CNN’s Barbara Starr reports that three U.S. sources from the military and intelligence communities are independently confirming that “military intelligence officials” are assessing what will happen to Pakistan’s nuclear weapons in the event of a coup against General Musharaf:

The United States has full knowledge about the location of Pakistan’s nuclear weapons, according to the U.S. assessment.

But the key questions, officials say, are what would happen and who would control the weapons in the hours after any change in government in case Musharraf were killed or overthrown.

Now this is interesting in comparison to Sy Hersh’s Watching the Warheads, where Hersh says both more and less than Starr about US efforts to secure Pakistan’s nuclear weapons.

  • More, in that Starr reports what seems to be a pretty bland analysis of what will happen to Pakistan’s nukes in the event of coup—this is like watching a car accident in slow motion—while Hersh imagines full-on contingency plans to use to elite special forces teams to exfiltrate the warheads.
  • Less, in that while Starr reports the current assessment asserts “full knowledge” of the location of Pakistan’s nukes, Hersh claims that a raging debate exists over whether we know where all the warheads are or not.

“The idea that you know where the warheads are at any given moment is not right,” a former “high level CIA official” told Hersh in a pretty typical passage, “As the operation approaches and the question ‘How certain are you?’ is asked, it becomes more difficult. The fact is, we usually know hours later. We never could do it in real time.”

So what’s the truth? My guess is that Starr’s story is closer to the truth with one caveat. “Full knowledge” depends on what you want to do with the information. The official talking to Hersh says our knowledge is terrible because we only know the location of the nuclear “a few hours later” not in “real time.”

Yet, if the goal is to contact the commanders in control of the weapons after a coup or assassination, a few hours is real time, as Starr inadvertantly suggests when she reports the key question of the estimate: ”[W]hat would happen and who would control the weapons in the hours after any change in government in case Musharraf were killed or overthrown.” [Emphasis mine, of course.]

So, my sense is that we probably have reasonably good sense of where the weapons are located, but a relatively poor sense of how custodial arrangements might shift in the event of a coup or an assassination.