The Problem of Redundancy Problem

Michael Krepon, in both his recent ACW article and in the Stimson Center Occasional Paper on the 1990 Indo-Pak crisis , does a thorough job of discussing the problems associated with some the conclusions raised in Seymour Hersh’s past investigative work on nuclear weapons and Pakistan. Since I cannot add anything new to that discussion, I would like to talk about a related issue: just how effective US “aid” might be in preventing Pakistani nukes from falling into the wrong hands.

Hersh mentions a disturbing possibility in his recent New Yorker article, that Pakistan would try to remove some nuclear weapons from the “count” that his report says has been turned over to the US. Such a possibility is far from surprising. In fact, Hersh’s listings of various areas of cooperation with the US (including giving target lists, mobilization plans, and security plans) are some of the least credible portions of the article. But the big danger might be that Pakistan actually does some of the things Hersh says they did. That additional danger includes the consequences of Pakistan establishing a sort of second arsenal outside of the normal nuclear depots. It could also include any plan to turn over nukes or their components to the US.

Complex Systems and Nuclear Security

Any of the potential actions Hersh talks about could cause what Scott Sagan calls the “problem of redundancy problem” for the Pakistani nuclear arsenal. I first read about this issue in Scott’s wonderful book, The Limits of Safety where he describes a problem the US faced with its early warning radar in Thule. Because Thule’s radar guarded an important corridor for Soviet ICBM’s, it represented in important target for a precursor attack, an attack only on that site by a short-range nuclear missile; perhaps one fired from a Soviet submarine. This was before bhang-meters were put into outer space to detect nuclear explosions. At that time, the US might not be able to detect a nuclear attack until the warheads started exploding over US cities if the base was destroyed. To prevent this, the Strategic Air Command started to position a B-52 on flying alert over the Thule radar base. That way, if Thule stopped reporting back to Headquarters, the B-52 could be contacted to check if Thule had been destroyed or if there had simply been a communications problem. All was going fine until the orbiting B-52 crashed.

Fortunately, it crashed seven miles or so away from Thule. If it had crashed into Thule it might have caused the US to respond to an assumed nuclear attack. This is an example of introducing redundant safeguards into a system if they are not totally independent. (It also shows how the connection can be not very obvious.) Scott has extended his ideas in the paper linked to above that discusses the connection between increasing security at US nuclear facilities and terrorism that is very analogous to some of the ways agreements between the US and Pakistan might increase the dangers of extremist elements getting a nuclear weapon.

Outside “Assistance” and Nuclear “Insecurity”

Consider the example that Hersh gives of how Pakistan might split its nuclear arsenal into a “normal” arsenal and a second “secret” arsenal that very few even in the Pakistani government know about. Again, I have no idea if this is true. (I strongly suspect it is not.) But for the moment let us assume it is. That means that a significant portion of Pakistan’s nuclear weapons have been removed from the normal procedures for their care, use, and securing them. How does that decrease their security? It increases the so-called “insider” danger in many ways: First, the missile or airbases where the delivery vehicles are kept must have procedures in case of a nuclear alert to let in a strange force purportedly carrying nuclear weapons into the base. If those bases have nuclear weapons in the normal arsenal on site, this could be used to allow extremists masquerading themselves as this force onto the base, allowing them to get that much closer to the normal nuclear weapons.

Second, for these secret arsenals to be truly different than the normal arsenal, the technicians who assemble the nuclear weapons (if the implication in Hersh’s article that the nuclear component is kept in multiple parts such as the pit and conventional explosive is true) will be unfamiliar to the guards at secret arsenal depot. That too increases the possibilities for unauthorized people get close to nuclear weapons. And there are other possibilities for extremists getting access to nuclear weapons either because of the implications of agreements with the US or directly because of the mechanisms set up as part those agreements. Consider the possibility that an arrangement has been made for US troops (such as these teams of JSOC and DoE personnel) could come and take nuclear weapons away. Here, too, the biggest danger might be a terrorist group somehow passing themselves off as such a team and being given a nuclear weapon.

Of course, US assistance to Pakistan in securing nuclear weapons can be very helpful but the direct insertion of the US into securing these weapons could cause a significant increase in the dangers associated with them. These possibilities, which are not obvious and only result from the complex procedures implemented in securing nuclear weapons, must be thoroughly studied. Until that time, the best way of helping Pakistan is to provide money, if needed, for physical security and consultations for other dangers.