Pakistan's nukes - how secure?

The recent attack by Islamist militants against the Pakistani Army headquarters in Rawalpindi has again brought the issue of the security of Pakistan's nuclear weapons into the world spotlight. I am currently writing a book on nuclear terrorism, Nuclear Terrorism: A Critical Appraisal, and have done much research on the topic. Given the latest bout of concern I would like to make a few comments both about the security of Pakistan's weapons and a possible scenario for their theft.

Attacking the Army headquarters is not like attacking a nuclear weapons facility. If I were a right-wing militia commander in the US I would rather attack the Pentagon than either Fort Bragg or a US nuclear weapons base. Of the three the Pentagon would be the softest target. The siege at Army HQ does not mean that Pakistan's nuclear weapons are insecure. That needs to be evaluated on its own terms.

In order to assess the safety and security of nuclear weapons we need to first know a little bit about nuclear weapons design.

We can divide a nuclear weapon into three main parts. The first is the physics package. This consists of the nuclear materials that are configured in such a fashion that they can be made to induce a nuclear explosion. Thus far Pakistan uses weapons grade uranium. The second is the Arming, Firing and Fusing (AF&F) system that includes the detonators, fuses and initiation components used to set off the nuclear explosion. This is done by imploding the physics package. The third consists of the actual casing in which the weapon is housed such as a missile re-entry vehicle. The overall package is called a warhead.

So far as we are aware Pakistan does not have fully assembled nuclear weapons or nuclear warheads in storage. These three components are stored separately as an added security measure. Given this, in normal circumstances, terrorists would not be able to steal a Pakistani nuclear weapon in one go (so far as we know).

This would require separate but co-ordinated attacks of great force. This is highly unlikely.

In the US, and likely elsewhere, the military maintains assembled nuclear weapons. Besides the obvious security features such as guarding units and alarms and such the US maintains what are called Permissive Action Links. These appear to be, for the most part, electromechanical locks that prevent unauthorised users from detonating a nuclear weapon even if they manage to get their hands on them. Only those with access to the required code can fire a US nuclear weapon.

We don't know much about PALs. What we do know tends to come from computer scientists interested in the history of public-key cryptography. The first were very primitive (according to the BBC the UK's first PALs were like bike locks!). The US has developed, over time, a number of PALs of increasing sophistication. These are labeled by category (CAT) CAT-A, CAT-B, CAT-C, CAT-D and CAT-F. The earlier PAL systems locked the AF&F system from unwanted users. It is considered that the early systems were by-passable.

The early based PALs were not integrated into the physics package.

However, the latest system, CAT-F, is. CAT-F PALs are pretty much unbreakable, it would appear. It would seem that the US has helped Pakistan to develop PALs for its nuclear weapons. Pakistani officials claim that their PAL system is every bit as good as US PALs. I could give Osama bin Laden himself a CAT-F PAL protected nuclear weapon and sleep very comfortably thereafter. In fact you could give A.Q. Khan a CAT-F protected Pakistani nuclear weapon and still sleep easily. So why worry?

The Pakistani claim, it would appear, is not correct.

Pakistan seemingly does not have a US-level PAL system. That is, Pakistan's Permissive Action Links are not CAT-F. If so, we have an empirical precedent for Pakistan over selling nuclear security. This appears to be the case because CAT-F PALs do not just lock the Arming, Firing and Fusing system. To certify that CAT-F PALs work actually requires the physical testing of the nuclear weapon design given their integration with the physics package. For instance, many US underground nuclear weapons tests were concerned with the safety and security aspects of advanced nuclear weapon design.