Geoff FordenBMD Hearts Lewis and Postol

Direct hit is like bowling—if you do not hit the pocket exactly, then many pins will continue to stand” Dr. Hans Mark, as quoted by Richard Lloyd in “Physics of Direct Hit and Near Miss Warhead Technology”, p. 12.

For every winner in a Defense contract competition, there is a loser. And those losers never give up trying to win a piece of the pie. So you can be sure that there were plenty of people cheering for Ted Postol and George Lewis inside the hallowed halls of the military industrial complex when they pointed out the obvious: there is more to a missile than just its warhead. (You can see a “White Paper” of their analysis here, which includes pictures.) The ranks of those organizational “losers” are filled out by the people who have invested their careers in fragmentation warheads and they were cheering because there are severe limitations to hit-to-kill; some of the biggest were pointed out by Ted and George in their article.

Instead of acknowledging the difficulties inherent in hit-to-kill technology, the Pentagon struck back by saying the shockwaves induced by the hypervelocity collision between the two extended bodies (the incoming missile’s warhead/rocket body complex and the interceptor) were sure to destroy the payload. The Missile Defense Agency’s spokesperson, Richard Lehner, stated that readings from test sensors “prove conclusively” that mock warheads “were destroyed and were no longer a threat.”

Hmmm…, what type of warhead and with how much certainty was it destroyed?

It’s true that shockwaves running up the missile body might destroy a unitary warhead (such as a nuclear bomb) but then again, they might not. The Pentagon’s scientists and engineers want a more certain kill mechanism than that, and who can blame them? Lloyd’s book is full of discussion about directly hitting warheads but failing to destroy all or even most of the (biological) submunitions it might contain. (Biological submunitions are very unpopular today among the nonproliferation crowd; I understand there was a vote taken and it was decided that biological weapons were not weapons of mass destruction. I disagree, but that is another story.)

One of the many possibilities missile defense researchers were trying to figure out was how to destroy—i.e. puncture—all the submunitions. Imagine how hard it would be to puncture a majority of submunitions if you are hitting the fuel tank several meters behind the munitions canister. It is very unlikely that any shockwaves propagating up the rocket body walls to have enough of a hydraulic ram effect would cause many of them to rupture. And if they cannot rupture thin liquid filled capsules packed shoulder to shoulder, how are they going to destroy a fairly robust nuclear warhead? On top of that, add in a warhead separation mechanism (to be blown later in flight). Could shockwaves reliably propagate across that?

Ted and George’s observations are not good news for hit-to-kill technology. The Pentagon’s reaction—basically denying the problem—is even less good news for the defense of our country. Fortunately, there will be plenty of losers from the last bureaucratic battle over missile defense technology—the one that selected hit-to-kill as opposed to fragmentation warheads—who will be eager to wade in and try to correct our missile defense systems.

UPDATE Just to be clear, Richard Lloyd’s book is about enhancing hit-to-kill technology when the interceptor comes close to hitting the warhead. It is a way of leveraging the advances that have been achieved in hit-to-kill technology in recent years. Ted and George’s paper still point out some of the problems left to be solved: ie leaving the last stage attached as long as possible etc.


  1. Rob Lewis

    “Biological submunitions are very unpopular today among the nonproliferation crowd; I understand there was a vote taken and it was decided that biological weapons were not weapons of mass destruction.”

    Apologies; I’m just a layman lurker, but I would be much obliged if you could link to something on this. Many thanks.

  2. John Schilling (History)

    I strongly suspect that at ICBM-type intercept velocities, yes, the combination of hydraulic ram, shock, secondary fragmentation, and the like would disable a hardened nuclear warhead or rupture most chemical/biological submunitions even with an impact on an empty tank 5-10 meters away. Postol & Lewis’s extrapolation from Patriot/SCUD intercepts is simply not relevant, and their analysis is conspicuously non-quantitative on that point.

    Of course, the Pentagon’s rebuttal is even less substantive, and if it is more in line with my own educated intuition I’d still like to see some actual math to back it up. And if there’s any doubt, a modest fragmentation warhead is cheap insurance.

    Much of missile defense is hard. But if one can get a KKV reasonably close to an ICBM, the relative energy availabe so grossly overmatches the hardness of any plausble target that there’s no excuse for failure at that point. We can still get failure, if proponents and opponents of various technologies are more concerned with proving the other guy wrong than getting the job done, but there’s no excuse for it.

  3. Mark Gubrud

    If you read Lehner’s response he carefully avoids asserting that any evidence shows the warheads were actually destroyed:

    “[Postol and Lewis] had no access to classified telemetry data showing the complete destruction of the target missiles.”

    Yes, everyone agrees the missiles were destroyed, but does any classified telemetry data, or any evidence at all, show that the warheads were destroyed? Lehner does not say so.

    Even less credible is the claim of

    “subsequent sensor views of the intercept that were not publicly released so as not to reveal to potential adversaries exactly where the target missile was struck.”

    Lehner does not dispute that the publicly-released videos reveal where the target missiles were struck, or the accuracy of P&L’s placement of the impact points. So, in case any potential adversaries are really interested, and this hadn’t occurred to them before, they now do know where the target missiles were struck. So, MDA can go ahead and release the sensor views now, if they really show that the warheads were destroyed in spite of the impacts missing the warheads—which, you might have noticed, Lehner does not actually say they do.

  4. Mark Gubrud

    The core of one of these kill vehicles is a fairly dense metal object, and in hitting a relatively thin and light rocket frame, particularly in a side impact like these, they will tend to punch nice holes through the targets and transfer very little energy into any shock traveling perpendicularly to the momentum of the interceptor. What shock does get launched will tend to be absorbed by crumpling metal. Fragments, too, will tend to fly in the same direction as the thing that set them flying. I’m not saying the warheads will enjoy a smooth ride, but they’re built tough.

  5. Tal Inbar

    Caveat emptor

    Just to add several points:

    Ah. The wonders of over interpretation of pictures strikes yet again. The photographic evidence that supported Postol and Lewis arguments, were NOT the last ones obtained by the KV’s sensors, hence it is IMPOSSIBLE to draw the conclusions drowned.

    Furthermore, why are they assume that the cloud of debris that will be created by some‎‎ hypothetical pyrotechnical device will continue to travel as a compound dense cloud – for MINUTES after its creation? Evan a small delta V (let’s say 1 meter per second) will give the debris the ability to move some kilometers apart, allowing the sensor of the KV AND the radar of the BMD system, to discriminate the real RV from any other debris. It is unlikely though that the KV’s sensor will be able to see a significant debris in its field of view.

    Regarding the hit point – in some tests there was a mock warhead that was SEPARATED from the target missile – and it was successfully hit by the KV.

    I tend to agree that in the HIT TO KILL strategy there is an inherent danger – if you miss the target – it will hit you. In this case, the same is applicable to Postol and Lewis – The missed altogether.‎

  6. George William Herbert (History)

    For the record:

    This criticism is only relevant to non-separating warheads, or hitting a ICBM upper stage after burnout but before RV separation.

    The first category is almost but not quite entirely occupied by “Scud”, and expanded out is entirely occupied by short range missiles.

    Even if a single center tank hit doesn’t disrupt the harder end of the warhead spectrum, a second hit on the largest remaining fragment would. Plus, if the warhead depends on the rocket body behind it for reentry stability, blowing the rocket body apart bodes poorly for warhead survival and performance.

    For the ICBM before RV separation case, if you hit it during final stage burn you at least cause the warhead to fall short of desired trajectory and target, by some large amount depending on how early in the process you hit it.

    For a hit after burnout but before separation, you’d at least cause the RV to tumble and not be released smoothly. Also, RV thermal protection systems for ICBMs are notoriously NOT that robust – peak heating and total heat input are much higher, particularly as the RV nears the ground, and TPS mass budget is fairly low. Fatal damage levels have been done by RVs hitting hail or raindrops on descent, though they’re designed to take some of that.

    Also, many RVs will do very poorly if they enter other than nose-first, and other than spinning at a particular rate. Hitting the upper stage before the release conditions are stabilized hurts those ways too.

    Mark Gubrud above repeats a common fallacy about hit to kill impact dynamics. At orbital speeds, any impact is an explosion – Even a stationary pop-up kill vehicle hitting a 6 km/s missile body is 18 Mj/kg, or 4.5 times the energy equivalent of TNT. Even the thinnest fuel tanks have enough mass density to convert the entire kill vehicle to expanding gas and small fragments, and that cloud expanding at something like the impact velocity then hits the far side of the tank with much higher surface area involved. These are not punched holes. They’re explosions, with KE playing the role that chemical energy does with conventional explosives.

  7. Mark Gubrud

    Lehner does make one point that Postol and Lewis will have to reply to. He says they left out 5 out of 6 successful warhead kills in tests with “separating warheads” as opposed to the 10 tests they analyzed in which the interceptor was supposed to pick out where the warhead was while still attached to the body of the missile. Why did they leave these tests out? I suspect the answer is because P&L’s whole point is the inability of the kill vehicle to pick out and correct to the correct aim point in the short time it has when the sensor returns data useful to determining where that is. Having a separated warhead, with or without MDA’s let’s-pretend “decoys,” actually makes the interceptor’s task easier. If all Iran or North Korea needs to do to defeat SM-3 is keep the warhead attached or at least shrouded the whole time, life is easy for them.

  8. Mark Gubrud

    George, I am aware of the energy density at orbital speeds and agree that the impact can be described as an “explosion” but your description of the process still leaves me puzzled how a significant amount of the KV’s momentum gets converted into perpendicular shock momentum given that the gas and debris cloud passes through the rocket at orbital speed and then is gone. Assuming a 45 degree debris cone, how much larger, really, is the area on the far wall (assuming that particular geometry) that the debris impacts and passes through, than the hole it left in the near wall?

  9. FSB

    ah….why is there a rocket body attached to the warhead in the first place?

    for ICBMs at midcourse there would be just the warhead.

    when do we see how good the SM3s are at hitting warheads when decoys are around, without the HUGE rocket body added to make the problem MUCH easier.

    completely UNrealistic testing.

  10. anon

    Good interview w/ Postol:

  11. yousaf (History)

    Quite apart from the technical flaws with nuclear missile defense, there are also serious conceptual problems. The following articles by Pavel Podvig, Panofsky and myself address some of these conceptual issues:
    The False Promise of Missile Defense
    “…it will be tough for Obama to let go of missile defense because until now, the discussion has been framed in such a way that it’s implicitly assumed that missile defense is a fundamentally useful thing—as long as it can be made efficient and built at reasonable cost without damaging the prospect for nuclear disarmament, of course. In short, the argument often is that the current missile defense system is flawed, but if those flaws could be solved, missile defense would be a great boon to international security. Missile defense proponents, of course, go much further, stating that missile defense provides a reliable (and some insist the only) way to counter emerging missile threats. The notion of missile defense as a good thing even entered the nuclear abolition debate earlier this year, with many experts ready to grant it a useful role at the final stages of disarmament, arguing that it might provide protection against those who cheat the system and attempt to build/use a nuclear weapon.
    The fundamental problem with this argument is that missile defense will never live up to these expectations. Let me say that again: Missile defense will never make a shred of difference when it comes to its primary mission—protecting a country from the threat of a nuclear missile attack. That isn’t to say that advanced sensors and interceptors someday won’t be able to deal with sophisticated missiles and decoys. They probably will. But again, this won’t overcome the fundamental challenge of keeping a nation safe against a nuclear threat, because it would take only a small probability of success to make such a threat credible while missile defense would need to offer absolute certainty of protection to truly be effective.”

    The Continuing Impact of the Nuclear Revolution
    “Defense of the nation, however well-intentioned, cannot be achieved by scientifically unsound means. President Bush should reconsider his approach to national missile defense and await the outcome of a balanced and thorough analysis of the fundamental issues.”

    The Myth of Missile Defense as a Deterrent
    “the U.S. infatuation with missile defense will cause other nations to desire this expensive and destabilizing technology. Following the U.S. lead, both China and India now have missile defense test programs. It doesn’t take much imagination to anticipate Pakistan’s response. There will be legitimate pressure for Islamabad to attempt to redress this perceived Indian defense by producing more missiles and nuclear weapons. In response, India and subsequently China will likely increase their own stockpiles—in turn increasing pressure on U.S. and Russian strategists to respond. So rather than reducing the value of nuclear weapons, missile defense actually increases it.”
    “…there’s now an urgent need for an informed, unbiased reappraisal of U.S. strategic thinking on the conceptual basis of nuclear missile defense policy.”
    What Missile Defense?
    “missile defense comes in two different flavors. The first is tactical missile defense, such as the U.S. Patriot system, which protects a theater of battle against short-range conventional rockets. The second category is strategic, or national, missile defense: systems meant to guard against adversaries’ nuclear-tipped missiles. While the first of these types is conceptually sensible, the second is not and may even make the world a more dangerous place.
    The reason for this is quite simple. A 70 percent effective tactical missile defense (to pick an optimistic number) makes a lot of sense. If 10 conventional missiles are headed your way, stopping seven is undeniably a good thing. Stopping seven of 10 nuclear warheads, however, is less decisive since even one will visit unacceptable devastation upon the United States. Just one nuclear-tipped missile penetrating your missile shield is about the equivalent of a million conventional missiles making it through.
    So even after the United States has set up and activated a national missile-defense system, it still will not have neutralized the perceived threat from Iran. Not only that, but Washington’s strategic calculations toward Tehran will remain unaffected: The United States will still need to be just as worried about Iran’s missiles, since the destruction of even one U.S. city or region is simply too high a cost to bear. For that security equation to change, national missile defense would need to intercept 100 percent of incoming nuclear warheads — an unattainable goal for any piece of machinery.”

  12. Mark Gubrud


    I’ll agree that some of the impacts P&L graded as misses were pretty close, but let’s consider your scenario: a side impact at several km/s on a tank section of the rocket.

    The areal mass density of the KKV is maybe 30 times that of the first wall. So let’s consider the situation when the KKV has penetrated the rocket wall about 3%, so that about 3% of the mass of the KKV is now violently mixed with 100% that of the wall through which it is passing. The velocity of this hot mix of gas and debris is already the average of that of the rocket and KKV, i.e. it is moving into the rocket body at half the relative speed of the KKV. In the next instant, it mixes with the next 3% of the KKV, and now it is moving at 75% of the KKV’s speed. And so on. The debris, gas, etc. basically gets pushed ahead. A shock wave is propagating into the KKV body, but probably slower than, or at comparable speed to that with which the KKV and debris is proceeding into the rocket.

    As you point out, the debris spreads out as it passes through the rocket and hits a somewhat larger area on the far wall, but only a few times larger, even assuming a 45 degree cone (expansion speed equal to KKV velocity). There the same process repeats, and the debris passes through, out and away from the rocket and warhead, rather than the picture of an explosion propagating up toward the warhead.

    A shock can propagate along the rocket body, but the coupling that generates it occurs only at the edges of the holes left by the KKV and debris cone, and this shock is then launched into the edge of a thin sheet metal structure which is not designed to transmit or withstand high-G accelerations.

  13. John Schilling (History)


    Generally speaking, you are making the mistake of using a zeroth-order and mostly qualitative approximation of a process that starts with about four orders of magnitude more energy than necessary to achieve the desired effect. “Assuming a 45 degree cone”, in this context, only works if you are prepared to claim that 100.00% of the effects of the primary impact will be confined to that cone.

    And even with that dubious assumption, you err when you say the far-wall impact is “only a few times larger” than the primary. The primary impact site will be on the order of one KKV diameter, the far-wall impact on the order of one tank diameter. That’s two orders of magnitude difference in area.

    A bit more detail, assuming an SM-3 KKV hitting a Titan II upper stage, dead-center perpendicular at 12 km/s. The primary impact will be on a 20-cm diameter spot, and will result in A: about 19.5 kg of undamaged KKV proceeding forward at about 11.7 km/s, and B: about 1 kg of mixed tank wall and KKV body remains (mostly singly-ionized aluminum plasma) expanding uniformly from the KKV leading edge at an RMS velocity of about 4.5 km/s.

    That leads to an average cone half-angle of 22.5 degrees – note “average”; some of the debris will be ejected 90+ degrees off-axis and may hit the warhead directly. But we’ll simplify and assume a uniform debris cloud – which hits a 2.5m diameter spot on the far wall, at 11.7 km/s average velocity, over a period of 200 microseconds. That’s 1 kg of plasma vs. ~70 kg of tank wall, which during the impact period can only move the tank wall ~1.7 cm. The tank wall will be converted to a debris cloud ejected mostly-harmlessly to one side (at ~350 m/s), but during the actual secondary collision period will still form a barrier to the impacting plasma.

    So, 1 kg of plasma brought to a dead stop from 11.7 km/s, joined by ~1.4 kg of vapor ablated from the far wall during the collision, and reexpanding uniformly at ~5.5 km/s RMS velocity.

    This will have an effect roughly equivalent to detonating 10 kilograms of TNT, four meters from the warhead. Or, looked at another way, it will overpressurize the remaining tank volume to ~110 psi, fragmenting the entire upper stage and driving the forward-bulkead fragments into the base of the RV.

    Hardened warhead with a thin standoff plate (the forward tank wall) vs. 110 psi blast overpressure. It isn’t obvious who wins that contest, especially since even this version is a very simple first-order approximation of one kill mechanism. But I’m not betting on the warhead.

    And as I said earlier, I’d prefer not to bet at all but to work with the results of more detailed analysis and better still actual test results.

  14. Mark Gubrud


    I agree that you describe a plausible mechanism for coupling a significant fraction of the energy thermalized in the first impact into an “explosion” that propagates forward – namely that the plasma accelerates out and ahead from the first impact and then impacts secondarily on a much larger area, so that it is “brought to a stop” and scattered in the rocket frame. Whether this kills the warhead is, as you say, not obvious. But a significant amount of blast and debris could be generated in the rocket’s forward direction.

  15. FSB

    Mark and John:
    …and none of this applies, of course, to ICBM warheads which will not have a HUGE rocket body attached to make it easy to hit, and can never, realistically, be discerned from actual warheads if decoys are present. (Read interview with Postol given by anon above).

  16. Mark Gubrud


    MDA counts 5 out of 6 hits with separating warheads. This may mean that, ironically, in the absence of real decoys separation makes the KKV’s task easier: even though the target is smaller, the KKV can hit it, and it doesn’t need to figure out where (i.e. at one end rather than on the centroid) on that little blur it needs to hit a second later.

    But of course, as soon as you add real decoys…

  17. dd (History)

    Closing all the loops and solving all the problems to hit even the upper stage of a target is a hell of an accomplishment. Rome wasn’t built in a day. This is Technology Development. The critics seem to be saying that since they’ve only solved 95% of the problems of hitting an incredibly hard to hit target, they should just stop and give up now.

    If you don’t like the concept or the politics of missile defense, fine, argue those points. But arguing that missile defense is somehow bogus because the technology isn’t perfect yet is lame.

    There were multiple sequential failures of the early Corona launches, therefore space-based reconnaissance was bogus and should have been cancelled, I guess?

    I also know first-hand, in a sort of “economic stimulus” argument, that the sensor, processor, rocket, and control theory tech development attacked by MDA to solve this problem pays off across our nation’s industrial base and economy.

    And finally, I’m not sure how you get more ethically pure in the DOD than working on purely defensive weapons that aren’t designed to kill anyone, and solely meant to protect a populace. Ethically speaking, this isn’t exactly Napalm or Agent Orange they’re working on at MDA. With all the Military-Industrial Complex things out there to critique or complain about, I’ve never quite understood the stinkeye directed at missile defense. As “stretch goals” go, there are worse ones than this.

  18. Anne Bazuin (History)

    When two bodies collide, the resulting situation depends of the following conditions prior to collision:
    Magnitude of speed and direction of speed of the bodies.
    Mass of the bodies.
    Rigidity of the bodies.
    Assume following conditions:
    A kill vehicle (rigidity R1) and a rocket with a warhead mounted on top (rigidity R2) are both in exo-atmospheric ballistic curves when they collide.
    The kill vehicle (mass M1) hits the rocket (mass M2) with a speed of 6 kilometer per second perpendicular in the side.
    Due to the enormous speed of the collision, a piercing or cutting, with little fragmentation, is most likely. The most rigid parts of the bodies will tend to remain relatively in shape, that is the hard core of the kill vehicle and the warhead.
    Post to collision, the kill vehicle and the warhead (if not exploded due to the shock of collision) will continue to move in ballistic curves different from their curves prior to collision.
    The kill vehicle’s hitting capacity depends on the sensivity of his infra red sensor or radar sensor, the specific program of his on board computer and the speed of his maneuvering system. In the last second prior to colission, that is on a distance of 6 kilometer, his system has to react very fast and with extreme selectivity.
    Results of interception tests published until now, show little chance that the kill vehicle hits the warhead precisely.
    The flying time of a ballistic missile from Iran to Israel (and from Israel to Iran) is ten minutes only. There is little time between detection of the start of the rocket bearing a warhead and the launching of a interception rocket with a kill vehicle. And there is little chance of a precise hit.
    An e Bazuin, 14 june 2010

  19. Anne Bazuin (History)

    Now consider the following conditions:
    The kill vehicle hits the rocket with warhead precisely frontal in a straiht line.
    Both have a speed of 6 kilometer per second and have almost equal mass.
    The result will be a total fragmentation, and the smallest particles may be only a few molecules large.
    The rigidity of the bodies plays no role, nor the question wether the collision is elastic, plastic or somewhere in between.
    This is my favorite hit.
    Anne Bazuin 16 june 2010