will survive, and if there is still any chance at all of avoiding this, then there simply can’t be any good reason why Now is the Time.

    Put yourself in the position of Col. Petrov, or of Yeltsin when the Norway rocket went up, or of Khrushchev and Kennedy and their advisers in 1962. You’ve got an attack warning, but it could be false, or you’re looking at confrontation that could escalate out of control or even provoke a preemptive attack, but then again reason might prevail and you might make it through. Well, if the attack is real or if the crisis is unstoppable, you’re dead, and so is everything you value, no matter whether you retaliate or launch first. But hope springs eternal, so you bet the warning is false, or you bet you can find a way out.

    The Cuban Missile crisis was dangerous indeed. It could have blown up over this missile, or combat could have escalated when the US Navy depth-charged a Soviet submarine, or when a U-2 was shot down over Cuba, or when Kennedy was urged to order a preemptive invasion of the island… Not to mention both sides taking the occasion to conduct high-altitude nuclear tests to discover their electromagnetic and atmospheric effects — Great timing, guys. Yet we’re still here.

    I find the fact that we still have the threat of nuclear annihilation hanging over us, 20 years after The End of The Cold War, to be insane and intolerable. I’m all for getting rid of nukes ASAP, and I believe this is doable, and OTOH, I’m deeply worried about the implications of ever returning to the frontiers of tense and complex technical confrontation, shortening timelines and, God forbid, automated decisionmaking.

    But I’ve come to believe that, as long as ordinary human beings are making the decisions, MAD is like one of those giant rocks standing upright in the desert. No doubt something could push it over, and given enough time, something will. But for now, it takes a harder shove than you might think just by looking at it.

    • Andrew Tubbiolo (History)

      Mark, I think your point about no day being a good day for everyone to die is the basis for MAD, and I also agree it has a stability. I think what will tax the limits of MAD in the poly polar world will be when tomorrow will look bleak with our without a nuclear exchange. How will the future look to a Pakistani under threat of having his nation overrun by India? What would a future of Islam having to live under Hindu rule look like to a patriotic Pakistani? The stunted regional powers of the future won’t have a lot to fall back on, defeat may come to one side faster than any combatant may plan for. This will tax MAD as it was never taxed when it was used in the era of superpowers.

  4. Gridlock (History)

    Not shown; the reasons why the US Navy was depth-charging them in the first place.

  5. John Schilling (History)

    If I put myself in the position of (say) Boris Yeltsin facing an ambiguous warning of a possible attack, I’m pretty sure I *don’t* want to have to complicate matters by wondering whether the various green lights on the status board are actually coming from some latter-day Petrov who’s hearing Voices from God telling him to pass me false information about the unfolding crisis. Or Jack D. Ripper, hearing a different sort of voice…

    Humans, being humans, will never be entirely reliable in such matters. But where humans controlling nuclear weapons are concerned, I’m not a big fan of anyone substituting their private judgement for the official chain of command. I’m pretty sure that even the average “rogue state” dictator can be trusted to Not Blow Up The World, if he’s getting accurate and reliable information from his subordinates and if he is confident his orders will be obeyed. It’s when people don’t have or don’t trust the facts, that they do stupid impulsive things that we’ll all regret.

  6. Bulldog (History)

    The ICBM launch on October 26, 1962 was actually a Titan II not an Atlas. Doesn’t change the point you’re making, though. Source: Titan II, A History of a Cold War Missile Program, David Stumpf, p. 72.

  7. Jonathan McDowell (History)

    The Atlas launched on Oct 26 was flown to Kwajalein as an antimissile target, probably to calibrate the Nike Zeus radars. There was also a Thor Agena space launch from Vandenberg that day, and on the east coast (much less threatening as that’s known to be just a test site, not operational, and not launching toward USSR) a Titan II.
    But you have to remember in these days the launch rate of such missiles was very high, several Atlas launches per month typically from Vandenberg. MUCH MUCH WORSE THE SAME DAY was the launch from Johnston Island of a Thor with a LIVE NUKE for a high altitude nuclear test. Way to not be provocative…

  8. Mark Gubrud (History)


    I’m pretty confident that Pakistan’s 100 or so nuclear weapons will be enough to deter India from attempting to overrun Pakistan in any conceivable circumstance, and I’m not sure I can see any reason why the Indians would ever want to anyway. A shooting war between the two sides could perhaps escalate out of control, but they fought one just a decade ago and it didn’t. One has to worry, though, about the dangers of the sort of technological confrontation (first-strike missiles, warning systems, missile defense, etc.) that they are creating over there, imitating the example set by Uncle Sam and his sparring partners.

    One may object that my expressed faith in the power of nuclear deterrence implies that we really should not get rid of the damn things, or at least that we can’t or shouldn’t until underlying conflicts and confrontations are resolved first. But my belief is that arms control, disarmament, conflict resolution and peace-building go hand-in-hand, and that global momentum toward nuclear disarmament will go a long way toward defusing supposedly intractable conflicts, and enable progress to be made on conventional arms control, disarmament and common security.


    You have a point, but it’s more or less orthogonal to my point. I suppose you have put your finger on why Col. Petrov was not given a medal and promotion. But notice that he also was not subjected to any harsh discipline, as he probably would have been had his superiors not also been human, living in constant dread of the whole business, and doubtless privately sympathetic and even grateful to Petrov for sacrificing his career to avoid that roll of the dice. If you have only one career to give to national and global security, you could do worse.

    I’m glad to hear that you agree that “even the average “rogue state” dictator can be trusted to Not Blow Up The World,” but I’m not sure I get why you think doubts about the information he’s getting would make him more likely to push the big red suicide button.

    Pavel Podvig has argued that it is not a bad thing for Russia’s early warning system to be as decrepit as it is; he says that the Russians would never believe an attack warning from their system, whereas Americans would be more likely to credit one from ours. Thus, we expose humanity to the risk of annihilation due to a technical malfunction, while the Russians remain more positively and humanly in control of their doomsday machine.

  9. John Schilling (History)

    The reason I believe “rogue state dictators” become more dangerous when they are getting bad information is, simplistically speaking, that the thing which might well make them “push the big red suicide button” is the belief that they’re dead already and the only decision left to them is whether to take their enemies down with them. Bad information can make them believe that when it isn’t so.

    Or the other way around, but in fact it seems that the United States is rather cautious about implementing regime change in “rogue states”, and the rest of the world even more so. Of 2002’s Axis of Evil(tm), Iran and North Korea are still going strong. Likewise, closer to home and fortunately still a long way from nuclear, Chavez and Castro. Well, *a* Castro at any rate…

    But it is neither stupid nor insane for them to fear that maybe they have been picked to be this decade’s Hussein. So if there’s a major crisis brewing, every general or colonel who is telling El Presidente, “my sources all say we’re still OK, the guy with the binoculars and the shortwave outside of Whiteman AFB is reporting all the B-2s are still on the ground” or the like, *and is trusted*, is a good thing. If El Presidente doesn’t believe the guy, he isn’t doing any good even if he is telling the truth this time.

    And if he choses to sacrifice his career over the matter, as you note he only gets to do that once, whereas there will likely be many crises.

    Oh, and a corrollary: if the United States, or any other major power that decides to take down a nuclear rogue, thinks that opening hostilities with attacks (hard or soft) on enemy C3I targets is a good idea, they’d best be damn sure they can do it right. A dictator who still has one working phone line to a nuclear missile base, and no clue what is going on except that he is under attack, is not a good thing to share a planet with.

    As far as trustworthy early warning systems are concerned, their value depends on what you plan to do with them. Podvig notwithstanding, I don’t think any of the major nuclear powers plan to do the launch-on-warning thing, or are at all likely to. There’s lots of things one can do with an extra ten minutes’ warning that will reduce, rather than increase, risk. Again, the warning has to be believed, as does the absence of warning.

  10. krepon (History)

    On August 11th, I had the honor of participating in a panel discussion at STRATCOM’s second symposium on deterrence in Omaha. Our panel, which was on the similarities and differences between space deterrence and nuclear deterrence, was chaired by Lt. Gen. Larry James, who is the Commander at Vandenberg. I asked Gen. James to inquire with the Wing historian about what transpired during the Cuban missile crisis, who provided the following information:

    “The pinnacle of the Cuban Missile Crisis ran from 18-28 October 1962. Between those dates, Vandenberg launched two missiles on 26 Oct, a Thor-Agena D and an Atlas D in support of the Army’s Nike Target Program being conducted at Kwajalein.

    Vandenberg was not alone is launching missiles during this same period. On the East Coast, five missiles were launched from Patrick AFB between 18-27 Oct. They fired two Atlases, a Pershing I, and a Titan II.”

    The Wing history provides no information on the motivation behind these launches.


Pin It on Pinterest