The Future of the U.S.-Soviet Nuclear Relationship (1991)

Page 25

Page 26

removing less survivable ones, while continuing to invest as necessary in the forces retained.

One further factor should affect future nuclear weapons choices: there is little need to match specific characteristics of Soviet nuclear deployments. Deterrence of nuclear war is not impaired if the nuclear deterrent systems of the United States and its adversaries are substantially different in type and deployment, so long as they are survivable. Different nations may best meet their nuclear systems' increasing need for survivability in diverse ways.

In this discussion, we turn first to the question of estimating the appropriate numbers of central strategic nuclear systems and then to sample survivable force structures.

Page 27

such a case, survivability and reliability would be lower than if the United States had acted when it received warning of an attack. Using this assumption, the overall ratio of weapons to targets (in actuality there are different ratios for different nuclear missions) has been estimated to lie in the 3-5:1 range. ³ With this ratio, covering a list of 3,000 targets would require a total of 9,000-15,000 strategic nuclear weapon systems.

Clearly, if the conditions for use are less adverse than assumed, the United States could strike far more targets in the Soviet Union. In particular, more nuclear targets could be struck. While the United States has not had a disarming first-strike capability at any time in the last 25 years, it could come much closer to disarming the Soviet Union if, for instance, Soviet submarines were in port, Soviet bombers were not on alert, and the Soviet leadership were vulnerable. In that case, the United States could anticipate, albeit not with great confidence, that any Soviet retaliatory strike would be weakened and any damage to the United States would be limited. The same would hold true for a Soviet first strike against the United States. The possession of large, partially vulnerable forces on both sides has thus created an unknowable potential for instability in crises fueled by the perceived value of striking first rather than in retaliation. This potentially disastrous instability may be partially alleviated, though probably never fully remedied, by lower numbers, in conjunction with more survivable systems and better operating procedures.

The current numbers of central strategic nuclear weapons have seemed excessive for some time. The START agreement will lower them by about 25 percent (see Appendix A). These numbers will still be very high for a purely deterrent posture. Given the prospects for changed conditions outlined previously, significant further cuts should be possible. These additional cuts, for the reasons noted earlier, would alleviate the risks and costs of maintaining forces excessive to need at no cost in lessened deterrence. The principal task facing planners will be to work out the steps needed for a secure transition.

In what follows, we consider two possible lower levels of strategic weaponry. We believe each level would be stably associated with a possible future set of political and technological circumstances, which we briefly outline.

Page 28

tion in verification and in communicating deployment plans in both countries continue to improve.

The Soviet Union maintains its forces within CFE limits.

Both countries begin to posture their nuclear forces to make the remaining forces more survivable, so that nuclear weapons systems worth targeting decrease both in absolute numbers and as a fraction of the whole.

Neither country deploys a strategic defense system. Development and test activities are carried out subject to agreement, whether the current one embodied in the Antiballistic Missile (ABM) Treaty or some agreed modification.

Neither China, France, nor Great Britain builds up nuclear forces beyond current levels. These are currently below 1,000 in France and Great Britain. Chinese force levels are not known for certain, but could be of the same order of magnitude. No other major power (for example, Germany or Japan) acquires nuclear weapons.

These circumstances, taken together, describe a situation in which the United States and the Soviet Union could bilaterally reduce their nuclear forces well below START levels. It is of course no more possible to derive uniquely the specific number of weapons systems needed for deterrence under the evolving political climate described than it was possible to derive the pre-START levels from the then-existing situation. Despite this, whatever next step is taken would have to specify some reduced overall force level.

The desired level will depend in part on the nature of the remaining forces. In what follows, we assume that forces on both sides will become more survivable and less highly MIRVed (see below). Under that assumption, the targets for the reduced forces would not include alert deployed nuclear forces such as silos or mobile missiles, since these would cost more of the attackers' remaining weapons than they would be worth. In addition, we include no targets outside the Soviet Union; deterring China or possibly other nuclear powers may be taken as a lesser, included case. We further assume that the remaining categories of targets (nuclear submarine and air bases, major conventional military bases and command and control centers, war-supporting industries) would be covered up to approximately 60-70 percent of total capability.

These assumptions lead to a total of 1,000-1,600 targets, depending on the mix of yields of the nuclear systems, distributed as shown in Table 1. Lower yields might require allocating more warheads to a given target (a target may thus represent more than one aimpoint), but could mean significantly less collateral damage and fewer total casualties.

The discussion and “saturation” curves in Appendix B show how these estimates were made. Clearly there is considerable leeway in choosing cut-off points in each target category.

Page 29

TABLE 1 Sample Target Allocation

Target Category	Aimpoint Range
War-supporting industry and energy sources	300-600
Conventional forces	400-600
Naval bases, airfields, etc.	200-300
Strategic command and control	100
Total	1,000-1,600

The weapons-to-target ratio under the same assumptions could decrease to about 3:1 while providing much the same assurance of retaliation as today. ⁴ The total number of strategic warheads under these assumptions would thus be 3,000-4,000, a reduction by a factor of 2 to 3 below the START limits. At these levels, despite the reductions, little would change in the (very uncertain and very high) casualties a nuclear war would cause, although there would be less worldwide environmental damage and fewer casualties outside cities. This is because we have not changed the assumption of targeting war-supporting industries, which are located in and near cities. ⁵ There is little question that a nuclear exchange between the United States and the Soviet Union would remain catastrophic. We see no escape from that circumstance even at much lower levels of weapons unless cities, and the industries and leadership targets located in and near them, are avoided altogether.

Page 30

Confidence-building measures that further enhance the assurance of warning of deviations from this situation are in place.

Efforts in the strategic defense field, to the extent they are carried out at all, are carried out cooperatively or sufficiently transparently to obviate the possibility of surprises.

The other nuclear powers accept appropriate strategic arms limitations.

The nuclear nonproliferation regime is preserved or enhanced.

Under such circumstances, it is difficult to posit any particular target set or number of weapons as most appropriate. There might be less need to deter the Soviet Union than some other country. Weapons systems might not be pretargeted at all. If the Soviet Union remains the main country to be deterred, if only because it remains the main military power outside the United States and its future remains uncertain, a minimum number target set may be put together by adding: (1) the most important 200-300 defense-related industrial targets, which comprise over half of the total Soviet industrial capacity, both defense-related and other; (2) the main airfields, naval bases, and other force projection targets, totaling about 200-300; and (3) a small reserve. This adds up to about 500-800 targets. An optimistic view by historical standards of survivability, availability, reliability, and similar factors leads to a weapons-to-target ratio of 2-3:1. Achieving this ratio would require a continued qualitative development program. Together, these assumptions lead to a total number of strategic weapons of about 1,000-2,000. ⁶ Clearly this is a highly uncertain number.

We have postulated, as noted above, that neither side deploys effective missile defenses. While defenses against nonnuclear tactical missiles are under some circumstances feasible and desirable, U.S. and Soviet strategic nuclear missiles continue to have an advantage over any defenses thus far considered. If partial, possibly tactical, defenses are acquired, and if their sophistication increases, a vigorous research and development (R&D) program will be needed to provide the basis for confidence in the continuing ability of the central strategic forces to penetrate against possibly poorly known, unacknowledged partial defenses. It might become impossible to maintain this confidence at the postulated 2-3:1 ratio of weapons-to-targets mentioned above. The issue, however, is beyond the scope of this report.

We note that the number of strategic nuclear weapons systems can be drastically reduced without going to a purely counter-city targeting doctrine, one consequence of which might be that nothing short of a nuclear attack on U.S. cities could be deterred. In the end, if the day comes that the relationship between the United States and the Soviet Union is like the relationship between the United States and its Western European allies, no nuclear weapons will be needed to deter the Soviet Union, whether aimed at its cities or at its military capabilities. Until then, a general targeting doc

Page 31

trine threatening both can be retained and still accommodate the type of large-scale reductions described here.

The eventual number of systems as a function of time and circumstances is less important than ensuring that the process of managing that number is cooperative to the extent possible and accompanied by efforts to improve the stability of the deterrent situation through improvements in survivability and operational practices. The two ranges presented—about 3,000-4,000 and 1,000-2,000—illustrate what might be achieved given appropriate political and technical developments.

Page 32

warheads would have to be delivered eight at a time, with the eight falling within a fairly restricted area.

At the same time, the submarines represent a significant investment and will not be replaced soon. If submarines are to carry half of the force—a reasonable assumption for planners in the absence of a sure understanding of future vulnerabilities—it will be desirable to reduce gradually the number of warheads on each missile until eventually each carries a single warhead.

For instance, if 1,800 warheads of a force of 3,000 were to be carried on the currently procured 18 Trident submarines, the 100 warheads per submarine would be distributed over 24 missiles, for a little over four warheads per missile. If 600 warheads of a force of 1,000 were to be carried by the same number of Trident submarines, there would be a little over 30 warheads per submarine or a little over one warhead per missile. This might be considered an unnecessarily expensive force per warhead, although the total cost would be less than that of the current force. On the other hand, the de-MIRVed payload would increase range, footprint, and operating area, all of which would enhance the survivability and operational flexibility of the force.

Besides submarines, the current strategic forces include fixed, land-based ICBMs and strategic aircraft. The land-based ICBMs would presumably continue to be vulnerable to a first strike and should therefore be de-MIRVed so as to remove incentives for such an attack. Their relatively low operating cost would keep them economically competitive even if de-MIRVed. They are not as survivable as submarines, but they can extract a cost in weapons that an attacker would have to use against them exceeding, or with some development, at least comparable to, the destruction achieved. To maintain a mix of systems with different vulnerabilities and other characteristics, we recommend that some fixed land-based ICBMs be retained even at the lowest force levels discussed. The present 1,000+ ICBMs could go down to 500 de-MIRVed missiles if the overall level is 3,000-4,000 warheads, and down to 200-300 if the overall level is 1,000-2,000 warheads.

As an alternative to the current basing in hardened silos, the United States could deploy either mobile land-based missiles or a multiple aimpoint system, in which many empty silos are constructed for each missile. Both systems have the advantage of extracting a greater cost from an attacker. A well-designed mobile system could require 5-10 attacking warheads to assure the destruction of a single missile. A well-designed multiple aimpoint system could exceed that number and maintain that advantage against a surprise attack without any tactical warning, as long as an attacker remains uncertain about the position of the actual missiles.

This is not the place to evaluate the relative merits of these systems. They are more costly than the current silo basing and less well understood than the submarines, both of which, in addition, have already been pro

Page 33

cured. While we believe that R&D on these systems should continue as a hedge against developing vulnerabilities in existing systems, we would not recommend their deployment under the circumstances we have projected.

Dual-capable aircraft, both long-range, U.S.-based and shorter-range sea and forward land-based, offer different characteristics from missiles: they can be recalled over a period of time after launch, are less detectable at launch than missiles, and under some circumstances may be able to carry out a variety of missions. On the other hand, they can be destroyed by air defenses and are vulnerable unless kept on short enough alert that they can take off on receiving tactical warning. Currently about 30 percent of the force is kept on tactical alert, which could be increased to 50 percent or more if needed. Aircraft involve other inefficiencies, in particular the need to allocate weapons to destroy enemy air defenses, although these drawbacks may be minimized by stealth techniques.

The flexibility gained by including aircraft in the strategic forces is thus accompanied by added technical risk and cost. Aircraft will probably continue to be included for their flexibility, as well as to complicate an attacker's plan, and as a hedge against weaknesses in the missile force. U.S.-based intercontinental bombers, whether they deliver bombs or cruise missiles, could be allocated 20-30 percent of the warheads at both the contemplated levels.

A comparable Soviet force optimization process might emphasize mobile ICBMs as the primary deterrent, with lesser dependence on SLBMs and aircraft. The Soviet Union would have to move away from its highly MIRVed silo-based ICBMs to retain a highly survivable force at reduced levels. Assuming they do, given the technologies now on the horizon, neither force would pose a serious threat to the other, except possibly through the vulnerability of their command systems. In particular, the primary components of the two forces, mobile Soviet ICBMs and U.S. SLBMs, would not threaten each other at all, since each is essentially invulnerable under current conditions.

Page 34

for a U.S. nuclear presence in Europe of fewer than 1,000 air-delivered nuclear weapon systems and no short-range ground-based nuclear systems. Implementing this step is presumably contingent on the almost certain withdrawal of Soviet short-range nuclear systems before or as the Soviet ground and air forces withdraw from Central and Eastern Europe.

Once these withdrawals are completed, the Soviet nuclear forces closest to Europe will be those in the Western military district of the Soviet Union. The INF Treaty ban on ground-based missiles in the 500-5,000-kilometer range, which applies to Soviet territory, means that the only Soviet missiles to threaten Europe will be those limited by START. Evidence suggests that the Soviet military also wishes to retain the option of air-delivered weapons, but under negotiated restrictions on the number of dual-capable aircraft, including naval air, in Europe.

Under these circumstances, and more generally given a political framework along the lines discussed above as part of the committee's proposed next step after START, we believe the level being considered by the NATO Strategy Review is appropriate and compatible with our recommended policy of nuclear cooperation. The main purposes of American nuclear forces in Europe are to deny incentives for nuclear proliferation and to help provide a secure transition to whatever new nuclear configuration the European powers and security organizations adopt. The eventual U.S. nuclear deployment in Europe needed for these purposes will depend on future events and decisions that cannot be forecast with confidence. This is good reason not to reduce the U.S. nuclear presence in Europe to zero now. It may go to zero later, depending principally on the contribution that such a reduction would make to supporting European cooperation in security matters.

There are two other nuclear forces to be considered in Europe. By the end of the decade British nuclear forces are slated to be increased to around 500 SLBM warheads based on three or perhaps four new submarines. Present plans also foresee replacement of their only other significant nuclear system, the British air-delivered gravity bombs, with a smaller number of bombs and air-to-surface missiles, to be acquired from the United States or produced jointly with France. The eventual total will be under 1,000 weapons, and perhaps considerably less than that, given uncertainties about fiscal constraints, British political developments, possible future arms control, and intra-European cooperation agreements.

France plans to increase the number of its strategic nuclear warheads to around 700 over the next decade, largely through modernization of the submarine-based force. There are pressures on the government to retire the land-based strategic missiles, to substantially cut (perhaps to 60-80) or even eliminate the new ground-based short-to-intermediate-range surface-to-surface missile Hades, and to invest in a new air-delivered tactical missile. These pressures result in part from domestic reassessments of the new stra

Page 35

tegic environment that call into question the existing French nuclear strategy, in part from perceived new opportunities for European cooperation, and in part from German concerns about the range and utility of Hades and the need for greater coordination on nuclear issues in Europe.

In Europe as in the United States, retained nuclear forces, along with their command and control systems, should be as survivable against attack as possible. Survivability of land- and air-based systems against a Soviet opponent is harder to achieve than in the United States because tactical warning times are shorter. For that reason, the bulk of British and French forces are sea-based, and eventually U.S. forces committed to Europe may be as well. Any remaining air-based forces may have to rely on longer-term strategic warning.

In summary, the committee comes to the following conclusions:

Page 36

We recognize the difficult challenge of verification, especially for the current sea-launched cruise missile (SLCM). We believe that means can probably be found to adequately assure that SLCMs on surface ships are armed with conventional, not nuclear, warheads. Over the long term, this will be easier if modernized SLCM systems are developed only with nonnuclear payloads.

This will leave only the question of dual-capable aircraft on carriers. The numbers and deployment conditions for these aircraft should be negotiated. While there should be no peacetime storage of nuclear weapons on board, the dual-capable option should be retained on carriers, as in Europe, as an element of force reconstitution to deal with future crises, such as one in which the Soviets abandoned the cooperative regime. These aircraft may be subject to general agreements to reduce or constrain all U.S. forward-based, nuclear-capable tactical aircraft still dedicated to deterring regional threats and providing political reassurance.

From these arguments flow three specific recommendations:

Page 37

the most readily transferable element (by sale, theft, or civil war) of nuclear weapons capability—the weapons themselves.

In approaching this problem, it should be recognized that U.S. and Soviet civilian and military fissile material production complexes and weapons fabrication facilities differ significantly. For environmental and safety reasons the United States has ceased producing plutonium for military purposes and is not reprocessing civilian reactor fuel for recovery of plutonium. The Soviet Union is still doing both. U.S. law separates civilian and military production processes, while Soviet processes are not as clearly separated and Soviet weapons manufacturing and handling facilities are more widely dispersed. In addition, there may be a question regarding undeclared facilities in the Soviet Union. The regime considered here thus implies an asymmetric verification burden on the United States unless the asymmetries are substantially reduced by negotiation. At the same time, redressing such asymmetries, in particular gaining Soviet agreement to cease producing plutonium for weapons, would be a clear gain for U.S. national security. We note that neither the United States nor the Soviet Union requires continued plutonium or highly enriched uranium production to meet current weapons requirements.

Reductions in the stockpile of nuclear weapons could be implemented in the following stages:

The nuclear weapons reduction regime could be further strengthened by requiring the declaration of any facility engaged in the continued fabrica

Page 38

tion of nuclear warheads. The facility could then be subject to continuous perimeter monitoring to ensure that any warheads produced were in fact replacements for permitted warheads on no more than a one-for-one basis. Tagging procedures could help ensure that the weapons turned in for refabrication were those permitted under the agreement. This type of arrangement would prevent the fabrication of undeclared material into weapons, while providing a mechanism to fabricate new weapons for modernization, reliability, or safety reasons from material from permitted weapons.

The most serious problem with this direct approach to controlling the nuclear weapons stockpile is the uncertainty about the size of the total stockpile of fissionable material at the time of the agreement. Declarations of these stockpiles could be useful as a confidence-building measure. IAEA access would help narrow estimates of past production. However, the complex operating histories of these facilities and the lack of critical past information would still prevent accurate estimates. Given the current large inventories of weapons, this would not be a problem in initial reductions. At lower levels, however, unless more complete transparency developed, this uncertainty could pose a significant risk if breakout from the reduction agreements occurred. If the problem proves to be serious as arsenals are drastically reduced, agreed quantities of weapons-grade U-235 or plutonium could be held in monitored reserve.

In such a regime, special provisions would have to be made for tritium, a fusionable isotope of hydrogen that is critically important for thermonuclear weapons. Although it decays relatively rapidly (5.5 percent per year), the weapons reductions projected in this study would assure that present tritium holdings suffice for several decades. If needed, future supplies could be manufactured in existing or new reactors fueled by monitored excess U-235 withdrawn from the weapons or peaceful stockpiles or by new, more environmentally benign, technologies outside the nuclear fuel cycle. Special provisions for supplying highly enriched uranium for naval propulsion would also be needed.

Page 39

At the same time, future testing might result in or create concerns about new nuclear weapon developments. In the absence of testing, basically new and newly optimized weapons on either side would be impossible. Consequently, new, more survivable delivery systems would have to be designed around existing warheads, and new reliability and safety problems might have to be dealt with in less than the optimum fashion by refabrication of existing weapons or modifications in nonnuclear components or operational procedures.

Politically, on the other hand, the failure of the United States, the United Kingdom, and the Soviet Union to negotiate a comprehensive test ban has become a symbol of the discriminatory nature of the NPT regime. As a consequence, the United States, which has borne the burden of this complaint because of its open advocacy of continued testing, may have lost some of its credibility as a leader in international efforts to control proliferation. There is concern that, when the future duration of the NPT is addressed at the 25th anniversary review conference in 1995, lack of progress on a test ban could jeopardize extension of the treaty. How difficult this issue will actually be remains to be seen, particularly if START is in place and the United States and the Soviet Union are making demonstrable progress toward substantial reductions in their nuclear arsenals in a cooperative security environment.

In the final analysis, most countries will make their decisions about the utility of the NPT regime or their maintenance of a nuclear option on the basis of their perceptions of their own security interests, not on the actions of the United States and Soviet Union or other nuclear weapons states on testing. The committee does not believe that a comprehensive nuclear test ban is critical to the policies recommended in this report and does not have a recommendation regarding one.