Foreign Relations of the United States, 1977–1980, Volume IV, National Security Policy

97. Memorandum From Secretary of Defense Brown to President Carter¹

Washington, October 23, 1978

SUBJECT

• The Triad and Its Alternatives

At our meeting on September 29th, you asked why—since Minuteman is becoming vulnerable, and since MAP involves thorny questions of its own—we should not simply abandon the ICBM leg of the triad, and strengthen the remaining two legs as necessary.² You were also interested in other possible basing systems. This memorandum is to give you my thoughts on those issues.

In simplest terms, we could indeed abandon the ICBM leg and move to a dyad of SLBMs and bombers. Or we might retain a triad—or something like a triad—by replacing silo- or MAP-based ICBMs with one of the newer concepts (air-mobile ICBMs, land-mobile ICBMs, air-launched ICBMs, or perhaps even SLBMs carried in coastal submersibles). With any of those alternatives, we could probably maintain an overall balance with the Soviet Union in terms of static indicators and in terms of dynamic exchange calculations. In addition, though it’s too early to estimate costs for the novel basing schemes, we can say that at least the SLBM/Bomber dyad need not be more expensive than an SLBM/Bomber/MX–MAP triad of generally equivalent capability as measured by most (though probably not all) criteria. But on the other hand, with the dyad alternative, and to a lesser extent with the triads involving the alternatives to MAP, we would give up features of both perceptual and military value that we have enjoyed in the past, and—if one puts aside some of the political and possible verification questions about MAP—would continue to enjoy with the MX in a MAP deployment.

The effect of the US giving up, under successful Soviet pressure, a military capability of considerable value which the Soviet Union retains would, in my view, have disastrous consequences both internationally and domestically.

Though precisely what we would give up in terms of military value depends on the particular alternative, there are two types of losses: the first is the loss of specific characteristics (which determine what kinds of things we will or will not be able to do); the second is a loss in simple diversity (which determines our confidence in being able to do it, come what may, in the face of new and unexpected Soviet developments, or of unforeseen occurrences in a combat environment).

A less drastic alternative would be to retain Minuteman in silos until its reliability is unacceptably degraded by age, perhaps then replacing it with a new missile, and to plan to launch it before Soviet R/Vs destroy it. This presumes reliable attack assessment, and a decision sequence that works. The Soviets might not be sure that sequence would not work; I, for one, would be far from sure that it would. My major concern can be put this way: such a course would be a major step on the road to giving computers the power of decision on going to war.

LOSS OF CHARACTERISTICS

As for specific characteristics, some we value are:

• Independence from tactical warning. This means being able to ride out a Soviet strike. It’s desirable because we avoid the need for a “hair trigger” response to possibly ambiguous warning with all the complexities that could entail, avoid as well the risk of the hair trigger not working because the postulated decision process doesn’t function, and also because a Soviet attack on our warning systems cannot jeopardize forces that do not depend on warning in the first place.

MAP deployments of ICBMs—if their size is in the proper ratio to the size of the Soviet attack—would have this valuable characteristic. SLBMs have it already. Bombers do not—they can be wiped out if caught on the ground, or near their bases. (For simplicity, I’ll save the discussion of novel basing systems for later in this memo.)

• Endurance. This means being able to maintain some capability after an initial exchange—perhaps for days, or weeks, or even a few months. It is desirable not only to prevent further coercion, but as a counter to Soviet sheltering of political leadership (they’ve got to come up sooner or later), and against third countries who might suddenly find themselves in positions of great relative strength.

Again, ICBMs in a MAP deployment would have this characteristic. The SLBMs per se do too, but their problem in this respect lies in the dubious endurance of the systems to communicate with them. Bombers, because their bases are vulnerable, have little endurance, unless they have so many potential recycling bases as to prevent the Soviets from targeting all the bases.

• Quick Response Hard Target Capability. “Quick” refers not only to short flight times [1 line not declassified], but also to your ability to get in touch with the force quickly, and how quickly it can respond. “Hard target” primarily implies accuracy (though warhead yields are not irrelevant). There are two classes of hard target: one is the roughly 1100 Soviet silos we would expect under SALT II; [3 lines not declassified] (There are also [less than 1 line not declassified] hard communications points; we do not have enough accurate weapons to target these now.)

Thus far, ICBMs have had unique advantages in this respect—quick and reliable communications, a quick-firing capability, short flight times, and high accuracy. SLBMs are currently weak here—slow communication, mediocre accuracy, and limited warhead yield. [5 lines not declassified] Bombers have no quick response hard target capability. [2 lines not declassified]

• Good C3. “Good” here refers to communications links that can transmit adequate amounts of data as rapidly as needed, and are not vulnerable to enemy attack. Simple execution [1 line not declassified] requires one level of C3 sophistication; a “warfighting” capability (see next below) can require considerably more.

ICBMs have great advantages in C3. Though no C3 system can be invulnerable, [11 lines not declassified] The survivability of these aircraft—because of the vulnerability of their bases—is also limited in duration, thus limiting the endurance of this part of the C3 system. [2 lines not declassified]

• Warfighting Capability. Sometimes called “flexibility” or “responsiveness”, this refers to the ability to do more than execute a massive pre-planned strike—to be able to use limited amounts of strategic force against particular numbers and types of targets during a nuclear war on the basis of what had occurred before. This includes retargeting, assessment of surviving US and Soviet forces, and connection to a surviving NCA. We have considerably more of that kind of capability today than we had in past; further developments in this direction are possible, at a price. Generally desirable characteristics to this end are the ability to retarget forces and execute missions promptly and with high probability of success, to limit collateral damage, and to restrict the size of each mission to the minimum deemed necessary.

I should note that I doubt that this kind of a strategic war is among the likely possibilities, or that it could be fought without rapid escalation to an unlimited spasm. But there are many indications that the Soviets are structuring the forces to fight such a war. Their plans and exercises point the same way. If they think they have such a capability, and both we and the Soviets know that the US does not, adverse military consequences are possible, and adverse political consequences probable.

ICBMs are the best of our forces for such missions. We can retarget them quickly and launch as few as one if necessary. We are now at the point of adapting some of our Minutemen for retargeting [1 line not declassified] They have a high penetration probability, and their accuracy limits collateral damage. SLBMs are much less well adapted to warfighting: communications tend to be slower, and launch of a single missile tends to expose the position of the SSBN. Penetrating bombers are not well-adapted to such missions: flight time is long, and—since such raids are likely to be small—penetration is problematical, particularly if attacks on Soviet defenses were to be restricted in an effort to limit collateral damage. On the other hand, aside from their long flight time, ALCMs carried on bombers offer good penetration probability, accurate delivery, and the likelihood of low collateral damage.

In summary, by giving up silo- or MAP-based ICBMs, we would lose the part of the Triad that provides our best warfighting capability and has our most reliable C3. [1 line not declassified] the best prospects for endurance of both the system and our ability to communicate with it, and a system that does not rely on either strategic or tactical warning. Some of these capabilities can be preserved with some of the alternative mobile ICBMs.

LOSS OF DIVERSITY

The other kind of loss entailed in moving to a dyad has to do with diversity, which relates to our confidence in our strategic forces in spite of an uncertain future. Diversity benefits us in three ways:

• It protects us against things the Soviets might do—“solve” the ASW problem, deploy a very effective air defense, face us with an ABM breakout, etc.

• It protects us against failures on our own part—physical failures of equipment (such as in our SLBMs in the early ’70s), unsuspected vulnerabilities (such as the EMP problem we discovered in the late ’60s), shortcomings in our war planning (such as the fratricide problem we recognized in the mid-’60s), etc.

• It tends to complicate enemy planning—for example, Soviet attacks timed to catch our bombers will warn our ICBMs, and vice versa.

Eliminating the ICBM leg would eliminate one of our hedges against marked Soviet progress in ASW (which could jeopardize the SLBM leg), but not the other—our bomber force. It could also eliminate one of our hedges against Soviet development of a successful tactic against our bomber bases—possibly a barrage attack—or a much stronger Soviet air defense (either of which could jeopardize the bomber leg), but not the other—our SLBM force. Thus, failures of a single leg of a dyad of SLBMs and bombers would be extremely serious but need not be catastrophic. Failure of both—unlike failure of two parts of the triad—would be a total disaster.

COSTS

Before turning to the possibilities of novel basing systems, I would like to elaborate on my statement above that dyad and triad costs need not differ appreciably. By way of illustrating that, below are outlined three forces, all roughly comparable in capability (aside from the inherent differences in characteristics noted above) to each other and to the Soviet forces expected under SALT II. One is a triad, with 400 MXs in a MAP deployment; the other two are dyads, one heavy in the airbreathing half (100 cruise missile carriers with 60 ALCMs each), the other heavy in the SLBM half (768 D–5 Trident II missiles in 32 SSBNs).

[table not declassified]

These differences in costs are well within the accuracy of estimation.

One would normally assume that the failure of a single leg would affect a dyad more than a triad. However, it’s not that simple if costs are held constant—if the three legs of a triad are smaller than the two legs of a dyad. For example, if the most expensive and least effective leg of a triad were eliminated, and its funding applied instead to the more efficient two legs that remain, conceivably each could be made strong enough to match any two legs of the old triad.

How that works out depends on the relative costs and effectiveness of the three legs that happen to be involved. In the example above, if the ICBMs fail, either dyad (both of which omit the ICBM leg) is of course superior to the triad; if the bombers fail, the triad is better than the first dyad but worse than the second; if the SLBMs fail, the triad is worse than the first dyad but better than the second.

Thus, dyads and triads of roughly equivalent capability—insofar as it is possible to make them equal—need not differ appreciably in cost, and may not differ appreciably in hedging against failure of any single leg—it depends on which leg fails.

NOVEL SYSTEMS³

As for novel basing alternatives to the MAP system, I should caution first that none has been thoroughly studied (at least not recently), and experience teaches us that many concepts that initially seem to be ingenious ways of solving existing problems later turn out, after a closer look, to be flawed with new problems of their own. Furthermore, it is most important to understand that none of these novel proposals is really a substitute for the ICBM force; though they surely could add to our strategic nuclear force, none could preserve all the valuable characteristics of the ICBM force—excellent warfighting capability and reliable C3, and endurance, and complete independence from tactical warning.

• Air-Launched Ballistic Missile (ALBM): With this system, large aircraft would carry ballistic missiles of intercontinental range. (In a feasibility test of such a system, we dropped and fired a modified Minuteman from a cargo aircraft some years ago, but no guidance system was operative, nor was capable of doing so).

This concept, however, is much more closely related to the ALCM-carrying bomber than to an ICBM. It suffers from the bomber’s dependence on tactical warning and lack of endurance. As an alternative to the ALCM, it promises only better penetration (though we do not expect penetration to be a severe problem for the ALCM) and quicker reaction. On the other hand, its accuracy and the number of warheads carried per aircraft would be much inferior to the ALCM’s. And, of course, instead of hedging against a Soviet ABM breakout (as the ALCM does), it would hedge against a breakthrough in Soviet air defenses.

• Air-Mobile ICBM: With this system, ICBMs would be transported from airbases to dispersed launching sites, where they would be unloaded, erected, and fired. Through use of short take-off and landing aircraft, the potential number of launching sites is very large. The dispersion process could begin on strategic warning, though some continuing movement of ICBMs between sites would be necessary to defeat Soviet reconnaissance and targeting.

This system, though similar in many respects to ICBMs as we normally think of them, would not be completely independent of warning, and it could be difficult to provide communications as secure and reliable as the ICBM’s. The proper aircraft does not now exist.

• Land-Mobile ICBM: With this system, ICBMs would be carried in land vehicles, either in more-or-less continuous motion to prevent pre-targeting, or in a “dash on warning” mode (in which vehicles normally kept at central bases would disperse on tactical warning, possibly to shelters randomly chosen from among a large number prepared in advance). There was much interest in the ’60s in mobile versions of the Minuteman, but the concept was abandoned in light of the problems of C3, security, and—most of all—the prospect of public reaction to the idea of so many nuclear warheads continuously circulating on the roads or rails of the country. However, the land required and the speed of the vehicles is subject to tradeoffs with the hardness of the vehicle, or of its MAP-like shelter.

In comparison with ICBMs in silos or MAP, this system would suffer in terms of reliance on tactical warning, or in terms of C3 for versions relying purely on mobility rather than random shelters.

• Coastal Submersibles: This concept postulates very small (350-ton), inexpensive submarines, each armed with a pair of encapsulated MX missiles, and manned by a crew of 12. They would operate on the coastal shelf, with communications via a redundant grid of fiber-optics cables on the ocean floor. The submarine would be connected to plug-in terminals in the grid through an umbilical.

If feasible, this system would probably be further off in the future than any of the other novel basing schemes. Its potential advantage would be lower cost; the likely disadvantage, even if feasible, is vulnerability. Aside from its vulnerability to Soviet trawlers—tearing up the cables, attacking the submarines, or even making boost-phase intercepts of the missiles—I am particularly concerned about vulnerability to the so-called Van Dorn Effect, a subject of considerable interest when the Navy first considered such coastal submarines.

The Van Dorn Effect postulates that as the wave front from the detonation of a large-yield weapon in deep water to seaward approaches the rise to the continental shelf, its energy becomes crowded into a much smaller volume, and the result is violent motion throughout the mass, with waves at the surface hundreds of feet high. Theory predicts that a single weapon could generate a destructive wave front over a hundred miles in width. Though there is clearly no way of running any kind of full-scale test, vastly scaled down tests using sizeable amounts of chemical explosives in a lake gave results that agreed well with the theory. Should the coastal submersible be vulnerable to this effect, the entire force—and much of our coastline as well—could be wiped out with perhaps a dozen, large-yield Soviet weapons.

SUMMARY

We could abandon the ICBM leg and move to a dyad. By strengthening the airbreathing and SLBM components, we could build back our number of warheads, megatonnage, etc. But we would also be sacrificing some important characteristics that we would enjoy with the ICBM force, and we would be less well hedged against unsuspected developments on the part of the Soviets and unsuspected failures on our own part. And for a US force size comparable with that the Soviets are sure to have during the mid-80s, moving from a triad to a dyad would not save a significant fraction of our strategic force costs.

We could also consider—and we are doing so—novel basing alternatives other than MAP. None of these has been studied thoroughly enough for me to recommend a deployment decision as among them. The addition of such alternatives to a dyad will result in a force that may approach the full goals of a triad, though none would fully match the characteristics of a triad that included MX in MAP basing. The alternatives could avoid some of the verification and political concerns connected with MAP, but could have other political problems of their own.⁴

Harold Brown

1. Source: Carter Library, National Security Affairs, Brzezinski Material, Subject File, Box 43, Missiles: 11/78–4/79. Top Secret. Carter initialed the upper right corner of the memorandum. Brown’s handwritten covering memorandum to Carter, October 23, reads: “This is the first of a series of memos on issues that have arisen in past discussions on the Defense program. They are deliberately structured as information rather than decision papers, because I will want to make specific recommendations separately at a later time. Though they may seem lengthy, they are worth reading in full.”
2. See footnote 2, Document 90.
3. Carter wrote: “MAP is very ‘novel’” in the right margin next to this heading.
4. In November 1 memorandum to Brown entitled “The Triad and Its Alternatives” Brzezinski wrote: “Attached is a copy of your memorandum of October 23 on the above subject. Please note the President’s comments.” (Carter Library, National Security Affairs, Brzezinski Material, Subject File, Box 43, Missiles: 11/78–4/79)