Jeffrey LewisB83 Mod X

Several readers have mentioned a piece that Bridge Colby and I wrote in The Diplomat (“How To Worry Kim Jong Il”) arguing that the Obama Administration should seriously consider modifying the B83 to hold at risk certain targets in North Korea.

After receiving a number of angry emails (and few thoughtful criticisms from good friends), I went back to reread the piece Bridge and I wrote.  Allow me to make two observations: First, some people are over-reacting.  Second, people are holding us to a higher standard than the one to which they hold the Obama Administration.  That’s interesting.

It is difficult to write about a piece written with a co-author.  One doesn’t want to speak for a co-author or revisit agreed language.  In this case, however, the fact of my support for a modified B83 has drawn as much attention as the arguments themselves.  So let me make a few arguments, while being clear that I am speaking for myself.

Revisting RNEP

Let’s take the politics out of this for a moment.  Many people ask how our proposal differs from the Bush Administration’s Robust Nuclear Earth Penetrator (RNEP), which seems to me to be a deliberate attempt to put politics in the center of the discussion.  But fine, let’s say this is just like RNEP.  Now, take poor old George W. Bush, preemption, his war against Iraq and everything else you don’t like about him out of the debate.  What was RNEP?

RNEP was a study to see if one of two weapons, the B83 and the B61, could be modified to penetrate into hard rock.  The RNEP study considered changes that might make the bomb casing stronger, lengthen the casing, or change the configuration of ballast. Changes to the Arming, Fusing, and Firing (AF&F) system were permitted; changes to the physics package were only permitted if they did not require underground nuclear testing for certification.

The changes in the existing B61 or B83 envisioned by RNEP were not nearly as dramatic as those envisioned for the WR1 — the Bush Administration’s Reliable Replacement Warhead design, which the Obama Administration has stated is not “new” under the definition in the Nuclear Posture Review — or any number of ongoing Obama Administration Lifetime Extension Programs including the B61 LEP and the common replacement for the W78/W88, none of which have resulted in outraged editorials about “new” nuclear weapons.

The RNEP debate in 2005 centered on a sled test, in which something like a B83 Joint Test Assembly would be slammed horizontally into hard rock. The Bush Administration hoped this would provide enough data to determine whether the concept was feasible.

I keep saying: If you object to this proposal on the grounds that it is new, then you have a problem with the definition of new nuclear weapons in the 2010 Nuclear Posture Review.  Take it up with Barack. If, on the other hand, you believe that the definition of “new” nuclear weapons is the correct one and should be emulated by other countries, then you have no grounds to object to the modification of an existing warhead to meet an existing military requirement without resort to testing.

A Deterrent Benefit

I don’t think Bridge and I argued that conducting a sled-test would solve all our woes with North Korea.  I do think, however, that there is some deterrent benefit to being able to hold at risk Kim Jong Il’s leadership facilities and nuclear forces, some of which are likely located in hard and deeply buried facilities immune to conventional attack.  Or, to be more precise, I believe deterrence suffers some difficult-to-measure harm from not being able to hold those targets at risk and that steps to remedy that gap may be more reassuring to South Korea than other proposals, such as the deployment of fighter-delivered tactical nuclear weapons that Gary Samore mentioned.

There are, of course, limits to benefits from nuclear earth penetrators.  As David Wright has noted, we don’t even know if the modifications we propose would do the trick.  The sled test, for instance, might well demonstrate that a B83 simply can’t be made to penetrate into hard rock.  In that case, of course, there would be no sense in proceeding.  Or perhaps the necessary modifications would be so great as to require a nuclear test.  I would not proceed in that case, as the op-ed makes clear.  But it seems very difficult to argue that we shouldn’t even conduct the sled test that would answer these questions. Perhaps, after all, a new casing would be feasible.

Extreme Circumstances

Another colleague angrily asked how we omitted any mention of fallout.  Here I observe that my coauthor is more constrained in what he can and should say about yield and other technical characteristics of nuclear weapons. I would note that Linton Brooks, when asked  “Is there any way an RNEP of any size that we would drop will not produce a huge amount of radioactive debris?” answered “No, there’s not. … I really must apologize … if we in the [Bush] administration have suggested that it was possible to have a bomb that penetrated far enough to trap all fallout. … I don’t believe the laws of physics will ever let that be true.”  I don’t see any reason to disagree with Linton’s statement about the terrible effects of using a megaton-class weapon against a target that would likely be located in or around an urban area. Nor do I see any reason to doubt David Wright’s observation that “the enormous amount of radioactive fallout a B83 earth-penetrator would create” would make it difficult to use such a weapon.  We were quite explicit in stating that such a weapon would not be more usable than other nuclear weapons and that, given the obvious constraints on using any nuclear weapon, “the United States should make clear that it would also invest heavily in conventional penetrators and functional defeat approaches that would provide the president with credible, indeed more desirable, non-nuclear options for holding at risk targets in hard rock.”

But stating that a nuclear weapon might only be useful in the most extreme circumstances, the standard set in the 2010 Nuclear Posture Review, is not quite the same thing as saying it would never be useful or that the option would have no value at all.  Of course it is hard to put a precise value on such a capability, but we attempted to balance the value of such a capability with nonproliferation concerns by eschewing both nuclear explosive testing and “new” designs as prohibited by Administration policy. If this Administration is seriously contemplating the deployment of B61s to South Korea, I think the balance we struck looks even less Strangelovian.

In Conclusion

There are other important arguments, but I’ve gone on at some length already.  We can continue the debate in the comments and maybe even a second post.  Of course, reasonable people might disagree with our judgment about the balance of deterrence and nonproliferation goals.  I know some of my colleagues on this blog do.  (Hey, we’re not Pravda, after all.)

I want close with what I think is a paradox.  It sometimes seems to me that the only ethical justification for using a nuclear weapon would be in response to a nuclear attack and against a target was both essential and not vulnerable to any other type of weapon.  (Perhaps you disagree on grounds that it is not the only justification; surely then it is the best.) If I try to imagine what sort of target that might be, only hard and deeply buried targets — targets such as a bunker housing a Kim Jong Il who has ordered the use of nuclear weapons or one sheltering his remaining nuclear forces — come to mind. Perhaps all nuclear weapons, then, are objectionable.  That’s certainly a reasonable thing to believe.  But if we admit the need for nuclear deterrence that in any way, shape or form isn’t pure city-busting, isn’t an earth-penetrating B83, the most controversial of all nuclear weapons, really the least objectionable?


  1. InsiderThreat (History)

    And lo, someone finally makes some sense. This is what happens when people meet in the middle, avoid politics and focus on analysis. If it met with objections, I am hard pressed to see their merit. But we shall see what happens as this string develops.

  2. pdisney (History)

    Well done, Jeffrey. It’s tough to defend against criticism without getting defensive; tougher still to remain coherent and actually shed new light on the debate while doing so. This post is a substantive contribution to a debate that was getting crippled with noise.

  3. Cameron (History)

    Your arguement sounds like a textbook definition of deterrence. It looks to me like people are saying that making a weapon more practical in it’s use is the same as making it more likely to be used. The constraints on nuclear weapon use (normative and functional) that exist are not, and never have been, based on ease of use.

  4. Stephen (History)

    I am curious as to why your (and seemingly everyone else’s) proposal for an earth penetrating weapon does not seem to incorporate a guidance package. When attacking deeply buried, hardened targets, accuracy is as important as penetration. Without a guidance package a delivery vehicle would have to make a lower level pass over the target, increasing the chances that the plane will be shot down before releasing the weapon.

    • Jeffrey (History)

      The answer is that I disagree with the statement that “When attacking deeply buried, hardened targets, accuracy is as important as penetration…”

      That is true only for conventional penetrators. A nuclear EPW doesn’t dig all the way down to the target. It only penetrates deep enough to couple the shock to the surrounding geology, crushing any nearby bunkers. So, like in horseshoes, close counts. If you look, for example, at the National Academies calculation regarding a 1 MT earth penetrator against a target in granite with two different circular error probable (CEP) assumptions of 10 and 100 m, respectively, there is no difference down to 200 m and very little difference below that.

      What matters is yield.

    • John Schilling (History)

      At lower yields and shallower depths, precision does start to matter. For targets at 50-150 meters hard rock cover, which is beyond the reach of plausible conventional weapons, a 10 kt nuclear EPW is credible if and only if precision-guided, where a dumb bomb would have to be in the 100+ kiloton range (mostly eyeballed from the National Academies charts).

      If our goal is to minimize collateral damage and fallout should circumstances call for nuclear strikes on deep targets, it would be appropriate to have the option of a precision-guided strike using only the ~10 kt primary. This could be done using either the B61-11 or proposed B83-X.

      If that’s not the goal, then perhaps we shouldn’t have retired the B-53. That would have been plenty deterring, to people who didn’t think we would balk at using it on account of the side effects…

    • Jeffrey (History)

      The B53 was unsafe. There was no choice other than to retire it.

  5. bradley laing (History)

    I see a problem: could any government exist in North Korea that did not have to be threatened with bunker-buster nukes?

    On paper, act least, Moscow has gone from being a Communist Party control city, to a weak Democracy under Yeltsin, to a some-thing-authoritarian under Putin. And the idea of deterring Russia has survived all of this, because russian leaders still might use their nukes as political leverage.

  6. Don (History)


    Your last paragraph sold me. Well played, sir!

  7. Malenki Sasha (History)

    I am having trouble understanding the role of RNEP as it seems to fall short in all categories.

    They appear to have the traumatic effect of nuclear fallout, if not worse, considering all the contaminated dust they put in the air.

    Can’t nuclear ballistic missiles do the same?

    Is this just another way to destroy something critical in a time of desperation without having to do a ballistic missile launch?

    • John Schilling (History)

      Nuclear ballistic missiles, ICBMs and long-range SLBMs at least, cannot cause a warhead to detonate several meters underground. The impact velocities are so high that no known material can survive; the warhead must detonate within microseconds of ground contact or not at all.

      Detonating the warhead several meters underground results in a spherically-symmetric ground shock, in which half of the warhead’s energy propagates roughly downwards. Detonating the same warhead on the surface, results in something like 95% of the energy release taking the easy path upwards. Consequently, destroying deeply-buried targets requires an order of magnitude less raw megatonnage if done using earth-penetrator bombs.

      That translates to an order of magnitude less fallout, and close to an order of magnitude fewer dead innocent bystanders. Push comes to shove, yes, we probably can get the job done with a sufficiently large salvo of ICBMs, but we’d very much prefer not to.

    • Cameron (History)

      It might be able to, but keep in mind that similiar hard targets in the cold war were to be targeted by several SS-19 Mod 2’s sporting up to 5MT per warhead. The fallout from the number of those that would need to be launched, plus the larger effects on the surface (larger CEP, larger blast radius, etc) makes a penetrator with a smaller warhead a more ‘useful’ weapon.

      To maintain a deterrent threat you have to convince the other side that you might use the weapons under your control. Smaller and more accurate weapons are more credible threats than hitting something with the nuclear version of a sledge hammer until if finally pops.

    • Jeffrey (History)

      I want to, again, emphasize that the B83 is one megaton, so it is not a small weapon. My objective is to hold at risk targets that are currently immune, not to make it easier to threaten targets we already hold at risk.

      I am dubious that “smaller” yields would reduce fallout enough to make any earth penetrator more usable than the weapons currently in the stockpile — many hard and deeply buried targets are located in and around urban areas. I simply don’t view the significant political barriers to using a nuclear weapon as susceptible to technical solutions like a reduction in yield.

      The significant cost to using a nuclear weapon is one reason that I thought it was important to include the sentence in the oped emphasizing the need to develop “more desirable” conventional options. (I really don’t want to speak for Bridge.) It would simply be better if we did not need to use a nuclear weapon for any target in North Korea.

  8. George William Herbert (History)

    As the B53 demonstrated, one doesn’t need to actually penetrate to squash bunkers. The B53 was designed to surface couple a lot of 9 megatons into the rock, which will splat underground bunkers some distance deep / away.

    The standard B83 will do so somewhat less effectively, of course, being only 1.2 megatons instead of 9 megatons. But not that much less effectively.

    The objection to surface and near-surface “penetrators” is fallout, but realistically you’re going to get some of that anyways. Particularly in rock, where it just never will get buried very deep. It’s arguably “cleaner” to go with a non-fissile-tamper two stage device with much higher yield than attempt to bury a dirty primary further.

    If push comes to shove, an un-retired B53Y2 secondary coupled to a MOP guidance kit and a whole B83 primary/secondary replacing the old B53 primary is probably the easy win. You get 5+ megatons of big-cleanish-hole Holding Deeply Buried Dear Leader at Risk- power.

    Someone will scream and pout at such a frankenbomb, and a certain amount of secondary/tertiary coupling modeling detail will have to be performed that will not be *cough* in public in more detail, but …

    Or, alternately, dig a bunch of Mark-11s out of the vaults, and fit them with GBU-28 guidance kits.

  9. Rudy (History)

    Very thoughtful, Jeffrey – well done.

  10. Eli Jacobs (History)

    Here are some of my thoughts on the initial article:

    • John Schilling (History)

      The B-2 does not require refuelling stops, ever. It does require refuelling, but that can be done anywhere there is a suitable tanker aircraft, with squadrons permanently based in California, Alaska, and Hawaii with forward-deployed detachments generally on Guam and Okinawa. I believe that a B-2 could take off from Whiteman on essentially zero notice with high confidence of refuelling support all the way to Pyongyang and back, and guaranteed abort to a secure airbase if the tankers do not arrive. Furthermore, the B-2 can orbit any holding point within tanker range essentially indefinitely[1].

      Similarly, the B-2 can be retargeted in flight. It is not necessary that the target’s location be known ten hours in advance. If the location can be estimated to within 500 km or so, ten hours in advance, a B-2 can be positioned to respond to later and more accurate information in about the time required for a standard pizza delivery.

      This is not to say that the B-2 is an infinitely responsive system. For the first ten hours of a crisis, no nuclear strikes can be carried out by B-2 bombers, and that is a significant limitation. But, if desired, a strike can be carried out on less than an hour’s notice, any time after T+10, without forward-deploying any strike aircraft.

      [1] 36-hour sorties are routine, and I believe 48 hours has been demonstrated. The ultimate limit on endurance would probably be the crew’s fresh water supply, but with 20 B-2s and only one Kim Jong-Il, there’s probably no reason to push the limits here.

    • Eli Jacobs (History)

      Thanks for the comment about B-2 in-air refueling – I didn’t know that could be done.

      Regarding the capability of retargeting, though, I don’t think that’s the problem with the ten hour gap. The problem is that we could have very high confidence in the location of Kim Jong-il at one moment and relatively low confidence ten hours later. The ability to update pilots cannot overcome the vagaries of intelligence – vagaries that would be reduced if the B-2s didn’t need to fly so far to reach their target.

    • Jeffrey (History)

      If you look very carefully at the history of leadership strikes, such as Dora Farm, you will see that your description is not quite right. It will take far more than 10 hours for intelligence to come in from the field, wind its way up the intelligence community, then be driven across the river for a Presidential decision. (SECDEF still has to come across the River, even if DNI doesn’t.) Rather, what happens is that we tend to learn that a leader is going to be someplace at a particular time — he’s sleeping at this farm or plans to meet a particular set of bad guys around that campfire. The forces are put into motion and the President has a deadline by which he or she needs to say yes or no.

      Dora Farm, for instance, basically took a full day for the President to say “ok.” (When conventional forces are operating in theater it is a different story, but since we aren’t like to predelegate the use of nuclear weapons to theater commanders to hunt Kim Jong Il, that seems beside the point.)

      For the B83, realistically, this is a scenario in which Kim Jong Il has used nuclear weapons against South Korea and the President is given a set of options that include targeting every bunker for leadership or remaining WMD forces. Ten hours doesn’t really make a difference. The travel time is the least important part of the process, not least because the forces can start preparing to move toward the target in advance of Presidential decision. (The crews need to go to the aircraft, the bombs need to be loaded and so-forth. Given that nuclear weapons have been used, they may even fly to an established go/no go point.)

      The alternative notion, where one just goes about playing “Battleship” with B83s against DPRK targets seems rather unlikely to me, in part because of the intelligence challenges. Dora Farm, for instance, was thought to have a bunker but it did not. If one starts dropping B83s, this is going to kill an enormous number of North Korean civilians, many for no purpose at all. I don’t think any President would take such a decision as lightly as that.

    • Jeffrey (History)

      I’ve cross-posted this comment on the CSIS PONI Blog.

      Thank you for commenting on our proposal. I want to state, again, that we did not argue that a modified B83 was a “panacea” for deterring North Korean provocations. I have a few comments that you may find helpful.

      I can’t speak for Bridge, but I hope he wouldn’t object to anything I say here.

      (1) I disagree that bomber range is a significant factor. You may have overlooked the possibility of in-flight refueling for the B2 bomber.

      During conventional operations over the former Yugoslavia, Afghanistan, and Iraq, B2 aircraft conducted many sorties operated out of Whiteman Air Force Base and back. Eventually, the United States forward deployed some B2s to Diego Garcia in the Indian Ocean and an undisclosed forward operating base. And, of course, the United States has exercised forward deploying B2 bombers to Guam (Polar Lightening).

      Forward deployment may be desirable to some policymakers, but is not necessary. North Korea, at about 6,500 miles from Whiteman Air Force Base, is further than Kosovo, but not quite as far as Afghanistan.

      (Along these lines, you might enjoy these YouTube clips: and I am so glad we didn’t have YouTube when I was an RA at CSIS; we accomplished very little as it was.)

      (2) As I noted over at the ol’ blog, the deterrent benefit here is really quite specific. There is some difficult-to-measure harm to deterrence from leaving some important targets in North Korea uncovered. There is probably some benefit from remedying that gap in terms of reassuring South Korea.

      I have a simple view about deterrence. I don’t want to spend a lot of time inside Kim Jong Il’s head. If he digs a tunnel, I want to be able to crush it. The means to crush it need to be cost-effective at the margin, or Kim can simply build more or deeper tunnels. Everything else is just metaphysics.

      To me, this is fundamentally an expression of deterrence by punishment. Tunneling reveals the value of that Kim Jong Il places on an asset whether it is his person, his nuclear forces or what-have-you. Calculate the hardness, pass go, collect $200.

      (3) I think if you reread our piece, you may observe that we did not “fail to consider the consequences of this capability for the NPT.”

      We explicitly struck a balance between our security commitments and nonproliferation commitments by imposing two constraints: (A) no explosive testing of the modified warhead and (B) remaining within the Administration’s policy widely-hailed policy prohibiting “new” nuclear weapons.

      One may certainly disagree with the balance we struck or argue that the Obama policy is too permissive. But we did take nonproliferation into account, just as we considered stability with Russia and China.

      We didn’t present the proposal with a public relations strategy, of course, but I suspect that if this program were executed like the B61-11 rather than RNEP, no one would much complain.

      (4) My preference for the B83 reflects its relatively large yield, which cures many ills related to targeting, and some serious concerns I have about the viability of the common B61 lifetime extension program, which I don’t want to make more difficult.

  11. Kingston (History)


    Thanks for the follow up post. A few thoughts:

    I don’t really disagree with your point that a modified B83 would not be considered a new weapon under the Nuclear Posture Review – for better or for worse. However, in my view that’s not what this debate should be about.

    For me the issue is whether your proposal makes sense for U.S. and South Korean security. I don’t think it does for the following reasons.

    First, you write:

    “there is some deterrent benefit to being able to hold at risk Kim Jong Il’s leadership facilities and nuclear forces, some of which are likely located in hard and deeply buried facilities immune to conventional attack. Or, to be more precise, I believe deterrence suffers some difficult-to-measure harm from not being able to hold those targets at risk and that steps to remedy that gap may be more reassuring to South Korea than other proposals, such as the deployment of fighter-delivered tactical nuclear weapons that Gary Samore mentioned.”

    As you note, this is hardly a ringing endorsement, and still I think it overstates the ability of earth penetrating weapons to hold HDBTs at risk. As others have already pointed out, reliable intelligence about the location of underground targets (and whatever is inside them at a given point in time) is difficult to come by, North Korea could build deeper facilities or shift more assets to mobile facilities, and the collateral damage from an earth penetrating weapon would be extreme, further undermining its credibility. Moreover, an enhanced earth penetrator would give North Korea a convenient excuse to accelerate its nuclear program, thereby undermining U.S. and South Korean efforts to denuclearize the peninsula. Of course without access to all the relevant classified intelligence information, I can’t rule out that some targets might be susceptible to destruction by a modified B83, but given that North Korea could simply dig deeper in response to the U.S. deployment of such a weapon, I still doubt its’ deterrent benefit.

    I also think you could be more precise about what exactly it is you mean to deter. Are you trying to deter North Korea from engaging in limited “peacetime” conventional provocations against the U.S. or U.S. allies, in light of mutual deterrence at the nuclear level (also known as the stability-instability paradox)? Put in another way, are you seeking to enhance deterrence of low-level conventional conflict by denying Kim Jong-Il a nuclear retaliatory capability (and a secure place to hide), such that the U.S. would have less fear of escalating in response to North Korean conventional provocations?

    If so, I don’t think a modified B83 would do the heavy lifting you want it to do, for the reasons I’ve already laid out: namely, it can’t effectively hold HDBTs at risk. Furthermore, given Pyongyang’s ability to wreak terrible destruction on Seoul with its conventional forces, I imagine the U.S. would be wary of escalating even if North Korea didn’t have nuclear weapons.

    If you’re trying to deter North Korea from launching a large-scale conventional attack against South Korea or threatening to use or using nuclear weapons in a conventional conflict with the U.S. or South Korea (i.e. all-out war), I doubt a modified B83 would add anything to deterrence, as the U.S. already possesses robust nuclear and conventional capabilities with which to eliminate the North Korean regime.

    Moreover, I think you also need to consider the implications of a modified B83 for crisis stability. In the event of a severe crisis or once a conventional war has begun, enhanced U.S. counterforce capabilities could prompt North Korea to take destabilizing measures to enhance the survivability of its forces by, for example, predelegating launch authority to field commanders. Likewise, the perceived ability to take out Kim Jong-Il in a hardened underground bunker could also create strong incentives for the U.S. to attack early in a conflict, lest the Dear Leader take steps to increase the survivability of his bunkers and forces, thereby closing the U.S. window of opportunity. While some of these crisis pressures would exist even in the absence of a modified B83, the additional capability would exacerbate them.

    Regarding the reassurance of South Korea, I agree that a modified B83 is a better alternative than redeploying U.S. tactical nuclear weapons to the Peninsula. However, that’s the wrong standard by which to judge your proposal, especially since (as far as I can tell) the constituency in South Korea in favor of redeployment is confined to a small number of conservative Korean officials and academics.

    I don’t doubt that North Korea’s recent provocations present real security challenges, but I don’t see how the deployment of a small number of modified B83s would be anything more than a temporary means of reassurance – North Korea can simply build deeper bunkers.

    In addition, I think there’s a risk to relying on specific military systems as a measure of the strength and vitality of the U.S. commitment to South Korea’s security. The more the U.S. and South Korea lean on such capabilities as a crutch, the easier it is to avoid difficult but much needed discussions about how the U.S. can continue to guarantee South Korea’s security even as we continue to reduce our reliance on nuclear weapons. I think the extensive discussions among the U.S., South Korea, and Japan that ultimately led to the retirement of the TLAM-N provide an excellent model as to how this can be achieved.

    I also weigh the balance of nonproliferation and deterrence goals differently than you do. Regardless of an existing mid-1990s era requirement for a hard rock penetrator, a modified B83 (particularly to deter limited conventional attacks) would signal a step back in the U.S. effort to reduce its reliance on nuclear weapons, thereby undermining U.S. nonproliferation objectives. I think this cost (along with the others I’ve highlighted above) outweighs the very minimal (at best) deterrence benefit a modified B83 might provide.

    Finally, given the current economic environment, I think you need to at least offer an estimated cost of the modification you’re proposing. Given that you don’t envision this as anything more than a niche capability, perhaps the cost would be minimal, but it would be helpful in assessing its affordability and opportunity costs.

    • Jeffrey (History)


      Yes, this is a niche capability — the advantage is covering a small number of uncovered targets. The only reassurance benefit comes from the actual utility of covering those targets. I wouldn’t propose this solely for the purpose of reassurance. This is a small benefit set against a very significant challenge. It is, however, consistent with my view that reassurance should be based on actual capabilities rather than symbolic measures.

      It would certainly be interesting to consider cost-estimates for the B83 against the possibility that North Korea digs even deeper tunnels. RNEP was relatively modest in cost, especially if it occurs as part of a B83 LEP that is planned anyway. I suspect that it will be cheaper than all the new bunkers Kim Jong Il would need to build, though I am happy to await such a calculation.

  12. Nick (History)

    Very informative discussion, but probably NNWS are wondering what happened to Article VI, in light of efforts to discuss or even improve nuclear EPW to destroy deep nuclear facilities beyond the reach of bunker busters.

  13. Peter Hayes, Nautilus Institute (History)

    Having worked in North Korea, I have seen first-hand in many locations the entries to subterranean North Korea. I have been inside some of these facilities. What is on the surface is epi-phenomenal. There are hundreds of major tunnels and huge caverns, very deep underground, with blast shelter doors; there are hundreds of linking tunnels; there are thousands of exit points. Having had every standing building in the northern part of Korea flattened by SAC after six months of the Korean War, they understand the power of strategic bombing far better than most Americans. They also understand the unique power of nuclear weapons, having accepted and cared for 3,000 Korean survivors of the Hiroshima bombing (the other 19,000 went to the South). There are also multiple Kim Jong Ils (look-alikes) in motion. I have a vision of a slow motion shuttle of strategic bombers from CONUS to PY and back delivering a megaton-sized ground, well, slightly underground burst every few hours, for days on end, in an attempt to destroy these Kim Jong Ils, with enormous collateral damage, while the Russians and Chinese apparently stand by, and all hell breaks loose in North and South Korea due to civilians attempting to flee from downwind fallout. If we succeeded thereby in killing Kim, we wouldn’t know; and moreover, is we killed him, we might create irresolvable difficulties related to war termination, surrender negotiations, and sending credible commands to North Korean forces to stand down and to not use whatever WMD they have pre-positioned and ready to fire. I also surmise that nothing is more likely to cause North Korean forces to unify and rally around a replacement leadership and to fight onto the last man and woman than use of American nuclear weapons to assassinate Kim Jong Il. Finally, I do not and never have understood the concept of “extra deterrence” that comes from adding additional annihilation capacities to those that already exist. How is this marginal deterrence measured? Is there such a thing as “surplus” deterrence? Are the putative marginal psychological effects of the marginal threats to be annihilated likely to be stabilizing or destabilizing? induce rationality or irrationality? In short, this exercise looks to me like a possible existing weapons capability in search of a mission; not a mission that is meaningful. Compared to the existing, highly trained special forces teams that exist and would be sent by the ROKs to hunt down, and kill or arrest the Kim clique at the start of a war, this seems a clueless and relatively ineffectual technological strategy that would backfire badly.

    • Jeffrey (History)


      You always make an original and thought-provoking contribution. Thank you!

    • Nah (History)

      “all hell breaks loose in North and South Korea due to civilians attempting to flee from downwind fallout”

      These bombs would be used in the context of a war on the peninsula, and thus all hell has *already* broken loose and civilians are attempting to flee the battle zone — and most likely, the North will have *already* used nukes, and thus the fallout from our strikes on bunkers in the North would hardly cause any “extra” panic.

      “If we succeeded thereby in killing Kim, we wouldn’t know; and moreover, is we killed him, we might create irresolvable difficulties related to war termination, surrender negotiations, and sending credible commands to North Korean forces to stand down and to not use whatever WMD they have pre-positioned and ready to fire.”

      On the other hand, killing him might *remove* the difficulties to war termination and surrender negotiations, and might eliminate the man who would order North Korean forces to use their WMD rather than surrender.

      “I also surmise that nothing is more likely to cause North Korean forces to unify and rally around a replacement leadership and to fight onto the last man and woman than use of American nuclear weapons to assassinate Kim Jong Il.”

      I don’t know why you would surmise that. There is at least as much reason to believe that killing Kim would cause the North Koreans to lay down their arms. They were willing to die for Kim, but why die for an unknown new leadership?

    • Jeffrey (History)

      There are two different arguments we seem to be having.

      Argument 1: Should the United States maintain some nuclear weapons for the purpose of deterring Kim Jong Il and, if so, what sort?
      Argument 2: If deterrence were to fail, and Kim Jong Il initiated the use of nuclear weapons against South Korea, would it make sense to respond in kind either with a B83 Mod X or something else?

      The B83 Mod X proposal — “do the sled test” — is based entirely on argument 1: the notion that, having accepted nuclear weapons provide some deterrence, we think there is a small deterrent benefit to being able to cover all targets. Some people think the benefit is too small to outweigh the supposed diplomatic costs. I don’t, but it seems like a reasonable discussion to have.

      Argument 2 is very interesting, but I can’t see how we could possibly resolve this out of context. I think Nah is right to observe that the context might produce very different answers. Overall, it is very difficult for me to imagine the sort of scenario that results a North Korean use of nuclear weapons. I can certainly imagine a few, but these are scenarios where we would not wish respond in kind. In general, I have difficulty imagining even the lawful and proportionate second use of nuclear weapons — but I will observe that my objections to using nuclear weapons against Hiroshima and Nagasaki are not widely-shared. Fortunately, resolving argument 2 is not necessary to determine whether to have the option to respond in kind, but only to determine the second-order question of whether to exercise that option in the “extreme circumstances” mentioned by the Nuclear Posture Review.

  14. ArkadyRenko (History)

    To Peter Hayes’ comment, the other plausible use for a nuclear penetrating weapon is to hit military targets, particularly bunkers containing NBC weapons. In this case, it allows the US to defeat North Korean stockpiles which are held in reserve to prevent a South Korean counter-attack across the DMZ. Here, it won’t be a shuttle attack to detonate every bunker in North Korea, only those bunkers which can be plausibly tied, probably with intelligence backing, to storing ballistic missiles. At the same time, it must be acknowledged any hard target warhead will be a tempting first strike weapon for the US if a crisis in Korean peninsula is tending irrevocably to total war.

    The second idea, that South Korea can send its SOF teams north; I am of the opinion that those missions will be worse than useless. Against the heavily defended North Korean leadership, be they dispersed across the countryside or buried hundreds of meters beneath rock. Small enough SOF teams, which can operate reasonably freely in North Korea, won’t have much of a chance to hit leadership targets. That means they’ll be wasted on that mission instead of doing something more productive, such as hunting down missile launchers.

    Finally, the North Korean leadership has shown no inclination to halt its weapon development. They understand two simultaneous facts: first, they face overwhelming conventional and nuclear forces, thus nuclear weapons are necessary tools for national security; second, nuclear weapons give them a tangible benefit, they are a bargaining chip which has and will be used to exact concessions. Nuclear weapons have both an defensive and offensive benefits.

    Thus, thinking that North Korea will unilaterally reduce its nuclear development if the US limits or constrains its weapons development misses the other simultaneous benefits of nuclear weapons for North Korea. Think about it this way, if North Korea only had a conventional threat, does anyone seriously think that South Korea or the US would really pay attention to it?

    • Peter Hayes, Nautilus Institute (History)

      Dear Arkady, I think you underestimate the capacity of ROK intelligence networks in the North, combined with the highly trained and exercised special forces in the ROK military who are mandated to take immediate control of essential infrastructure and leadership locations in the North. The ROK does not want to inherit a radiating ruin with an alienated, angry and insurgent population at war with the South. They want to take over, show who’s in charge, get most of the population and military to capitulate, so they can get on with reconstruction and connecting their networks with China and Russia all the way to Europe. Such an operation will be very fast, and very bloody, which is one of the reasons that we don’t want to know very much about it, nor have it run as a combined operation, as human rights will not be a major consideration during implementation. Separately, our (US-ROK) teams will go hunting down WMD at known or suspected sites. These are different missions, with different units trained for these missions. Both are far more precise, lethal, and likely to succeed than megaton-sized nuclear explosions bouncing the rubble. And yes, the DPRK’s conventional threat to Seoul remains its primary deterrent force; the DPRK’s capacities at this point are compellent, and belong in the political-symbolic rather than the military realm. Best wishes, Peter

  15. Mark Lincoln (History)

    This website is wisely titled ‘arms CONTROL wonk.

    We are stuck with nuclear weapons. That our current stockpile should be maintained is beyond question.

    The idea of modifying bombs to be target-specific is an interesting technical and strategic question.

    Your correspondents, and yourself have perceived the essential questions of feasibility and consequences.

    Your fundamental argument, that we need a means to ‘hold certain targets in the DPRK at risk’ constitutes a tautology.

    Is there the slightest reason to believe that if holding the entire DPRK is insufficient deterrent, that holding certain targets will change anything?

    The idea that we have to do something because doing it will cause it to be done is simply begging the question.

    The DPRK has faced nuclear annihilation for it’s entire existence. It is supremely unimpressed. There is absolutely no reason to believe few bombs targeting a few sites will alter that.

    If we build it, will they come?

    We have much more pressing needs for what money we have to spend upon our nuclear deterrent than building boutique bombs to impress people who do not care.

  16. Anon (History)

    A proper science-based criticism of the B83 Mod X is given by Dr. David Wright:

  17. Peter Hayes (History)

    Dear Nah,

    Thanks for your comments. It’s not obvious at all that the North Koreans will use their nuclear weapons at the start of a war, or outside of North Korea; and, if they did, that they would produce anything like the amount of fallout as megaton-sized ground-bursts. There is a bit of a debate going on about these issues in the public domain.

    If you or other readers haven’t seen them, you might find the following of interest:
    Unprecedented Nuclear Strikes of the Invincible Army: A Realistic Assessment of North Korea’s Operational Nuclear Capability, NAPSNet Special Report – 22 September 2011, Peter Hayes and Scott Bruce assert that North Korea’s options for a nuclear strike are severely constrained—so much so that the only credible use of the DPRK’s nuclear arsenal is to detonate a bomb within North Korea itself to slow down or to stop an invasion in the context of an all-out war with the United States and South Korea. They conclude that, “At this stage, North Korea’s outrageous nuclear threats against targets outside its borders are not backed up by actual capabilities. Countering the North’s rhetorical threat with more nuclear extended deterrence raises tensions instead of addressing the underlying problem of nuclear insecurity. Ultimately, the only way forward is to re-engage the North, and identify pathways that create confidence and reduce the mutual perception of the threat of massive destruction, whether by conventional or nuclear weapons.” at: for the Unthinkable: Countering a North Korean Nuclear Attack and Management of Post-Attack Scenarios NAPSNet Special Report – 6 October 2011.

    We also published a contrary view by Bruce Bechtol,
    “Bruce E. Bechtol Jr. provides this review of North Korea’s nuclear weapons capabilities, which reveals a two-track agenda consisting of both a plutonium (proven) and a highly enriched uranium (likely) program. Scenarios involving both of these programs show that North Korea—despite rather primitive capabilities—can deliver a nuclear weapon that would cause casualties in the tens of thousands. Bechtol states, “While a preemptive strike may seem like the obvious answer to a nuclear attack, North Korea’s ability to strike back with non-nuclear forces would likely mean a full-scale conflict possibly involving hundreds of thousands of casualties. High-level officials in Washington and Seoul have placed renewed focus on planning for nuclear scenarios on the Korean peninsula—but the bottom line is that preventing and deterring a North Korean nuclear attack must be a high priority.” at

    And, while you are at it, you may want to consult Bruce Bennett at RAND on this dismal topic, at:

    Leaving aside the argument of who might fire nuclear weapons first in Korea, I am satisfied with the conclusion by the USCINC in Korea that he sees no need to use nuclear weapons to defeat the North Koreans, period. This view is consistent with a long-standing view in the US military, which is why the US Army was the main push behind the effort to remove them in the mid-late eighties (witness the role played by General Jack Cushman, former CINC CFA in the US Army debate at the time about the dis-utility of theater and tactical nuclear weapons).

    A good public domain example of why US nuke preemption or retaliation were needless against all-out DPRK attack, is found in B. Jack et al, The South Korean Case: A Nuclear Weapons Program Embedded in an Environment of Great Power Concerns, volume II, Regional Rivalries and Nuclear Responses, Panheuristics Final Report to US Defense Nuclear Agency, DNA 001-77-C-0052, February 28, 1978, pp. II-85 to II-95, at:

    Finally, abstract assertions about the possible role for nuclear weapons that don’t take into account force structure, logistics, topography, training, exercises, etc aren’t really useful in determining the applicability and utility of nuclear weapons in Korea. The whole discussion seems detached from military reality on both sides of the DMZ to me.

    Fortunately, some updated studies along the same lines as PanHeuristics earlier work by competent military analysts will be forthcoming in the coming months.

    I see no reason to attribute awesome capacities to the North Korean leadership or its military; which is also no reason to shy away from calling its compellent nuclear threats are anything other than extreme, exterminist rhetoric, designed for political and psychological warfare against the South Korean leadership.

    That’s also no cause for us to respond in kind, even if our rhetoric is more clinical in nature.

  18. AshleyZ (History)

    Every tunnel must have at least one entrance on the surface. Would it be possible to collapse the entrances, under the theory that being trapped forever is as bad as dying right away?

    • Jeffrey (History)

      That’s called functional defeat. It works provided that the asset can’t be dug out or continue to operate from within.

    • Peter Hayes, Nautilus Institute (History)

      That’s why PGMs are so great, especially for the hardened artillery sites that threaten Seoul.

  19. Matthew Hoey (History)

    I found the article very interesting – but wanted to pass this brilliant investigative documentary along and offer some commentary.

    I personally do not buy the entire Cheonan sinking story. Lots of fuzzy fluffery.

    This is a solid, scientifically backed, deep look into the Cheonan events. Worth watching.

    Take into consideration that most LMB NatSec moves are being sold to the public and elected official based on the sinking.

    Also, the Yeonpyeong Island mess (in my opinion) was the product of lotsa ball busting on both sides.

    If ROK NatSec is based too heavily on these two events we should all be skeptical. Many Koreans question the moments that led to these headlines as well.

    Enjoy my friend the documentary my friend and keep up the great work : )