Michael KreponSafe Nuclear Weapons

It’s not easy to make nuclear weapons, build missiles to carry them long distances, and to produce highly enriched uranium or plutonium. It’s even harder to keep nuclear weapons safe so they do not detonate except under orders from a National Command Authority. If a single mushroom cloud appears at a time of crisis or warfare because of an accident, inadvertent or unauthorized use, escalation control will be extremely difficult and all of the presumed benefits of nuclear deterrence can be lost.

Nuclear safety and security techniques and practices are designed to prevent these eventualities. Gates and guards and personnel reliability programs help with nuclear security. All states with nuclear weapons employ these practices. Nuclear weapon design features and other safety techniques help provide insurance against accidental, inadvertent, or unauthorized detonations. Nuclear safety and security reinforce each other. Sometimes these categories merge. For example, authorization codes required to arm and use a nuclear weapon — permissive action links — can be considered as essential for both nuclear safety and security. Additional design features, including the use of insensitive high explosives, are required besides PALs to prevent unwanted mushroom clouds.

The United States has a “one-point safety” standard for all of its nuclear weapons. This standard means that the probability of achieving a nuclear yield greater than four pounds of TNT must not exceed one in a million for any event involving the initiation of the warhead’s high explosive at a single point on its periphery. The United States achieved this exacting safety standard after decades of effort, significant investment, and a learning curve derived from nuclear testing.

Warhead safety mechanisms can be put to the test by many kinds of accidents. They can also be sorely tested in the event that nuclear-armed states engage in conventional warfare. Prior to conflict – and to avoid conflict – nuclear-armed states have been known to signal resolve with increased readiness levels for nuclear-capable delivery vehicles, missile flight tests, and ostentatious missile movements. Pakistan, the weaker nuclear-armed state on the Subcontinent, has employed these signaling devices not only to warn New Delhi of the possible consequences of a clash, but also to spin up Washington to engage in high-level crisis management. India also relies heavily for deterrence purposes on medium-range, mobile, nuclear-capable missiles. Medium-range missiles on the subcontinent can have stabilizing effects because they can be kept far from potential zones of fighting and are hard to target.

Short-range missile systems constitute a new feature in deterrence equations in South Asia. India has flight-tested the Prahaar, a missile with up to 350 km range. Pakistani analysts assert that the Prahaar could carry nuclear weapons, but there are no indications from Indian political and military leaders that this is the case. Pakistan has flight-tested the nuclear-capable Nasr with a range of perhaps 60 km. The Nasr has been advertised as being capable of carrying nuclear warheads to reinforce “full-spectrum deterrence” and to offset the Indian Army’s advantages. Pakistani military leaders believe that short-range systems like the Nasr have “poured cold water” on the Indian Army’s ambitious but unfulfilled plans for “Cold Start.”

To have their proper deterrent effect, short-range missile systems and other tactical nuclear weapons need to be positioned close to the forward edge of prospective battlefields. In the event that crisis management fails and warfare begins, missiles like the Prahaar and Nasr will be fair game for Pakistani and Indian Air Force pilots, who are trained to employ aggressive tactics. If these missiles are struck, if they are carrying nuclear weapons, and if these weapons do not incorporate advanced safety features, mushroom clouds could result along with the dispersal of very hazardous fissile materials.

While much is known about how to make nuclear weapons, fissile material and delivery vehicles, very little information is in the public domain about mechanisms for nuclear safety. States that already have decades of experience can share best practices regarding nuclear security, but some critical information regarding nuclear safety remains highly classified. Even if long-time nuclear weapon states were willing to share sensitive information, potential recipients would not allow outsiders anywhere near their nuclear design information. In other words, there is a mutual taboo about information exchanges on certain topics relating to nuclear safety.

Understanding nuclear safety issues is hard for political and military leaders. Regardless of nationality, they are not trained in nuclear physics, mechanical engineering, and chemistry. They cannot make independent judgments about nuclear weapon safety and must rely on guarantees provided by technical experts and laboratory heads.

U.S. Presidents have relied on competing design teams at different nuclear labs to double-check assurances they have received and to confirm technical calculations. Because U.S. labs have had comparable design capabilities and are highly competitive, this has been an effective way to confirm assurances given. In addition, junior analysts at U.S. national labs have been encouraged to question the design decisions of their elders, and could do so without sacrificing their professional advancement.

In South Asia, as elsewhere, acquiring nuclear weapons to deter potential adversaries has been of paramount importance. Ensuring the safety of weapon designs typically takes time, in part through the testing of nuclear devices. Clashes over how best to proceed between out-sized personalities are not uncommon in the early stages of nuclear weapon programs. In Pakistan, for example, A.Q. Khan and Munir Ahmad Khan were adversaries, competing in an environment where one was favored at the expense of the other. Their labs worked on different means of creating mushroom clouds. A collaborative as well as competitive laboratory culture did not exist in their time; the situation is no doubt better now. Senior officials involved in India’s nuclear program have also had sharp disagreements. Differences of view were publicly aired concerning the advertised yields of India’s 1998 tests. Pakistani claims for its nuclear tests have also been questioned.

Are there stringent internal processes to cross-check, question, and resolve differences of view regarding nuclear safety mechanisms in India and Pakistan? Top-down decisions in South Asia are hard to challenge, and juniors typically do not question seniors. Indian officials, like their Pakistani counterparts, are tight-lipped about the standard of safety they deem necessary for their weapons – as is the case for other states possessing nuclear weapons.

One responsibility of national leadership is to ensure collegiality as well as competitiveness at nuclear labs. This mix can be put to good use developing improved safety features for nuclear weapons. India and Pakistan most assuredly possess nuclear weapons that will create mushroom clouds. Their numbers are growing, and that their means of delivery are diversifying. Do these weapons have safety mechanisms able to meet severe tests that could lie ahead?

Nuclear dangers will grow if there is a resumption of nuclear testing on the Subcontinent. Nuclear dangers will be reduced by avoiding intense crises and warfare. Re-thinking dangerous military practices is another way to improve nuclear safety and security. Weapons that are hardest to maintain control over in wartime and closest to live fire are, by definition, the least safe and secure. If weapon designs are not one-point safe, the best insurance policy against the accidental, inadvertent and unauthorized use of nuclear weapons is making them hard to find and keeping them at a distance from ongoing military operations.

(Note to readers: a shorter form of this essay appeared in Dawn, a Pakistani daily, on April 28th.)


  1. bruce goodwin (History)

    While I affirm that nuclear weapons must be as safe and secure as possible, I do not know what to do with your observations. Any “help” with safety and security provided to these proliferators would result in vertical proliferation by making these weapons more usable and so credible. Such help might also lead to nuclear testing for verification of efficacy. So, what do I do with your observations?

  2. Krepon (History)

    Thanks for your comment.
    The dilemmas and constraints you mention are quite real.
    My observations are not meant for you, but for colleagues in Pakistan and India. If safety mechanisms are not “fool proof,” and if a resumption of testing is not a good idea, then the utility and wisdom of weapons designed for battlefield use warrants deeper consideration.

  3. J_kies (History)

    You have a bifurcation of topics; ‘one point safe’ and PALs are not equivalent. If the persons on the subcontinent have non-one point safe designs they have innate hazards due to accidents without attack.

  4. Jonah Speaks (History)

    I believe the U.S. has already provided Pakistan with information on constructing Permissive Action Links (PALs), implying that assuring safety about current weapons is more important than “encouragement” to build or deploy more weapons. If I am a terrorist stealing a weapon, without PALs I only need to figure out how to push the right “button.” If the weapon is not one-point safe, I can shoot it with a machine gun and hope to get “lucky.”

    Is the one-point safety something that can be assured simply by knowing about or testing electrical wiring and conventional explosives that might be used with a nuclear weapon? Or is it intrinsically necessary to know how the nuclear physics package works (beyond what a proliferator already knows)? Would providing information on one-point safety be useful for miniaturizing nuclear warheads, controlling their yields, or otherwise improving the quality of the nuclear warhead?

    In other words, is the trade-off between nuclear safety and “encouragement” any different for one-point safety information than it is for PAL information?

  5. Bradley Laing (History)

    —-If Assad’s government collapses, do we get the full story on the the Box On the Euphrates was, and all the details of Assad’s secret nuclear program?


    Four years and 200,000 deaths later, the regime of Bashar al-Assad may be on its last leg.

    The Syrian Army, faced with low morale, internal divisions and rapidly decreasing popularity, is facing its most serious challenges since the start of the four-year long Syrian Civil War that has claimed the lives of more 200,000 people.

  6. Bradley Laing (History)

    “This chart shows every model of military submarine in service around the world”