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

101. Memorandum From Secretary of Defense Brown to President Carter¹

Washington, October 28, 1978

SUBJECT

• ICBM Vulnerability, MAPS, and Verification

As we discussed at our midterm review, the growing numbers and increasing accuracy of Soviet missiles will give them the capability to destroy [less than 1 line not declassified] of our Minuteman silos by 1982–1983.² In a paper sent to you last week,³ I discussed the unique value of the ICBM leg of our Triad and the risks of giving up our ICBM force—especially under this pressure from the Soviet Union.

I also noted at the midterm review that while we should improve our attack assessment system, we should not allow our ICBMs—at least not all of them—to be so vulnerable that it would be necessary to launch our force out from under an attack in order to save it. Our ICBM force has had the ability to ride out any conceivable attack on it. This is far preferable to a situation where, because the ICBM force cannot survive, we embark on a course whose outcome is that a computer decides whether we launch a nuclear attack.

The alternative to the sobering options of retreating to a dyad, or being forced into a launch-under-attack policy, is to provide survivable basing for our ICBMs. We still are studying the various alternatives, but already it is clear that there are no simple, quick or cheap solutions.

Leaving aside political and perceptual problems, as well as the very important problem of uncertainties in counting the number of Soviet launchers (or missiles) if they go to MAPS or some other mobile concept—the best solution, both from a technical and from a cost point of view, appears to be MAPS. This is not to say MAPS is an inherently attractive solution; the problems just referred to cannot be left aside. But all of the other basing systems we are considering either cost more or compromise away some important characteristic of the present ICBM force (e.g., no dependence on strategic warning), or both. In most cases they also, like MAPS, involve counting uncertainties if the Soviets build similar systems. You should not expect our studies of these alternative basing systems to reveal a better solution than MAPS. Rather, we are looking at whether there are acceptable systems which do not raise the same issues of SALT verification and intelligence requirements that MAPS does.

It is important to emphasize that the SALT verification problems posed by MAPS are only a special case of a general intelligence problem that would also exist if SALT limitations did not exist. That problem is the requirement that our intelligence be able to monitor the size of the Soviet RV force which a MAPS must face with sufficient certainty to give us confidence that we have enough shelters to absorb the threat and leave an adequate number of U.S. missiles surviving. Uncertainties in the size of the Soviet force can stem from a Soviet MAPS, Soviet cheating on SALT, or Soviet measures to conceal the size of their force in the absence of SALT limitations.

Recognizing that there will always be some uncertainty in our estimates of the size of the Soviet force, a MAPS system would be designed with a tolerance for considerable uncertainty as to the Soviet ICBM inventory. For example, in one MAPS design concept we would plan on 200 M–X missiles with 50% survivability. If the threat from Soviet missiles were doubled, survivability would be reduced to 15%, i.e., only 30 M–X missiles—but still a force equivalent to 100 Minuteman IIIs. We also could hedge against threat increases by adding additional shelters; for example, if as in the previous example the threat doubled, a 10% cost increment for added shelters would keep survivability from dropping below 25%, or 50 M–X missiles.⁴

In my view, a crucial point in discussion of MAPS verification in the SALT context is this basic fact about the concept: The system would, for strictly military reasons, be designed to tolerate a substantial concealed increase in the size of the Soviet force. The size of that tolerable concealed increase would have to be large enough that we would be confident our intelligence could detect a force larger than that with very high confidence—whatever concealment measures the Soviets might employ, including copying the MAPS concept. Meeting this strictly military requirement would in itself meet the basic test of arms control verification: That test is not absolute certainty, but high confidence that cheating (here, concealment of illegal missiles in a MAPS) would be detected before it could affect our security or undermine the strategic balance. For SALT it is very desirable—and may be politically essential—to have more confidence in counting the Soviet force than we require to meet this military requirement. However, it is important to recognize that we are talking about adding to a level of confidence that is, by definition, already adequate for preserving our military security.

With this perspective, it is reasonable to examine critically the SALT issues raised by MAPS:

1. Is MAPS legal under a SALT II agreement? (SALT says that we may not use deliberate concealment to impede verification by national technical means, including satellite reconnaissance.)

2. Can we design MAPS to allow verification and still keep the Soviets from knowing which shelters to target?

3. Is the Soviet Union likely to deploy a MAPS?

4. If the Soviet Union deploys MAPS, can we verify?

5. If the Soviets deployed MAPS, could they use it to “break out” from a SALT treaty?

6. If the Soviets deployed MAPS, and in fact complied with SALT provisions, would SALT critics claim the Soviets were cheating (and were also able to tell which of our MAPS shelters contained our missiles), thereby undermining public confidence in the system’s survivability, and perhaps thus push us to respond by filling up all our shelters?

These, like most questions raised about SALT verification, are difficult and complex. Our studies of the past few months have provided some important answers. In this paper I will discuss these issues using MAPS with vertical shelters (silos) as a frame of reference, but most of the discussion is also applicable to hybrid trenches (MAPS with horizontal shelters under cover), which I believe is also a viable, but somewhat more expensive, option.

1. Is MAPS legal under SALT II?

Yes. The verification provisions of the draft treaty say that we may not take deliberate concealment measures that impede verification by national technical means. Although MAPS obviously conceals the current location of launchers, we could and would as discussed below make special design provisions to permit verification of their numbers. Provided verification is possible through NTM, concealment of current location is not barred—otherwise it would be illegal to send SSBNs to sea, an act which certainly conceals their location.

2. Can we design MAPS to allow verification and still keep the Soviets from knowing which shelters to target?

Yes. In addition to the inherent verification opportunities afforded by our open society, we have ways of allowing the Soviets to verify our compliance with MAPS without endangering the security of the system. These have been studied by both SALT and military experts, and there is good agreement that they can provide an adequate approach to this problem.⁵

The first verification method involves insuring that the Soviets can monitor the number produced. We could do this by designing MAPS so that all missile-launcher combinations go through an assembly process which can be observed by satellite reconnaissance before they enter the shelter area. Thus the Soviets can count the missiles and launchers going into the shelter area without knowing which particular shelters (silos) they are going to. (This is similar to the way we verify SLBMs today, in that we make our count in the assembly area but do not know where the missiles are after the submarines carrying them are deployed.) The length of time required to assemble and deploy a missile and launcher before it is put in a shelter is long relative to the intelligence cycle time. The Soviets may worry that we would cheat by deploying a covertly assembled missile and launcher into a MAPS shelter overnight under cover of darkness. While this would be a difficult and risky undertaking, it is conceivable for MAPS (as for a variety of other strategic systems).

To help reduce this residual uncertainty about whether illegally produced launchers had been deployed secretly, we could also make a practice, or formally offer, to open a few groups of shelters on demand to prove that there is only one missile and launcher per group. We would open all shelters in a designated group except the one that was full. This occasional checkup could be accomplished by satellite; if the Soviets prefer, some of these inspections could be done by ground observers since there is no security risk in showing them empty shelters. I believe that countermeasures can be used to maintain security but still permit adequate verification. For a covered MAPS (hybrid trench) the missile would be assembled and put into the trench while under observation. Then the trench would be sealed except for a small opening for maintenance purposes.

No set of measures such as these can be absolutely cheat-proof—but there is a residual uncertainty about verification of any weapons system. This kind of measure would eliminate any basis for a good-faith claim that the situation for MAPS was in any way more uncertain than for other, more familiar systems.

3. Is the Soviet Union likely to deploy a MAPS?

Probably not. First of all, MAPS as such is an expensive defensive system which as such gives them no additional striking power. Second, if the Soviets decide to improve the survivability of their ICBM force, they already have developed (but not deployed) in the mobile SS–16 or derivative, a model for a “truck-mobile” system which should be more than adequate. However, we cannot be certain on this point. If they did decide to deploy MAPS, they certainly would have fewer problems than we relative to land availability and environmental restrictions, a factor which would ease several types of non-silo basing.⁶

4. If the Soviet Union deploys MAPS, can we verify?

The Soviet Union is bound by the same SALT provisions as are we. Therefore, if the Soviets deploy a MAPS they would have to make adequate provisions as to its verifiability through exposure of the process of production, assembly and deployment. Otherwise, they would be impeding verification by NTM (including satellite reconnaissance) by concealment.⁷ If they followed the example I have described for our MAPS, we could verify it with good confidence. [2 lines not declassified] The potential problem, of course, is that they might make provisions for verification that we would judge less than adequate. In that case, we would then have a major argument in the SALT consultative commission. This is a risk of the kind we face with other SALT verification questions, including a variety of alternative basing concepts, and have been dealing with in the SCC.

5. If the Soviets deployed MAPS, could they use it to “break out” from SALT?

The answer is yes, but only if they are able secretly to produce and hide more than a thousand ICBMs. A few hundred covert ICBMs would not change fundamentally the nature of the threat; in fact, our MAPS would be designed initially to accommodate a 2:1 increase in Soviet RVs over present levels (achieved either by cheating or by RV fractionation), and could be expanded with relatively small cost to accommodate a 3:1 increase. Therefore, in order to pose a serious threat to our MAPS, their covert production and storage operation must handle more than a thousand ICBMs and their forces in the field must be able to support twice as many launchers as the apparent, legal force. I believe we would detect such a massive operation. Just as important, given the nature of intelligence, they could not be confident we would not detect it. The Appendix⁸ describes our problems in counting their ICBM production. This is an extremely important requirement. It deserves more intelligence emphasis; our current capability is far from perfect, but it is also far from zero. This issue is not unique to MAPS. It is fundamental to our ability to assess strategic balance, with or without MAPS—in fact, with or without SALT. As explained above, that confidence is a necessary precondition of the military acceptability of a MAPS system—that we could detect a massive Soviet effort needed to prepare to overwhelm it with secret missiles. If the Soviets can covertly produce and store more than a thousand SS–16s, SS–17s or SS–18s, they can launch them from their storage areas. All these missiles are designed for “cold-launch,” and could be launched from a simple pad or a simple erector launcher, with no need for silos. A MAPS, to be sure, would provide in-place C3 and logistic support, but would do so at locations which are under our intelligence “microscope.” If the Soviets can solve the problem of covertly producing and shipping that many ICBMs and expanding the support base in deployment areas for 2–3 times as many launchers as the legal force (which I doubt), I believe they could solve the further problem of covertly producing and shipping the attending C3 and logistic support. Thus, to the extent this is a real risk, it is a risk with or without MAPS.

6. If the Soviets deployed MAPS, and in fact complied with SALT provisions, would SALT critics claim they were cheating, and that our MAPS missile locations were not secure?

The answer is probably yes—such claims likely would be made, no doubt by some of the very people now advocating MAPS. Would that destroy our confidence in the viability of the system and force us to respond by filling up our shelters and abrogating SALT? Not unless the credibility of our strategic strength had been badly eroded by other issues. Your Administration and two which preceded it have weathered similar charges about Soviet cheating on SALT I—and about alleged Soviet military superiority in various areas—and have not given in, nor have they ever lost public support as a consequence. We probably shall have to deal with similar charges that the Soviets are cheating on Backfire, cruise missiles, SS–16 production, RV freeze, and other Treaty limits.

To summarize, I believe that for strictly military reasons our intelligence must be able, even without SALT, to detect a Soviet buildup big enough to threaten MAPS’ viability. Most of the additional verification questions raised about MAPS in the context of SALT can be answered by the special design provisions we would make on MAPS to provide for adequate verification while maintaining security. In addition to political problems and impressions, that still leaves two possible situations that cannot be put aside. One is the possibility that the Soviets might succeed in covertly producing and storing more than a thousand missiles, and then suddenly “breaking out” from SALT. The other is the possibility that some in the United States will claim that the Soviets are doing so and demand that we abrogate SALT. Both of these are risks. The risk of covert production exists even if we do not have a SALT, but is aggravated by the convenience of silos for the sudden deployment presumed. The risk that SALT critics will claim that the Soviets are cheating and that we should abrogate is aggravated by MAPS but also exists without MAPS. These factors need to be judged in the light of the risks of Soviet misbehavior, or charges of Soviet misbehavior, on Backfire, cruise missiles, new types and other provisions.

Risks of this sort must be weighed against the perceived benefits both to SALT and to strategic stability. In the case of MAPS, the risks are real, and the benefit is the degree of confidence maintained in the survivability for our ICBM force. When we have fully explored the alternatives to MAPS to the same degree of detail, MAPS may well prove to be the best option in combined political-military-cost terms. Many of the alternatives present comparable verification/intelligence problems. For that reason I believe we cannot now prudently rule MAPS out, though I do not think we can now choose it to the exclusion of the alternatives.

Early in November I expect to have for your consideration a development and deployment plan for strategic forces over the next ten or more years, including options for future decision. In presenting such a program, I shall include considerations of the Triad and MAPS along with the fundamental issue of the strategic balance with the Soviet Union.

Harold Brown

1. Source: Carter Library, National Security Affairs, Staff Material, Defense/Security, Ermarth, Box 2, Defense (Items in the System): 9–12/78. Top Secret; Sensitive. An undated paper prepared in the NSC entitled “Comments on Harold Brown’s 28 October Memo on ICBM Vulnerability, MAPS, and Verification” is ibid.
2. Refers to the September 29 Mid-Term Defense Review meeting. See footnote 2, Document 90.
3. See Document 97.
4. In the right margin next to this paragraph, Ermarth wrote: “No numbers.”
5. In the left margin next to this paragraph, Ermarth wrote: “oh?”
6. In the left margin next to this paragraph, Ermarth wrote: “DoD said the same thing about MIRV.”
7. Ermarth drew a vertical line in the right margin next to the first three sentences of this paragraph and wrote above it: “inadequate.”
8. Attached but not printed is an undated one-page paper entitled “Monitoring of Soviet Strategic Forces.”