Jeffrey LewisDo We Need ICBMs?

The Commander of the US Strategic Command, C. Robert “Bob” Kehler recently offered support for the Triad — the trinity of bombers, land-based missiless and ballistic missile submarines — that some people have chose to characterize as tepid.

Now, admittedly by “some people” I mean two interns at the Heritage Foundation, one of whom now writes for Air Force Magazine.

Nukes of Hazard has posted Kehler’s remarks on the triad. As I suspected,  they are essentially the same as the ones he made in October 2011 and May 2012. To most observers, Kehler has continuously expressed support for the triad, not opposition.  Kehler’s heresy was to note that his support for the triad is contingent and immediate, subject to adjustment in light of changes to the strategic environment.   (Some people take insufficient enthusiasm very seriously.  It’s a good thing Kehler doesn’t command the KPA Strategic Rocket Forces because he’s got a thing or two to learn about sugar-coating things for 20-somethings.)

Rather than heresy, Kehler’s defense of the triad.  has been, in fact, quite orthodox.  He actually made the canonical case for the ICBM-leg of the triad, based on promptness:

And what the ICBM force gives to the president is the ability to respond promptly. I think that’s still a valuable component of the range of alternatives that we could offer to the president.”

This has always been the fundamental justification for keeping some portion of the ballistic missile force based on land.

The statement left me wondering: Are ICBMs really more responsive than SLBMs?

The idea that the President can communicate with an ICBM launch facility more quickly than a submarine at sea is one of those enduring notions within the strategic studies community.  Heck, I’ve even written it.

After a while, I started to wonder.  Is there any data to support that claim at all? In trying to answer that question, I discovered a very interesting little study.

In 1990, as the Warsaw Pact crumbled, eventually taking the Soviet Union with it, Congress asked the then-General Accounting Office to examine the triad.  GAO did something very unusual — research — and produced a report in 1992.  GAO also made available an unclassified summary that explains:

This is … the first study in at least three decades that both sets up a comprehensive framework for comparing numerous, dissimilar strategic systems on multiple measures and that uses test and performance data to compare the systems in question.

Moreover, we did not simply use the performance data cited by DOD for these systems, but instead validated those data through extensive research, analysis and questioning about the underlying validity of the data, including the quality and quantity of the testing of each system. For systems where we found that the claimed performance could not be sustained by such analysis, we have made the uncertainties explicit.

It turns out the claim that ICBMs were more prompt than SLBMs was just something people said. It seemed reasonable, of course, but there was no data.  NoneNadaZip.  When GAO looked at all the relevant data, including patrol data for ballistic missile submarines, they found little difference.  The two systems were essentially equal in terms of responsiveness:

Test and operational patrol data show that the speed and reliability of day-to-day communications to submerged, deployed SSBNs were far better than widely believed, and about the equal of speed and reliability of communications to ICBM silos. Contrary to conventional wisdom, SSBN are in essentially constant communication with national command authorities and, depending on the scenario, SLBM from SSBNs would be almost as prompt as ICBMs in hitting enemy targets.

Reading this stunned me.  The idea that ICBMs have an advantage in promptness is one of those bits of conventional wisdom I’ve never seen challenged.  To see GAO say “nope, that’s wrong” is one of the more disorienting experiences I have had in a while.

Now, obviously, the US has incrementally improved the command-and-control system for land-based ballistic missiles over the past twenty years — though not nearly so much as one might hope.  The Navy, too, has been improving command-and-control systems for Trident, including development of a capability to retarget SLBMs.  It is hard to compare improvements in command-and-control without the sort of data GAO used, but when STRATCOM proposed conventionally-armed Trident ballistic missiles, STRATCOM officials seemed pretty dismissive of concerns about how difficulties in communicating promptly with submarines.

That’s bad news for the ICBM force.  Although I think silo vulnerability is vastly overstated, land-based missiles just can’t compete compete with submarines, even with really expensive mobile launchers.  And then there is the issue of overflight.  It is an unfortunate reality that Russia happens to lie between the United States and most of all the interesting places we might someday want to nuke.  The Unites States is extraordinarily unlikely to light up the Russian EW system with one or more nuclear weapons en route to someplace else.  That leaves a prompt response to a Russian bolt-from-the-blue as all the ICBM leg has left. Well, that and the Senate ICBM caucus.

Bad news, if it is true, I mean.  Someone should do a study!  Actually, not just someone.  The Government Accountability Office.  Congress ought to ask GAO to revisit the 1992 Triad study making use of the same data-rich methodological approach.  It would be very interesting to see whether the promptness argument holds up to scrutiny.

Update | I’ve edited the post because it turns out I had read Kehler’s remarks on Nukes of Hazard this morning — they were just so close to remarks in May that I momentarily thought I was conflating the two.  I probably should have done the strike-through thing, but this is how I wanted the post to read in the first place.


  1. Bradley Laing (History)

    If someone did a war-game this month, based on the classified data only Presidents are likely to have, how big would be the response time gap between the submarines and the missile silos?

    And, not to be missed, if someone asked Putin’s generals to do another war game, based on what they think they know about US subs and US silos, what would they say would be the response time gap between the submarines and the missiles silos?

    Thirdly, I’ve read that during WW II, Navy regulars would find they were in an unknown situation and say “Captain to the Bridge,” delaying decision making until a higher authority got there. The Navy *reservists,* called up for the war, in the same position, would make a decision on the spot. The communication system work perfectly in both cases. Either you called for the captain, and delayed the decision, or you didn’t. But one method delayed decisions, and the other did not. You do not need a specific kind of radio or landline communication system for internal decisions to delay action.

    • Jeffrey (History)

      You have GAO’s answer based on classified performance data for both systems: SLBMs were “about the equal of speed and reliability” as ICBMs.

      They are the essentially the same.

    • Bradley Laing (History)

      But that was 23 years ago. Do we know that a change in some form has not sped up one method? (Fiber optic cables moving messages faster than previously, for example?)

  2. George William Herbert (History)

    My end concern with eliminating the ICBMs is advances in ASW.

    There are things areond the edges of what we can do now – adversaries less so, but a focused US or western ally program could achieve – which would make finding even deep slow subs easier. Possibly much much easier.

    It might take fifty or more years, or never develop in the practical real world, but if science allows it and technology dev is heading in supportive directions, it could be real and not that far away in long term planning cycle timescales.

    Of course, precision guided conventional ICBM warheads would do a number on any credible hard silo on land, too, so keeping an eye out is an all around problem…

    • Allen Thomson (History)

      My end concern with eliminating the ICBMs is advances in ASW.

      There are things around the edges of what we can do now – adversaries less so, but a focused US or western ally program could achieve – which would make finding even deep slow subs easier. Possibly much much easier.

      Care to expand on that?

      I spent a number of years back in the day engaged with efforts to understand advanced submarine detection techniques. Mostly they had to do with the possibility that various submarine-caused hydrodynamic disturbances, some of them not well understood, would propagate to the surface where they could be detected by radar, IR or whatever. The upshot, as far as I could see then, was that shallow-running submarines, like SSNs in the Persian Gulf or Formosa Straits, definitely had something to worry about. For SSBNs lurking out in the vasty deeps, not so much so — the famous Seasat Rev 407 image notwithstanding.

  3. Kenton A. Hoover (History)

    23 years ago, the US had an ELF “bell ringer” network that allowed SSBNs to be told to come up to VLF depth to get an EAM. The US ELF system was closed in 2004, so it might indeed be worthwhile to have GAO update the study. We do have almost global VLF coverage.

    In re ASW improvements, perhaps, but you don’t really know that such things will work unless you actively work to target SSBNs in wargame et al and have visible successes. Otherwise we’re back to a 60/70/80s hypothesis about a decapitation strike and that seems a bet with long odds. And if you saw signs that ASW was an issue you could easily put Minuteman back into production or perhaps even figure out how to put Trident into land silos.

    • Jeffrey (History)

      It has always seemed more likely that the Russians would figure out how to hold at risk silos before submarines.

      The GAO study is old, as I noted in my post, with C3I for both land- and sea-based legs receiving incremental improvements. Obviously, as I said, one cannot be sure without the sort of data GAO had, but we have some hints. I observed that STRATCOM officials dismissed the “promptness” objection to placing conventional warheads on Trident. That would suggest, to me, that the difference probably remains slight. Someone should do a study!

  4. John Bragg (History)

    I’m a bit confused. I thought that the clinching argument for ICBM’s in an arms-controlled environment was as a “warhead sink”? Basically, each ICBM soaks up one warhead that would otherwise be targeted at a US city.

    If the US has 100 or 300 or 450 ICBMs sitting out in North Dakota, then an adversary has to target that many silos to prevent a US second strike. (Assuming that he has mermaid commandos or something to take out the SSBNs).

    • Jeffrey (History)

      There are always arguments — insurance against an ASW break-through, warhead sinks, etc. (I’ve always hated the warhead sink argument since some SSBNs are always in port and those ports in much closer to urban areas than some god-forsaken wheat field in North Dakota. But, yes, people do make that argument.)

      I have always understood promptness to be the primary advantage. That is certainly the view expressed by GAO in 1992.

      In the mid-1990s, when the Clinton Administration considered the issue, they emphasized promptness first and foremost:

      “ICBMs provide the United States a prompt-response capability. START II requires the downloading of ICBMs to one warhead, but does not place a sublimit on the total number of single-warhead ICBMs. Approximately 500 Minuteman IIIs will be retained and downloaded to one warhead apiece. ICBMs also increase the cost ratio to an adversary of attempting a first strike. Retaining approximately 500 single-warhead Minuteman IIIs provides for a reduced but prudent ICBM force.”

      Lately, Kehler has been making the “survivable, prompt and flexible” argument outlining a specific strength of each leg of the triad.

      Perhaps the simplest way to make the point is to argue that a dealerted ICBM force could meet both goals — a hedge against an ASW breakthrough and serve as a warhead sink. But most proponents of the ICBM force believe that dealerted ICBMs will ultimately be eliminated. I would argue that is because, at the end of the day, the “promptness” is their bread-and-butter.

    • David Forscey (History)

      That logic makes sense in an era with limited warheads, but it falls apart if the adversary can simple manufacture more, right?

  5. Stephen Schwartz (History)

    And then there’s the fact that we’ve never really launched a fully operational ICBM from an operational silo. From Atomic Audit: The Costs and Consequences of U.S. Nuclear Weapons Since 1940:

    “In the 1960s, to allay concerns that operational missiles might not launch when needed, SAC initiated Project Long Life. On March 1, 1965, a Minuteman I missile was launched from its silo at Ellsworth AFB in South Dakota (the second and third stages were inert and the first stage was tethered to a long elastic cord). The flight lasted seven seconds and covered just 2 miles (3.2 kilometers). However, subsequent tests with the Minuteman II on October 19 and 28, 1966, and August 14, 1968 (the latter named Giant Boost), were unsuccessful, leading to a new effort in 1970 to launch eight fully operational Minutemen from a SAC base. This program never got under way because members of Congress over whose states and districts these missiles would fly declined to authorize it.”

    • Stephen Schwartz (History)

      By the way, according to the 1992 GAO report, “as of 1981, the Soviets had launched more than 200 ICBMs from operational silos.”

  6. Daryl Press (History)

    I think the “warhead sink” argument is stronger than John B. suggests. In some possible nasty future, someone thinking of destroying silos with nukes will probably need at least 3 per silo to reliably destroy a few hundred silos — regardless of the accuracy of the weapons (because I presume reliability issues will endure). That is, they’d need to use 3:1 at least if they want to have a chance of going 450 for 450, or something close to that. Hence, a single-warhead ICBM leg becomes nearly invulnerable to nuclear disarming attack unless the enemy has many times your force, or they find some other way of neutralizing the leg without plinking the silos, or they develop another weapon that can significantly contribute to the attack without using up nukes.

    On top if that, there’s the risk of flaws in single systems which could take elements of the force offline for a time — e.g., a flaw in the SLBMs or the SSBNs themselves.

    The reason this is difficult to think through, in my view, is that we’re making these assessments in the current, rather benign international environment, but we’re talking about systems that are supposed to suffice for many decades even if the world gets much nastier. Sure we could theoretically eliminate systems and reintroduce them later but that might grow increasingly difficult politically, and in some cases strategically dangerous.

    Last, re: Gen Kehler’s argument for “survivable, prompt, and flexible”…if one believed all three of those characteristics mattered (as I do), wouldn’t you also want at least one of the systems in your arsenal to embody all three characteristics? Which leg of the triad would that be today?

    • Daryl Press (History)

      oops — to clarify my question at the end: I’m not implying that the answer is “the ICBM force.” I think the answer is: “none of them.”

    • Jeffrey (History)

      Do you find SSBNs inflexible? Because they are very survivable and, apparently, quite prompt.

    • Daryl Press (History)

      Your post on SLBM promptness is interesting — I too had always assumed (and have had it suggested to me on occasion) that prompt communication to subs is trickier than the data you uncovered suggests.

      On flexibility: I think range-of-yield is a key part of flexibility, which is (if open-source reporting is correct) lacking on the sub force. (i.e., having two yield options — really big, and even bigger — is not sufficient, in my view.) My sense though is that developing the yield flexibility for the SSBN force that one has, for example, in the existing bomber force would require a major upgrade to SLBM accuracy (in three dimensions) to permit low yield warheads to be used effectively. That’s what I meant when I said that policy makers who value promptness, survivability, and flexibility might wish to have them embodied in a single system — and don’t.

      But smarter people than I have thought really hard about this — and some of those people value the same force attributes that I laid out, and some of those people are satisfied with the existing force. Hmmm. One can connect some dots, perhaps, about some capabilities that might currently exist in the sub force which are nowhere discussed in the open-source. Perhaps.

    • Alex W. (History)

      My intuitive reaction to the “warhead sink” argument is that it doesn’t seem like it takes many nuclear weapons targeted at cities to make life really, really awful.

      Maybe I’ve just played DEFCON too many times, but it seems like tossing a spare ICBM at New York would be an easy thing to do in a full-scale exchange, no matter how many you’ve got targeted at bases in Montana.

    • Allen Thomson (History)

      My intuitive reaction to the “warhead sink” argument is that it doesn’t seem like it takes many nuclear weapons targeted at cities to make life really, really awful.

      Maybe I’ve just played DEFCON too many times, but it seems like tossing a spare ICBM at New York would be an easy thing to do in a full-scale exchange, no matter how many you’ve got targeted at bases in Montana.

      Well, yes. And even the Montana (and North Dakota and Wyoming) attacks would lead to melancholy circumstances elsewhere, as they’d probably be 2-on-one surface bursts of multi-hundred kiloton bombs.

      Once again calling on my favorite study of such things, NAPB-90, see

  7. kme (History)

    One thing to consider is that the overall arrival time of a warhead at a target is affected by both the time for the launch order to be recieved and by the flight time of the missile to the target.

    Even if it takes a bit longer for the order to launch to make it to a submarine, they might well make up the difference by being closer to the target when they launch.

    • Jeffrey (History)

      By guess — but it is only a guess — is that might be why the patrol data mattered for GAO.

  8. Allen Thomson (History)

    Is there any discussion around that explains what (besides, I suppose, launch on tactical warning in the face of a counter-force attack against the ICBMs themselves) makes promptness a valuable property of ICBMs? And how prompt does prompt have to be for the scenarios that need promptness? Five minutes? Fifteen?

    • Jeffrey (History)

      I actually wrote, then deleted, a paragraph on promptness. I keep pointing out that no one complains that the military campaign in Afghanistan wasn’t prompt, even though it took something like 28 days.

      There are specific scenarios in a nuclear exchange that I can imagine would give rise to various definitions of prompt – launch under attack as a survivability measure, holding at risk fleeting targets by barrage, etc.

    • Daryl Press (History)

      Wouldn’t promptness matter in retaliatory counter-force strikes? (i.e., to prevent further attacks?) I’m not principally thinking of a peer or near-peer, but rather a regional actor with a small force.

      Perhaps the ICBM is not the best weapon for that mission — for the overflight reasons Jeffrey noted among others. But in my mind, that merely underlines the point about the existing arsenal seemingly lacking a system that is prompt and flexible. (ICs and SLs are the first but not the second; bombers are the second but not the first).

  9. Peter Hayes (History)

    What matters is relative promptness (compared with adversary) and the nature of the coupling between deployed forces and the command center on each side. During the Cold War, in case of US, we had decentralized nuclear forces in unified and sub-unified regional commands constantly interacting with periphery of highly centralized SRF of FSU. An error in our system (a submarine probe going aground, or the KAL007 incident, for example) was dampened by the decentralized nature of the US C3I system; bu the same error led the FSU command to respond instantly and to propagate the error, leading to enhanced risk of global nuclear war. What matters is how the two systems interact, not how one or the other behaves–including in the promptness dimension. Des Ball and Paul Bracken (and to some extent, folks at Mitre Corporation) approached this issue in the 1980s; but to my knowledge, no-one analyzed its implications for the risk of nuclear war in any depth.
    With regard to Stephen Schwart’z comment above about the limited experience in firing missiles from land-based silos–and the fantastic nature of the scenarios involved and unknown system-reliability of such capacities–to my knowledge the US also never fired land-based long-range missiles in a realistic north-south mode that took into account the Coriolis effect, Earth’s oblateness, geomagnetic anomalies that would affect accuracy of missiles, etc. We know we can nuke Kwajalein in an east-west trajectory; or the Azores Islands; but fire accurately over-the-poles, either pole…we really have no idea. The MX tests launched from Vandenberg that were to provide precisely such experience were cancelled in 1985 when Australia’s peace movement mobilized to force the Labor Government to reverse its provision of transit airfelds for US ARIA aircraft that were to collect the telemetry from the incoming RVs from the sono-buoys in the Broad Ocean Area test sites (without which one had no measure of accuracy of splashdown points, mid-ocean south of Tasmania and New Zealand). This resulted when yours truly published an FOI-released document 1in 1984 on the pending tests that led now-Ambassador to US then-Defense Minister Kim Beazley to ask how come no-one including him knew about the tests, proving thereby that sometimes, the pen can be mightier than the missile.
    chapter 13

    • Stephen Schwartz (History)

      “…to my knowledge the US also never fired land-based long-range missiles in a realistic north-south mode that took into account the Coriolis effect, Earth’s oblateness, geomagnetic anomalies that would affect accuracy of missiles, etc.”

      The same holds true for the Soviet Union/Russia.

      James Fallows provides a clear and succinct explanation of these problems in his 1981 book National Defense (pp. 144-57). He also argues that the errors these factors could generate in inertial guidance systems would create significant uncertainty as to the odds of success of a first strike on ICBMs by either country, dramatically reducing confidence that such attacks could be effective and thus diminishing the likelihood they would ever be attempted.

      Other useful sources:

      Matthew Bunn and Kosta Tsipis, “The Uncertainties of Preemptive Nuclear Attack,” Scientific American, vol. 249 (November 1983), pp. 38–46; and Andrew Cockburn and Alexander Cockburn, “The Myth of Missile Accuracy,” New York Review of Books, November 20, 1980, pp. 40–44.

      This quote from Dick Garwin (in the aforementioned NYROB article) is especially good:

      “In every ICBM you have an inertial package. Accelerometers and gyros and things like that are mounted in your missile. You’ve got to fire your missiles from operational silos to points in your enemy’s country. Now, obviously you’ve never done this before and so you have to base your calculation on test shots—in our case from Vandenberg to Kwajalein lagoon, that is, east to west; and in the Russians’ case from northern European Russia to Kamchatka in the northern Pacific, west to east. Judging from how far each test shot falls from the target, you adjust your accelerometer or your gyro, to compensate for the inaccuracy, until in the end your test shots are landing within the prescribed area. But every time you fire a new-model missile over the same range or the same missile over a slightly different range, the bias changes. Sometimes it is greater, sometimes it is smaller, but it never has been calculated beforehand.

      So you have to go back to readjusting-the gyros and so on, to try and eliminate the novel bias. But if we were firing operationally, both we and the Russians would be firing over a new range in an untried direction—north. And a whole new set of random factors would come into play—anomalies in the earth’s gravitational field, varying densities of the upper atmosphere or unknown wind velocities. They may adjust and readjust in testing and eventually they might feel sure that they have eliminated the bias. But they can never be absolutely certain. We certainly cannot be: and although we are less well informed about the Russian ICBM test program than our own, there is no reason to suspect that they are any more successful than we are at dealing with the problem. If you cannot be sure that you would be able to hit the enemy’s silos, then there is no point in even trying—because the idea is that one side could wipe out the other’s missiles before they are launched in a first strike.”

  10. Tom Nichols (History)

    All of these comments focus on the relative capability of ICBMs as offensive weapons. That misses the point about land-based ICBMs entirely: their value isn’t that they are ICBMs, it’s that they’re land-based, and specifically by “land” I mean “the territory of the United States.”

    I used to tell my students that ICBMs were more reliable than subs, too. Now, I think that’s probably not true, but more to the point, it’s completely irrelevant whether Minuteman responds marginally faster or slower than SLBMs, or whether it presents more alternatives for operations. The deterrent value lies not in what the land-based force does, but in where it *is*. An enemy wanting to take a shot at eviscerating the U.S. nuclear deterrent is going to think a lot harder about an attack on CONUS, which is a far hairier proposition than trying to sink submarines or suppress bombers (which are really the leg of the triad that no longer make sense, actually).

    If you really intend to cut the strategic force safely, and really *mean* it when you say that those forces are a last-ditch, deterrence-only, we-are-not-f*cking-around kind of weapon, then leave them in the United States, and force an enemy who’s even considering a nuclear strike to really put the Big Boy Pants on and think hard about killing millions of civilians instead of thousands of submariners. Cooler heads will prevail if we remove the option of Tom Clancy-ish wet dreams of hunting boomers on the high seas during a crisis.

  11. David (History)

    There are two pieces to promptness. The communication piece IRT how quickly an LCC or SSBN can receive a message. The focus of Dr. Lewis’ post is on that piece. The other piece is how quickly a weapon can be placed on a target. I believe the ICBM has an advantage there, no?

    • Moe_DeLaun (History)

      Not necessarily. A foward-deployed SSBN could launch a salvo on depressed trajectories from within a few hundred miles of an opponent’s coastline and leave the defenders very little time to react. This in essence is how the SSGN’s are being used operationally (e.g. USS Florida (SSGN 728) during Odyssey Dawn.)

    • Moe_DeLaun (History)

      In fact, this is what Sea Launched Prompt Long-range Strike (SLPLRS — “Sliplers”?) is all about, and would be one element of Air-Sea Battle. Use a conventional SLBM to target a mobile ASBM launcher, for example. Of course, by the time we’ve sailed into that bay, the sky’s on fire…

  12. Alex W. (History)

    Just an observation: we seem to be talking about the importance of the Triad, of the need for “warhead sinks,” of the relative responsiveness, as if there was some genericized enemy out there that had the capability and interest in trying to pull a first-strike attack.

    Is the “enemy” Russia? It’s hard for me to see who else fits the profile. Does anybody really think that they’re interested in pulling that sort of thing in this day and age? China doesn’t seem to really fit the profile, strategically.

    All of which is just me saying that as a naive outsider, all of this worrying about theoretical deterrence models with first-strike anxieties seems a little out of touch with which century we happen to be living in, and bizarrely abstract given that the number of countries that fit the threat profile can be counted on basically one, maybe two fingers.

  13. Mark A. Gubrud (History)

    Seems to me the first “leg of the triad” to cut off is the manned bomber, before a next generation appears which turns out to be an unmanned nuclear-armed drone, either remote controlled via vulnerable satellite links, or potentially enabled to autonomously select among targeting options, as human pilots are sometimes expected to do.

  14. Jon (History)

    I am sure with current technologies that promptness of response between Minuteman III and Trident II are not significant. I see the value in promptness with the icbms lies within the ability to use them before they are destroyed. This is really the only significant aspect of their promptness. Our icbms are to deter Russia. So if Russia were to attack our icbms with their icbms, we could launch ours before they are destroyed, which means Russia is wasting their icbms on empty silos and the US in response has just destroyed 400-420 targests on Russian soil.

    • kme (History)

      This seems perilously close to saying “the value of our ICBMs is in deterring an attack on our ICBMs”.

    • Jon (History)

      No, not at all. It just means our icbms are survivable against a first strike.

  15. Stephen Schwartz (History)

    By the way, Jeffrey, the full GAO report has much more to say about promptness and communications with regard to ICBMs vs. SLBMs:

    – “Assuming strategic warning, and compared to ICBMs, no operationally meaningful difference in time to target [for SLBMs] was found.”

    – “The silo LCCs and alert, submerged SSBNs are both in continuous, instantaneous communication with the national command authority (NCA). … Both LCCs and SSBNs have equipment to constantly monitor transmission paths to detect outages.

    …Another measure of C3 performance is how quickly SSBNs and silo LCCs can receive Emergency Action Messages (EAMs)—nuclear attack orders. (The speed measure inherently includes reliability, since a garbled EAM is not considered to have been received.) According to 1988-1990 data from ‘POLO HAT’ and other C3 operational tests, [REDACTED] compared to an average of about [REDACTED] minutes to SSBNs in the Atlantic, and about [REDACTED] minutes to SSBNs in the Pacific (with the longer time apparently due to more relays). Under conditions that assume the availability of satellite transmissions—a more realistic scenario in today’s world—the average EAM transmission time to LCCs was 3.3 minutes, compared to [REDACTED] minutes for Atlantic SSBNs, and [REDACTED] minutes for SSBNs in the Pacific. These differences may further narrow under DOD plans to reduce the number of C3 aircraft for relays, by deploying fewer with more powerful transmitters. [NOTE: Elsewhere in the report, GAO writes, “In effect, under standard operating conditions, communications connectivity to alert SSBNs—about one third of the SSBN force—is essentially the same as that to silo LCCs, albeit through radio frequencies rather than land lines. Transmission of messages may take slightly longer (measured in seconds), due to the fact that the LCCs can receive real-tme radio voice communications, while submerged SSBNs receive signals via the very low frequency and low frequency bands. However, according to the Navy, there are more electronic links through which messages must pass between the peacetime NCA and the silo LCCs than there are to submerged SSBNs, with the result that messages are received at least as quickly by the SSBNs, if not more quickly.”]

    The longer average time period for EAM transmission to SSBNs is not a survivability problem for them because they would not be under attack, whereas every minute is critical in the case of the (locatable) silo LCCs, which must receive the EAM to be ready to launch the ICBMs prior to the detonation of incoming warheads…. [NOTE: Elsewhere in the report, GAO writes that “the Navy has implemented a ‘minimum reaction posture’ (MRP) program, to decrease SLBM launch reaction times, which includes sending a pre-EAM message [REDACTED]. This pre-EAM message permits the SSBN to begin launch procedures, helping reduce the launch time of the first SLBM to just [REDACTED] from the time the EAM is received. Preparations include: manning battle stations; spinning up the SLBM guidance systems; moving to launch depth and angle, and preparing to hover; and enabling the navigation sonar system.”]

    In operational terms, the relatively small average difference in EAM transmission times to LCCs versus SSBNs is not great, and certainly far less than generally understood. At the same time, it is clear that there is a difference, however small, between EAM transmission times to LCCs versus SSBNs, that could be taken into account in nuclear warfighting plans. Moreover, all ICBMs can be launched essentially simultaneously, while not all SLBMs can be launched at the same time from an SSBN [REDACTED]….”

    The report goes on to compare time-on-target between SLBMs and ICBMs:

    “GAO calculated the time-on-target (TOT) averages for SLBMs versus ICBMs by adding: the above-cited average EAM transmission times for simulated nuclear war conditions; the average launch time required for SLBMs versus ICBMs [REDACTED]. Adding these components together produced—for a ‘worst case’ surprise attack scenario—[REDACTED; 1.5 paragraphs].

    Whether these TOT differences are large enough to be operationally important is open to debate. According to a 1990 study by the Institute for Defense Analyses (IDA), any time difference less than two hours is not operationally meaningful, even in the case of ‘time-urgent’ targets. Other analyses contend that the very quick average response time offered by silo-based ICBMs contributes to deterrence, and that for some very time-urgent targets, the somewhat slower TOTs for SLBMs, on average, would make them less effective than ICBMs.

    Since GAO did not have access to nuclear warfighting details, these views could not be fully evaluated. However, GAO finds that the TOT differences between ICBMs and SLBMs are considerably less than is widely understood. Moreover, in the 1992 threat environment, it seems questionable that an SLBM, hitting a target [REDACTED] after an ICBM would, is a meaningfully less useful weapon, as some in DOD contend. Since it takes hours to reload Soviet silos, a [REDACTED] TOT difference makes no difference in terms of their reuse—which would be prevented by either an ICBM or SLBM. Similarly, assuming that the United States responds to an attack (rather than initiating or pre-empting one), it is difficult to think of an enemy leadership or non-silo military asset that would have to be destroyed in the first 60 minutes of a strategic conflict—as opposed to the first [REDACTED]—to critically affect the outcome of the conflict.”

  16. Allen Thomson (History)

    FWIW, this from Senator Levin speaking on June 7, 2000. Since I have no idea where he got the numbers, it’s hard to know just WIW.

    Congressional Record, V. 146, Pt. 7, May 24, 2000 to June 12, 2000

    “Right now the U.S. maintains 2,300 warheads on launch-ready alert: 98 percent of the Minuteman III and Peacekeeper land-based force on 2-minute launch readiness and 4 Trident submarines, two in each ocean, on 15 minute launch readiness.”

  17. Aaron Tovish (History)

    If I wasn’t worried about alienating all of you from the get-go, I would say you are all mad as hatters!
    The only person who even begins to ask the right questions is Alex W. You have all most definitely played DEFCON far too much.
    I and a few other brave souls have tried to bring to your collective attention the Robock et al studies on nuclear famine. To the best of my knowledge, our efforts have been studiously ignored. What is the problem?
    Have you all carefully considered these findings and found them to be manifestly in error or somehow irrelevant?
    If nuclear policy is not about assessing the risks associated with different nuclear postures, including abolition, and then pursuing the least risky, then what is it all about? Nuclear famine shows that the stakes are MUCH higher than previously reckoned. That ought to trigger a thorough re-evaluation of policy. Yet one sees no sign of this.
    Please, someone, enlighten me on why this is so.
    If you continue to ignore this question, then I will discount anything and everything you babble on about henceforth.

    • Daryl Press (History)

      1) It’s true that I (and perhaps others?) don’t know as much about these studies as perhaps I should. I know these claims have been made; I know that other scientists have disputed them, based on things like the observed consequences of volcano eruptions. I decided I couldn’t be an expert on everything, and I haven’t dug deeper into this topic for now.
      2) But if this topic is serious, those making these arguments need to do a better job making their research findings seem like real analysis and not just partly-informed advocacy. At your suggestion, I spent 15 minutes on Robock’s website and skimmed through two of the top pubs he posted. They’re filled with glossy pictures of hungry kids, and graphs of the alleged consequences of nuclear war, but nowhere do I see detail discussion about his methodology. Specifically, given that I’m not a climate scientist, I’m in treated in: what targeting assumptions is he making? Nowhere to be found. Furthermore, there’s a lot of vagueness about scenarios and assumptions in the text that exists, which are big red flags to me. Additionally, he says things with certainty that are clearly wrong or at least clearly unprovable (and unknowable): e.g., that the nuclear winter thesis ended the cold war. All these things make me highly dubious about his work, which would be really unfortunate if in fact his research is sound. But if it is, he should be crystal clear and he should carefully qualify things: e.g., “we find that an attack using fifty 1-megaton warheads on urban areas in South Asia would have the following effect on average temperatures…”
      4) Regarding the posts in this thread: I feel strange defending them all, because I disagree with more than half, but I haven’t seen anyone post anything in this thread that suggests they think large-scale counter-city war is a good (or acceptable) thing. Most posts are offering different people’s perspectives on whether land-based ICBMs have better or worse deterrent effect than other basing modes. Of all the people, I’ve gone the furthest in suggesting that weapons should also be assessed on their ability to carry out the missions they’re supposed to carry out — but if you look carefully I was talking about counter-force (retaliatory) and not burning down cities.
      If you think that people on this blog haven’t paid sufficient attention to important findings by Robock, he or you should package that analysis in a clear analytic fashion that permits real scrutiny — on the war plans side of things, as well as the climate science.
      I’m not sure why I took the time to write all this, other than I read your post as a sincere question about: why is this “nuclear winter/famine” stuff getting ignored?

    • Aaron Tovish (History)

      [I hope this reply appears after Daryl’s, since it is a reply to his remarks on my comment.]
      Daryl, thanks for having the intellectual integrity to address this matter of catastrophic climate change. I will respond to your points in order:
      (1) Of course, no one can be an expert on everything. But when these new claim put into doubt the very foundations of your existing area of expertise and there appears to be a collective unwillingness to ‘dig deeper’ surely it is wrong-headed to just shrug ones shoulders. (And perhaps the collective shoulder shrugging is mutually reassuring that we can all carry on business as usual?)
      It is not true that other scientist have disputed the catastrophic climate change findings. Robock has been begging scientists to come forward with well researched and formulated criticisms. It would be ‘better’ after all if they were completely wrong and the world was not being placed in such terrible danger. (See my closing point as well.)
      (2) Dear me! You read two introductory articles but fail to follow-up the references to peer reviewed pieces in which the methodology is spelled out in detail and the results (with uncertainties specified) are stated in strictly scientific terms. 15 minutes! The fate of the Earth may be (is) at stake and you form judgments after such a cursory examination? Shame on you.
      Specifically, as to the claim that nuclear winter studies helped to end the Cold War, this can be attributed to Gorbachev. Gorbachev may be fabricating (why would he), but you cannot slight the author for citing him. And is it so odd to think that the nuclear famine studies might also be sobering to our current crop of leaders if all those around them weren’t in denial about it?
      (3) Not about bombing cities??!! Are you really asserting that you can start to use nuclear weapons in any numbers without radically increasing the odds that cities will soon become targets? The DEFCON game is a GAME, reality has a way of getting out of hand. When the world is wired with nuclear weapons ANY use is an opening for massive use — by which I mean 20 large cities get incinerate (20 large cities roughly equals 100 Hiroshimas in terms of firestorms).
      When our mayoral delegations asked defense officials whether cities were excluded as targets the main response we got was, “We do not discuss targeting.” The British assured us, “We do not target cities as such.” Now that’s REALLY reassuring! Target industrial centers, target transport center, target political-military centers, and how many cities have been partially or totally incinerated?
      (Your final remark.) I will ask Robock to send you the ‘analytic articles’ you desire but are too busy with ‘counterforce’ arguments to research. But he has to do this work on a shoestring budget because all the money is going to the defense department for war gaming that IGNORES nuclear winter/famine, that limits its evaluation of attack damage to radiation and blast, as if the ensuing firestorms are irrelevant. One can hope that somewhere in the bowels of the Pentagon some competent and independent scientists have been commissioned to look at this seriously, but the public record on this is not hopeful. Take the starting point and end point of the last NPR: the head of the last posture review came to the 2009 PrepCom to let folks know about the process, he was asked by Steve Starr if the new studies on CCC would be taken into account this time around. The answer was that it was not in the posture review mandate so: NO. When the declassified review came out, I searched it for any key word I could think of that might lead to a mention of CCC. NADA.
      The entire nuclear enterprise now smacks of reckless endangerment of the highest order; and the failure to seriously pursue the establishment of a nuclear-weapon-free world is nothing short of criminal negligence.
      So I am sorry I wasted some of your precious time on the fate of the world. Please go back to your DEFCON games, and let’s hope the next time you look up the world is moving decisively for universal abolition of these monsters.

    • Jeffrey (History)

      I’d be interested in hearing any suggestions regarding in which field, if any, this sort of rhetorical approach is considered persuasive or even helpful.

      As someone who spends a lot of time trying figure out how to ensure human and environmental costs receive equal consideration along with vague appeals to security, I often find myself scrutinizing this rhetorical style of “damn-the-facts-the-fate-of-the-planet-is-at-stake.” I can’t help but think of Jean Genet’s tube of Vaseline.

      Why don’t you just point to the best peer reviewed study and we’ll discuss it like grown-ups?

    • Aaron Tovish (History)

      Jeffery, I am prepared to completely drop the ‘rhetorical approach’ once it is clear that a non-rhetorical approach will get a full and fair hearing. I and at least two others have tried to raise this matter ‘politely’ many times in the past and it was as if we were talking in a sound-proof box. I have been intentionally provocative as a means of waking those who are sleepwalking.
      Links to all the key scientific publications can be found at (thanks to Steven Starr). But Daryl was already in the right place (I assume) when he looked at , Robook’s site. There are a score of papers there nearly all published in peer review journals in the last four years. Some, I will admit, are heavy going for the non-expert.
      I propose that a discussion about the very foundations of nuclear policy formulation be started up on your website. Let me provide an analogy. A bunch of working guys get together every Friday night for a friendly game of poker. The stakes aren’t so high; on a bad night, a guy might lose his week’s wages. On evening one guy arrives late and takes his seat with the usual stack of chips in front of it on the table. He immediately notices that the usual jovial atmosphere has been replaced by a tension in the air thick enough to cut with a knife. The cards are dealt and he throws a chip into the pot. The others do so only very reluctantly. “What the heck is going on?” he asks. Finally, one of the players coughs out, “Tonight each chip is worth $100,000.” The guy literally flies back from the table, “Are you all completely nuts? I cannot play for those stakes, and neither can you. You have to call this off right now!” Would you say he is over-reacting?
      Even if one concedes (which is dubious) that nuclear weapons have instilled a level of caution that has significantly reduced the chances of another world war, is that reduction worth running the risk of catastrophic climate change of either the nuclear winter or nuclear famine variety? To evaluate that trade-off (which is actually more complicated) one MUST know the costs. What we are saying is that ALL previous policy reviews have been working with severely incorrect costs, so they can no longer serve as guidance except toward potential catastrophe.
      In short, at a certain cost balance, it is better to make the world ‘safe’ for conventional war than ‘unsafe’ for nuclear war. During the Cold War, the orthodoxy was, of course, reversed. We are in a different world and we have different information about costs, thus it should be no surprise that new conclusion will be drawn.
      When costs are properly taken into account then abolition reemerges as a competing policy or, more to the point, as the clearly superior policy. Now the ‘problem’ with abolition is that it is a collective multinational act rather than a unilateral national policy act. All the more reason to get to work on it right away and with determination and commensurate resources. I have a lot more to say on this subject, but I will leave it at this for now. Aaron

    • Jeffrey (History)

      You keep avoiding requests to identify a specific paper we can discuss.

    • Mark A. Gubrud (History)

      Jeffrey, I think this is a topic you should come back to.

      Robock’s site does display dozens of papers. Some are glossy presentations, some are brief reviews, others are major studies published in peer-reviewed journals. The latter examine the issue from a number of different perspectives which are fragmentary relative to the global problem: i.e. impact on corn production in the US, or in Venezuela. The issue is complicated, and requires consideration of multiple effects on multiple ecological, economic, social and political systems in many different scenarios. I’m not sure which single paper gives the best and most comprehensive overview. I’m sorry I don’t have the time at the moment to dig into this, and after all, a blog comment space ages rapidly.

      Robock is also not the only author on this subject. It is a subject that is clearly very important to assessing the degree of threat to human civilization posed by even a limited nuclear war scenario. I’d be happy to look into this subject for you if you’ll put up what I find in a future post. Or, I encourage you to look into it yourself.

      I was quite disappointed by Daryl’s admission that he spent only 15 minutes looking at only two publications listed on Professor Robock’s website before posting a comment which reads as if designed to convey to time-pressed readers (e.g. virtually the entire community involved in policy making) that Robock appears to be some kind of wingnut whose work can be dismissed with a smirk and a disingenuous “I’m not an expert on this.” It’s hard to see how anyone could claim to be an expert on the nuclear weapons threat to human security without covering the questions (at least) of nuclear winter and famine.

    • Jeffrey (History)

      It’s a deal: I’ll put up what you find in a future post as an annotated guide to climate science and nuclear war or something like that.

    • Aaron Tovish (History)

      Mark and Jeffery, thank you for taking a constructive attitude toward this potential game changer. I look forward to Mark’s take on the matter and you can count on me to be a constructive interlocutor.
      I would suggest this as a framework in the initial phase of discussion:
      (1) Do nuclear explosion on or over concentration of flammable materials (unlike almost every test explosion), trigger firestorms?
      (2) Do large firestorm inject materials into the upper atmosphere and what percent of what burned is lofted to those heights?
      (3) Are some types of particulates particularly effective at blocking sunlight, which it turn increases their own heat and loft ‘half-life’? (And how does this compare to, say, volcanic ejections?)
      (4) What is the half-life residence time for the bulk of upper atmosphere particles; what is it for the longer half-life particles?
      (5) What impact does the reduction in sunlight reaching the Earth’s surface have directly on plant grow and indirectly through climate change?
      (6) How do these impacts compare with previous reduction in food crops both globally for one year and locally for five-ten years, i.e. has there ever in recorded history been such a shock to human food supply?
      (7) Who is at risk, and what are their chances of survival depending on the length of the supply contraction?
      (8) For those who do not starve outright, how does malnutrition effect their resistance to epidemics and reduce their reproductive fertility?
      (9) Some general discussion of other than human costs, i.e. ecological, economic, historical, etc.
      There is also the still unexplored impact of ozone depletion, reckoned to be in the 40-70 percent range, i.e. far greater than any other man-made ozone holes.
      I think it would be useful to also consider the impact of a conventional world war lasting one or two years — not an easy task either. The same for a prolonged regional war — there the Iran-Iraq conflict would be a good guide.
      Once we have at least some sense of the relative costs, we can turn our attention to the chances (probabilities) that those costs would be incurred given difference nuclear use policies (including no use in a NWFW, or at least very low probability of use compared to a nuclear-weapon-beset world). From that we can compare the cost expectancies of the various policies. Let the dice fall where they may!
      Sound reasonable? Aaron

    • Jeffrey (History)

      That is an absurdly large agenda. Let’s start with Mark’s literature review and see where that puts us.

    • Aaron Tovish (History)

      Perhaps, but don’t be surprised if Mark’s review ends up covering precisely these points. One of the values of wonkiness is it’s thoroughness, no?

    • Jeffrey (History)

      Thoroughness does not preclude making choices about what is, and is not, important.

      There is a difference between dilettantism and wonkiness, even if both exhibit a certain affinity for the fine details.

  18. Nick Ritchie (History)


    A bit late to the game on this one, but more meat for the grinder is available in the Senate Committee on Governmental Affairs’ hearing on ‘Evaluation of the US Strategic Triad’ in June 1993 in response to the GAO report.

    Full text here:

    SECDEF Perry unsurprisingly rejects GAO’s findings.

  19. JohnLopresti (History)

    I appreciated commenter Ritchie’s link to the transcript from the hearing, in the US Senate, in a (1993) committee chaired by Ohio senator J Glenn. As commenter Ritchie summarized, Dr. Perry’s remarks trended into refutation of the original post’s cited GAO report.

    Also interesting is the commenter a Thomson’s linked (2000) page from the US Congressional Record, containing remarks by US Senate armed services committee chair senator Levin. That committee’s website often is quite sparse.

    It would be interesting to know apt research methods for finding documents such as those two provided by these two helpful commenters. Congressional committee transcripts and records are notoriously available behind obscurantist online facades which shift with political winds every two years, as well.

    To paraphrase the remarks, or rather, the responses for the record, after the hearing, by Dr. Perry, the ICBM part of the ‘triad’, in 1993, had a worth which was larger than merely the math of adding 1+1+1 to obtain a ‘triad’; Perry called that benefit a ‘synergy’. That is, having (at least) 3 elements of the defense deployed provided more than the net worth of each standalone.

    Beyond these elementary impressions of mine, it appears clear from reading the document set pertaining to this thread that technology has shifted much.

    Ultra low frequency sonars were in their infancy in the 1970s. The space shuttle’s 30 years contributing to guidance systems development began only 31 years ago. Further, I suspect many new-tech-based linkages among defense systems remain entirely classified.

    Even the commenter-linked FEMA study annex, mapping fallout anticipated during the 1980s-1990s, looks dated, and of perhaps dubious classification status?

    There are lots of unknowns in the discussion of ICBM worth.