2
Military Issues

DEFINING CONVENTIONAL PROMPT GLOBAL STRIKE

It is well accepted that the United States needs conventional strike capability, that such capability should be global in its reach, and that it should be usable on a timely basis. It is less clear, however, precisely how that should be translated into what the Department of Defense (DOD) calls “requirements.” How much reach is enough? How fast is prompt? The DOD has expressed its preferred answers, but the Committee on Conventional Prompt Global Strike Capability was asked for an independent assessment.

How Should “Global” Be Defined? How Much Reach Is Enough?

It can be argued that strike capability could be needed against targets anywhere on the globe. However, even those who favor capabilities-based planning rather than planning tied to specific scenarios will acknowledge that it is unlikely that strikes will be necessary on outposts in Antarctica, Patagonia, or Tasmania, or even that the ability to threaten such strikes will be necessary. Further, if the world situation should change in this regard, there would likely be time for redeployment (over days or weeks). Because of such considerations, the laws of physics, and economic considerations, the committee concluded that it did not wish to limit consideration to systems that could strike any point on the globe within 1 hour, or even a number of hours. It was willing to make trade-offs on the matter. Thus, in what follows, “conventional prompt global strike (CPGS)” implies long-range coverage but not necessarily “global” coverage at any given time.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 27
2 Military Issues DEFINING CONVENTIONAL PROMPT GLOBAL STRIKE It is well accepted that the United States needs conventional strike capability, that such capability should be global in its reach, and that it should be usable on a timely basis. It is less clear, however, precisely how that should be translated into what the Department of Defense (DOD) calls “requirements.” How much reach is enough? How fast is prompt? The DOD has expressed its preferred answers, but the Committee on Conventional Prompt Global Strike Capability was asked for an independent assessment. How Should “Global” Be Defined? How Much Reach Is Enough? It can be argued that strike capability could be needed against targets any- where on the globe. However, even those who favor capabilities-based planning rather than planning tied to specific scenarios will acknowledge that it is unlikely that strikes will be necessary on outposts in Antarctica, Patagonia, or Tasmania, or even that the ability to threaten such strikes will be necessary. Further, if the world situation should change in this regard, there would likely be time for redeployment (over days or weeks). Because of such considerations, the laws of physics, and economic considerations, the committee concluded that it did not wish to limit consideration to systems that could strike any point on the globe within 1 hour, or even a number of hours. It was willing to make trade-offs on the matter. Thus, in what follows, “conventional prompt global strike (CPGS)” implies long-range coverage but not necessarily “global” coverage at any given time. 27

OCR for page 27
28 U.S. CONVENTIONAL PROMPT GLOBAL STRIKE Decision; prepare Actionable warning for “instant (triggering event) response” Tw Ts Ta Te T0 Strategic Execution Time of effect Forces alerted; warning; ordered final planning preparations begins begin Execution time FIGURE 2-1 Time line for a case in which execution follows a substantial period of warn- ing, contingent decision, and preparation, with the final decision made quickly. NOTE: Ts—time of strategic warning; Ta—time at which forces are alerted; Tw—time of actionable warning; Te—time of execution order; T0—time of effect. Intelligence and command-and- control activities are critical factors in the overall time line for any CPGS system. The Figure 2.1 time between actionable warning, Tw, and the order to execute, Te, includes the time for bringing weapons to ready-to-launchb&w SOURCE: Adapted, with permission, from status. Paul K. Davis, Russell D. Shaver, and Justin Beck, 2008, Analytical Methods for Assessing Capability Options, RAND, Santa Monica, Calif., p. 61. © 2008 RAND Corporation. How Should “Prompt” Be Defined? Is “Within 1 Hour” a Good Criterion? The same type of question asked about the definition of “global” can be asked about the definition of “prompt.” Obviously, an instantaneous strike capability would be desirable if achievable. But what is feasible, valuable, and affordable? What Is Feasible? As discussed in detail in Chapter 4 of this report, global strike operations involve a long and complex end-to-end chain of activities. Figure 2-1 explains this chain of activity schematically for a case of particular interest. Two of the most important factors are execution time and the class of targets that can be effectively attacked. “Execution time” is the time between the President’s order to execute an attack (at a time Te) and when the target is affected (at a time T0).1 The figure envisions earlier strategic warning (at a time Ts) and an early contingent decision (at a time Ta) that initiates detailed preparations. When “actionable warning” occurs (i.e., when the triggering event occurs, at a time Tw), the final decision to proceed is made quickly (perhaps in tens of minutes, based on precedent, at a time Te), and execution begins. 1 This assumes that neither aircraft nor missiles are launched prior to the President’s authorization to strike.

OCR for page 27
2 MILITARY ISSUES Both experience and an examination of possible futures with the DOD’s and the committee’s scenario-based analysis suggest that this case will be the norm for CPGS: crises and opportunities will seldom arise as complete surprises with tar- gets that merit a conventional “strategic” military strike “popping up.” The more important cases are likely to be ones in which events develop over time—perhaps many hours, days, weeks, or even months. As a result, considerable preparation, deliberation, and tentative decision making can occur ahead of time, even if final developments move rapidly—that is, even if execution itself is and needs to be fast (i.e., prompt). In such cases, execution time itself may be the limiting factor for timeliness. For example, there could be an advance policy decision to strike at a terrorist leader if he could be located, or at shipments of nuclear materials to a potential proliferator if these materials were identified and located. This said, the other factors are also critical to the mission’s success (and will sometimes deter- mine timeliness as well). If CPGS is to be pursued, all aspects of the end-to-end process must be pursued vigorously. Enablers (e.g., intelligence, command and control, and targeting) for CPGS capabilities are discussed later in this chapter. First, however, execution time, target type, and some other key operational fac- tors are addressed. In addition to execution time, a second key measure is the class of targets against which a CPGS can be used. These classes range from the less difficult (soft point targets such as people or stationary vehicles) to the more difficult (hard and deeply buried target complexes). Other factors (such as whether targets are moving) are also important, but Figure 2-2 uses these first two, execution time and target class, to make some important distinctions. Figure 2-2 is to be understood as follows. Consider first the solid contour marked “A.” This contour indicates that tactical aircraft or cruise missiles are usable for global strike in situations that permit execution times greater than a few hours and which involve targets that may be hard, and may even be buried, but that are not large-area deeply buried targets. This is the region below and to the right of contour A. If the launch platforms are not both located where they need to be and poised for action (e.g., if an aircraft carrier is in the appropriate general region but conducting exercises that impede mission execution), the execution time might be more like 10 hours; and if ships or submarines would need to deploy to new loca- tions, the execution time might be days (i.e., more like 100 hours than 10 hours). With existing systems, even the most distant target can be reached in the time it takes for a bomber to fly from Guam, Diego Garcia in the Indian Ocean, or the continental United States (CONUS)—perhaps 10 to 20 hours of actual flight time, plus whatever additional time is needed to generate tanker support. Moving upward in the figure to the light contour marked “B,” note that bombers could have the same range of execution times if they were forward deployed, but if not, execution time would be on the order of 10 to 20 hours (as just described). Bombers, however, would have more capability against complex deeply buried targets because of their great payload.

OCR for page 27
30 U.S. CONVENTIONAL PROMPT GLOBAL STRIKE Target Character (increasing difficulty) Large, deep, No capability underground, B ambiguous Bombers Large, deep, underground A Tactical aircraft, cruise missiles C Ballistic or hypersonic Hard point cruise missiles target Point target 10 2 1 10 Execution Time (hours) FIGURE 2-2 A depiction of capability space for global strike. NOTE: The dashed lines with two-headed arrows indicate that the boundary moves rightward unless the platforms in question are appropriately deployed and postured. SOURCE: Adapted, with permis- sion, from Paul K. Davis, Russell D. Shaver, and Justin Beck, 2008, Analytical Methods Figure 2.2 for Assessing Capability Options, RAND, Santa Monica, Calif., p. 52. © 2008 RAND color Corporation. Suppose that the mission in question requires shorter execution times. Instan- taneous would be ideal, but according to Figure 2-2, the only feasible region is for relatively simple targets and times on the order of 1 hour or more; that is, the dotted contour marked “C.” The only weapon systems being contemplated for this region of the capability space for global strike are ballistic missiles and hypersonic cruise missiles.2 2 In special circumstances, any of the contours could move leftward somewhat. If orders were received ahead of time and missiles were at short range, strikes could occur in tens of minutes. If tactical air forces were already in the air and merely had to be diverted to a new target, they could fly 500 miles or so in roughly an hour. For planning purposes, however, more realistic times should be used. Many factors come into play, including maldeployment; being engaged in other missions or not being at a high state of alert; distance from the target; delays associated with receiving, interpret- ing, and verifying a presidential directive; or the need for detailed at-the-time planning to avoid air defenses, avoid collateral damage, and orchestrate (for some appreciation of such difficulties, see Michael R. Gordon and General Bernard E. Trainor, 2006, Cobra II: The Inside Story of the Invasion and Occupation of Iraq, Pantheon Books, New York, Ch. 9). Thus, the committee has treated “a few hours” as a reasonable analytic lower limit for normal air breathers and 1 hour as a reasonable lower limit for ballistic or hypersonic cruise missiles.

OCR for page 27
31 MILITARY ISSUES What Is Valuable? Would an Execution Time of 1 Hour Be Important? If a 1-hour execution time is feasible, then the next question is whether it would be of sufficient value to justify the required investment funds and priority. The answer is not immediately obvious; numerous past strikes have been accom- plished without such extreme promptness. Nonetheless, the DOD has argued strongly that the capability for prompt strike would be very important, and it has defined “prompt” as an execution time of 1 hour or less. The view of the Secretary of Defense has had firm supportive testimony from the Secretary of State and the Joint Chiefs of Staff. The committee, however, was charged to make an indepen- dent assessment, and it concluded that credible instances can readily be imagined for which a CPGS capability would be quite valuable. Specific scenarios can be useful in judging whether a capability is needed. The committee identified three scenario classes (see below), the concreteness of which aided discussion, each of which creates an argument for short execution times such as an hour.3 These scenario classes are as follows:4 1. Terrorist leaders. The first scenario class envisions striking a gathering of terrorist leaders expected to be meeting in a particular location for a short period of time. This kind of information is sometimes obtained by electronic intercepts or by intelligence from operatives on the ground. Such a scenario is analogous to actual events from the past decade. In the 1990s, it was sometimes known that al-Qaeda leaders would be meeting, but the time and place were not known pre- cisely until late in the game. In one instance, the President had such intelligence but decided not to strike because of concerns about collateral damage. In another instance, the President authorized a strike in advance, but when the cruise missile subsequently struck the camp, Osama bin Laden was no longer there (having left perhaps several hours earlier according to some reports).5 During the first phase of the invasion of Iraq in 2003, intelligence reports purportedly located Saddam 3 Members of the committee differed about which illustrative scenarios should be regarded as the most important and plausible. Some members resonated most with scenario classes 1 and 2, while having deep reservations about class 3. Others resonated with developing deterrence-enhancing capa- bility for class 3 scenarios, while being less persuaded about classes 1 and 2. Overall, the committee believed, of course, that scenarios are merely examples intended to provide concretely some of the many possibilities for which capabilities may be needed. Reasonable people may differ about whether CPGS would be a sensible option to use in any particular scenario, but forgoing the capability would mean not having the option even when it clearly would be needed. 4Whether a CPGS capability could actually be employed in the various scenarios would depend on contextual details, such as the strategic environment and political circumstances and, of course, on whether the capabilities ascribed to the CPGS options are successfully achieved and supported by enablers such as described later in this chapter and in Chapter 4. 5 Many aspects of the event are described in Richard A. Clarke, 2004, Against All Enemies, Free Press, New York, Ch. 8. These include the early intelligence, days of preparation and decision making, the detection by Pakistan of some forward-deployed naval forces prior to the attack, and complicated politics.

OCR for page 27
32 U.S. CONVENTIONAL PROMPT GLOBAL STRIKE Hussein at the Iraqi presidential compound at Dora Farms. Aircraft and cruise missiles already in the region were reassigned to strike at the reported site, but Saddam was not there. These historical instances illustrate that execution time can be a critical variable.6 2. WMD shipment. A second class of scenarios is illustrated by the striking of a transshipment of one or more weapons of mass destruction (WMD), perhaps by truck or ship. Experience tells us that intelligence may exist about when a ship- ment is planned or may be en route, or where loading, unloading, or temporary stops may occur. Details may be lacking until late—perhaps when those doing the transporting stop for rest or maintenance, or when delays occur at a port, bridge, or border, including stops associated with routine inspections. In some cases, it might be possible for local police or military forces to make an intercept; in other cases, they might not be available to do so, or political considerations might pre- clude such action. Similarly, political factors might preclude the use of CPGS, as might the risk of collateral damage (caused, for example, by the dissemination of radioactive or noxious materials at the target). In other cases, a prompt strike with U.S. assets might be necessary and appropriate. One such scenario was sketched by former Secretaries of Defense Harold Brown and James Schlesinger in an article supporting CPGS capability.7 3. Immediate response or preemption of imminent attack. A third class of sce- narios is illustrated by imagining that the United States, U.S. forces, or allies are about to be (or have been) attacked and that an immediate response is needed to prevent dire (or additional dire) consequences. Possibilities include conventional or nuclear missile attacks on U.S. cities, on U.S. forces or other military assets, or on U.S. allies. When such scenarios are examined in operational detail, the importance of prompt strike becomes evident: a response time of less than 1 hour might be essential to avert great losses (e.g., loss of numerous satellites crucial to U.S. command and control). In such cases, the CPGS might be the leading edge of a much larger military campaign. That is, the use of CPGS would not have to be large or decisive in itself. It could be very important despite being small. History is again useful. The very earliest part of the military campaign against Iraq in 1991 was a small, specialized, and focused attack on Iraqi air defenses. More gener- ally, military campaigns often begin with relatively small attacks on particular military targets that have leverage—in defeating or disrupting the adversary’s air defense; command, control, and communications (C3); or most-feared weapons. And for some of these, timing is critical. With today’s forces, such small leading- edge attacks might be conducted by Special Operations Forces (SOF) or other 6A sober reading of history also reminds us how frequently human intelligence is wrong. 7 Harold Brown and James Schlesinger. 2006. “A Missile Strike Option We Need,” Washington Post, May 22. As an example of how complex a decision to use CPGS could be, consider that an “easy” strike on some platform carrying WMD might not be possible for fear of releasing toxic chemicals or radiation.

OCR for page 27
33 MILITARY ISSUES forward-deployed assets. In the future, such attacks might be accomplished with prompt global strike assets—with less need for forward basing and less concern about failure due to long flight times or detection. Is 1-Hour Capability Affordable? Having established that using a 1-hour criterion for promptness makes sense from the viewpoint of feasibility and that credible cases can be envisioned for which it would be valuable, the remaining issue is cost (and related trade-offs). Without elaboration here, since options on the matter are described later in this chapter and in Chapter 4, it can be stated that achieving 1-hour capability is rela- tively inexpensive (in DOD terms), although it could be quite expensive if the less-expensive CPGS options are ruled out for one reason or another. Conclusions on Definitions That Make Sense In summary, the committee concluded that defining conventional prompt global strike should not be too stringent with respect to the “global” criterion, but that the 1-hour criterion in defining “prompt” was sensible. That said, it could imagine systems (as discussed later) that might not quite make the criterion of promptness but that would be close enough so that they should be considered among the options treated. ATTRIBuTES AND TEST CASES FOR COMPARING OPTIONS Attributes Having discussed execution time and target classes, let us next consider the longer list of factors important to CPGS capability, displayed in Table 2-1. The first three have already been discussed (the second is a subset of the first); the others have not. Test Cases for Operational Scenarios The previous section and Table 2-1 outline a way to compare options tech- nically, but a more “operational” comparison is also needed, one more closely related to military utility in actual missions. One way to do this involves construct- ing a spanning set of scenarios construed as test cases. A spanning set needs to stress the alternative systems with respect to timeliness, target class, geographic depth, adversary air defenses, and the need to search and find targets at the time that weapons arrive. Drawing on the three scenario classes described earlier but thinking about the various operational issues to be considered, the committee constructed a set of six

OCR for page 27
3 U.S. CONVENTIONAL PROMPT GLOBAL STRIKE TABLE 2-1 Variables Important to Conventional Prompt Global Strike Capability Feature Meaning Execution time Time from the order to execute until effects are achieved (assuming no contingent launch of aircraft or missiles) Final positioning and planning Time from the order to execute until systems are located and time supported properly and mission planning is accomplished. May be a part of execution time. Relevant prompt global strike Target classes (e.g., soft point targets, hard point targets) targets Lethality of weapons The probability of achieving the target damage sought, assuming successful delivery Ability to attack moving targets Ability to attack targets that are on the move when attacked Volume of fire The number of weapons available for use in a given strike Controllability Safety and security with respect to the weapons and their employment Geographic coverage The targets that can be reached without repositioning, e.g., without substantial sailing time or rebasing of bombers Risk of operational side effects The likelihood or potential for negative effects on, e.g., nuclear deterrent or Special Operations Forces operations At-the-time strategic factors Need or non-need for overflight rights, allied cooperation, or forward basing Availability Earliest initial operational capability Development risk The likelihood of serious development failures or slippages; a function of the technical readiness level of components, full-system testing, industrial base, and organizational competence Confidence in high reliability Assuming “successful development,” how likely would the system be to have high reliabilities (e.g., 0.95 rather than 0.6)? Evolutionary potential Value of the option in laying the technical and operational groundwork for subsequent, more advanced systems test cases as a reasonably good (but not exhaustive) spanning set. All are variants of the three scenario classes mentioned above, but they have been reformulated to focus on operational issues. They can be summarized as follows: 1a. Immediate response to threat (soft targets, near-surface targets, moder- ate air defenses); 1b. Immediate initial response to threat (soft targets, deeply buried targets, advanced air defenses);

OCR for page 27
35 MILITARY ISSUES 2. Attack of deeply buried WMD facility (hard, deeply buried targets, advanced air defenses);8 3a. Intercept of terrorism-related WMD transshipment (soft targets, near- surface targets, moderate or no air defenses, mobile targets); 3b. Attack of terrorist leadership (soft targets, possibly under ground, and possibly moving; moderate air defenses); and 4. Attack of key nodes as leading edge of military campaign. This set of test cases (scenarios) stresses the capabilities of candidate systems in different ways and, by and large, highlights the strengths and weaknesses of the alternatives. The quality of the testing, of course, depends on numerous details such as the quality of air defenses imputed to the adversary. The test cases should be stressful, but reasonable. A REPRESENTATIVE SET OF OPTIONS FOR CPGS Seven Options Analyzed The committee analyzed seven of the very large number of possible CPGS options, informed by results of the DOD’s ongoing analysis of alternatives (AoA) and the committee’s independent thinking. These options were as follows: 1. Existing systems: A combination of tactical aircraft, cruise missiles, and long-range stealthy bombers, supported as necessary with tankers; assets for com- mand, control, communications, computers, intelligence, surveillance, and recon- naissance (C4ISR); combat search and rescue (CSAR); and suppression of enemy air defenses. Some of these systems have very high accuracy and large payloads and, if pre-positioned in areas where targets are likely to emerge, have relatively rapid response times—but almost always measured in a few hours, not less. The exceptions are not a good basis for planning (see footnote 2 in this chapter). 2. Conventional Trident Modification (CTM): The option proposed in the President’s budget would use existing Trident missiles on nuclear-powered ballistic missile submarines (SSBNs). The intent is to have, on each of the 12 deployed SSBNs, 2 of the submarine’s 24 missiles armed with conventional rather than nuclear warheads. In its initial version, the CTM would have four kinetic energy projectile (KEP) warheads per missile, in a normal Trident reentry body augmented with an inertial navigator, Global Positioning System (GPS) receiver, and a modest maneuvering capability to strike precisely at GPS 8A 2005 National Research Council report entitled Effects of Nuclear Earth-Penetrator and Other Weapons concluded, among other things, that many strategic hard and deeply buried targets can only be attacked directly with nuclear weapons. See National Research Council, 2005, Effects of Nuclear Earth-Penetrator and Other Weapons, The National Academies Press, Washington, D.C.

OCR for page 27
36 U.S. CONVENTIONAL PROMPT GLOBAL STRIKE locations on the ground. The accuracy required is quite high, and testing to confirm that it has been achieved is essential. Lethality can be achieved against soft, relatively small-area targets with diameters up to about 10 meters (more if multiple warheads are appropriately targeted across an area), and can be significant against some point structural targets, as discussed in Chapter 4. The lethality of warheads would evolve over time (e.g., with a bent-nose warhead, as described in Chapter 4, which would improve effectiveness against some terrain-protected targets). 3. CTM-2: An option developed in the committee’s activities, this would remove the third stage of the Trident missile, thereby increasing payload capacity and conventional-payload options, while still achieving 4,000 nautical mile (nmi) ranges. Developing the CTM-2 would require relatively little modification of the existing missile, but it would require a few additional flight-tests to validate the modified guidance and control software. The base configuration would be as with the CTM—each SSBN would have two or more CTM-2 missiles. An initial ver- sion of the CTM-2 would use the CTM’s KEP warheads or the large penetrator reentry vehicles (RVs) that are more effective against hard or buried targets. Even the initial version could exploit bent-nose technology, as described in Chapter 4. Later versions could include ballistically delivered unmanned aerial vehicles (UAVs) with intelligence, surveillance, and reconnaissance (ISR) packages and weapons to reacquire and attack mobile targets. 4. Submarine-Launched Global Strike Missile (SLGSM): Conceived by the Navy as a potential mid-term system following up on CTM experience, the SLGSM would be launched primarily from existing nuclear-powered guided missile submarines (SSGNs) but could also be launched from SSBNs to extend coverage; it would have intermediate range and a single RV payload larger than the total payload of the CTM (it could also carry CTM-like payloads, of course). Because the SSGNs have a variety of missions, there would be some operational side effects related to the use of Special Operations Forces. 5. Boost-glide missile (initial version, CSM-1): The Air Force has proposed the Conventional Strike Missile (CSM) based on the proposed Minotaur launch vehicle described in Chapter 4. This would have boost-glide capability, extend- ing range and permitting large maneuvers to avoid unintended overflight. The Air Force proposed that its initial boost-glide missile have an 800-second glide in the atmosphere. The committee believes that to reduce technical risk and field a capa- bility quickly, the initial design should be limited to the delivery of a penetrator warhead or larger KEP warheads, rather than the munitions proposed for the next option (the “second version” of CSM). 6. Boost-glide missile (second version, CSM-2): The Air Force proposes a second, later variant of CSM with a 3,000-second glide in the atmosphere to increase its range and maneuverability. It would have more capability, more development risk, and a longer time to deployment than CSM-1. In the Air Force concept, both the initial and second versions would have the capability to slow

OCR for page 27
37 MILITARY ISSUES enough to dispense a wide variety of the Air Force’s arsenal of air-launched muni- tions (e.g., Small Diameter Bomb, Joint Direct Attack Munition [JDAM], and submunitions of various kinds, including the BLU-108 with Skeet warheads and possibly modules for ISR, battle damage assessment [BDA], and reattack). 7. Hypersonic cruise missiles: Hypersonic cruise missiles are being inves- tigated and, if developed, could be used for CPGS if forward deployed (e.g., on SSGNs) or launched from long-range aircraft. Such hypersonic cruise missiles would have only medium range but would have considerable capability for ter- minal-phase dispensing of smart munitions and ISR modules, and if slowed to subsonic speeds, could execute a search to reacquire and strike moving targets. Why Not just use the Minuteman? Although the committee considered all of the many options used in the DOD AoA and invented additional options and variants on its own, an obvious question for some readers may be, Why not just use the Minuteman missile? The committee investigated the option, but concluded that, although technically viable, the Minuteman III is not a realistic contender for reasons involving the complexity of issues described in Appendix I of this report. In short, the required renegotiation of the Strategic Arms Reduction Treaty, combined with “not-in-my- backyard” issues, presents substantial challenges to deploying intercontinental ballistic missiles (ICBMs) in locations such as Hawaii or Guam, particularly in the near term, as compared with other challenges associated with CTM deploy- ment. Furthermore, reducing the number of conventionally armed ICBMs to be deployed in Hawaii or Guam or using mobile units rather than silos will not offset these concerns. Like the Air Force, the committee gauged the pros and cons of land-based conventionally armed ballistic missile systems and decided against recommending the Minuteman as a near-term option. ANALYSIS Functional and Technical Analysis With the foregoing as background, the committee’s independent analysis of CPGS alternatives follows. Time and resource constraints precluded an in-depth study, but the analysis provides useful insights. Table 2-2 summarizes a qualitative technical comparison of the options, focusing only on military-technical issues. (The knotty “nuclear ambiguity” issue identified by some observers, including some in Congress, is discussed in Chapter 3.) Table 2-3 summarizes the committee’s rough estimates of earliest availability and total 20-year system costs. The estimates are informed by material from the DOD and from contractors and by some committee members’ long experience in capability development.

OCR for page 27
50 U.S. CONVENTIONAL PROMPT GLOBAL STRIKE The CTM succeeds in the test cases shown in Table 2-4 (columns one and four) because of its promptness, but it has no ability to attack hard facilities, only limited capability against terrain-protected targets, and no ability against moving targets.16 The score for the “Attack Terrorists Fast (Leaders or WMD)” test case (third column in Table 2-4) is mixed because the targets might or might not be moving. In the latter case, the CTM and CTM-2 would be capable. Given a UAV, the CTM-2 could have capability against moving targets as well. The SLGSM and CSM-1 have no moving-target capability and so are rated the same as CTM and CTM-2. The CSM-2 (with UAV) and hypersonic cruise missiles would have moving- target capability, but, depending on load and possible maneuvers, might or might not be able to cope with advanced air defenses. Such defenses would probably not apply for the “Attack Terrorists Fast (Leaders or WMD)” test case, however. For the other cases, the table shows results as “good,” but “marginal/mixed.” These systems should do relatively well if loaded with payloads not requiring them to slow up in the terminal phase. With payloads requiring the slow-up, it is unclear whether they would be able to deal well with advanced air defenses. It is a matter of speculative disagreement. On balance, this “Attack Terrorists Fast (Leaders or WMD)” test case seemed to merit a score of “good.” By and large, all of the planned options would achieve significant capability in most of the test case scenarios. The moving-target cases pose the most serious challenge. It is imperative to understand, however, that the evaluations are for “good cases,” such as when a deeply buried target has a special vulnerability, or when moving targets are out in the open where collateral damage is not a concern. None of the options would have capability in all versions of any of the scenarios, but the analysis indicates that the “good cases” are plausible enough to plan for. One of the factors complicating this analysis is that effectiveness depends on weapons loading, and most of the options have multiple potential loadings. The DOD will want more explicit analysis regarding the ability to adjust loads (and concepts of operations) based on strategic warning. For example, if a plausible threat to nuclear SSBNs arose, SSBNs could return to maximally secure opera- tions; or more likely, if the need to attack hardened sites became evident, unitary penetrator weapons might be favored over KEPs. How long would it take for reloading? Some weapon systems could have a mix of weapon types at a given time; others could not. 16 The committee evaluated the CTM as it was described to the committee, but noted that the CTM could be given more capability against terrain-protected targets relatively easily and that eight, rather than four, RVs could be carried on a given missile, thereby increasing the volume of fire and capabil- ity against area targets or targets for which geospatial location information was slightly unreliable. Flechette density could also be increased, with more numerous, smaller, flechettes.

OCR for page 27
51 MILITARY ISSUES ENABLERS OF CONVENTIONAL PROMPT GLOBAL STRIKE So far this chapter has focused on the potential effectiveness of the CPGS platforms and weapons. It has pointed out that in many and perhaps most of the plausible circumstances of CPGS employment, there would be warning time per- mitting intelligence collection and synthesis, precise targeting, preliminary deci- sion making, and other preparations. In such cases, the measures discussed (e.g., execution time, lethality) would be especially important. In some cases, however, many of the “enabling activities” would also need to be done quickly—if not in minutes, then in hours or a few days, rather than weeks and months. Whether or not additional time was available for them, the enabling activities would be criti- cal, not merely desirable. The material that follows touches on such enablers, including the challenge of ensuring that aimpoints are properly geolocated. Challenges for Enablers The use of long-range missiles to deliver conventional weapons accurately enough to damage targets requires information that is significantly more difficult to provide than that needed for nuclear weapons or for conventional weapons delivered by aircraft or short-range missiles. The primary issues relate to the following: (1) command and control (C2) that reserves decisions to the National Command Authority, while delivering information to the weapon system quickly for a short overall execution time; (2) the provision of the information necessary for accurate weapon delivery to a specified aimpoint; (3) the accurate location of aimpoints; and (4) target detection. The following paragraphs discuss these issues in order. At the outset it is noted that the timely information available for early versions of CPGS, such as CTM, will be sufficient for many important targets of interest. However, it will be some years before the information will be sufficiently timely for a broader set of targets, especially those that are moving while being attacked. That is, one should assume considerable evolution over time and consider the possibility of follow- ing a dual track; deploy quickly a system able to expand the present capability and use existing technology (including technology for enablers) while pursuing work toward longer-term capabilities as technology advances (and, possibly, new missions are identified). Command and Control The command-and-control problem for the CTM is greatly simplified by adapting modestly the existing C2 system used for the nuclear-armed Tridents— the system developed for controlling the major strategic nuclear system of the nation. Its CTM variant17 would be both effective and cost-effective. Such is not 17 The proposed program would add a number of hardware and software provisions to prevent

OCR for page 27
52 U.S. CONVENTIONAL PROMPT GLOBAL STRIKE necessarily the case for alternative future systems, which are being conceived as strictly conventional, in part to get away from the concerns expressed by some observers about mixed loads of weapons (i.e., mixes of nuclear and conventional). Such future systems, as currently conceived, would not be tied into the nation’s C2 system for strategic (albeit non-nuclear) weapons, even though decisions about the use of such strategic weapons will almost surely be made at the highest level of the government. It is unclear to the committee whether physical control of the weapons will be strong enough, and whether the mechanisms will exist for an efficient matching of targets and weapons—something routinely considered in the existing nuclear C2 system. The committee believes that any CPGS option should be regarded as “strategic” and demanding of extremely high safeguards. Accuracy The accuracy required for CPGS is on the order of meters in each of three dimensions (two dimensions on the surface of the ground and one “vertical” dimension relating to target elevation and fusing altitude). Such results have been achieved with short-range systems after several iterations of guidance- system details (in Chapter 4, see the subsection entitled “Guidance, Navigation, and Control Accuracy Issues”), but the CTM or other future systems will need to achieve excellent accuracy from the outset. Even with results from testing, ensuring that a single or a handful of shots will achieve such accuracy every time is very challenging. Several studies have shown that bias errors will be trouble- some in some cases for the CPGS systems.18 Given the limited payloads and small radius of effects for conventional weapons, there is little room for error in any of the three dimensions. These CPGS systems will require GPS updates to correct inertial guidance systems toward the end of flight and perhaps will need to include additional systems to correct for bias errors (a capability not available for early systems). Also, the requirement to protect the weapon itself from heat and other atmospheric effects implies correct compensation for aerodynamic effects over significant time periods, during which velocities and orientations are changing significantly. Development and full-scale, system-level testing are essential. That testing is one of the highly attractive features of the proposed CTM program and would be important whether or not the CTM were deployed. If CTM research, launch of a nuclear weapon rather than a conventional weapon when a conventional-weapon attack was intended. Safety, in this case, would rely not only on rigorous procedures but also on well-tested hardware and software safeguards. The committee found this aspect of the CTM’s design to be es- pecially attractive. 18Accuracies are usually quoted in circular error probable (CEP) terms, with CEP measured around the intended impact point. Bias errors are another matter: the intended impact point may be displaced significantly from where the actual target is located. One contributor to that error is target location error (TLE), and it is not a function of the guidance system or the number of times that the missile has been flight-tested.

OCR for page 27
53 MILITARY ISSUES development, testing, and evaluation (RDT&E) are terminated, other systems will start from a much more primitive technological base. Aimpoint Location Also challenging is determining the location of the target to be struck. CTM and its early evolutions will be limited to striking locations known before launch—locations that contain the right targets and have reliably established coordinates. Targets will need to be stationary for at least 1 hour to be struck effec- tively, unless their arrivals at fixed locations can be accurately predicted. Future versions of the weapons might allow an in-flight update of the target location, but the targets would still need to be stationary for the times between last update and impact, which might be tens of minutes. Any error in the actual location of the target (target location error, or TLE), including bias errors, needs to be significantly smaller than the weapon’s range of effects, which means that TLE needs again to be on the order of a few meters. While such accuracy is difficult in the two dimensions of ground position, it is even more difficult currently for the vertical dimension. That situation is improv- ing, as discussed below. Horizontal accuracy can be improved significantly by measuring position relative to a known “fiduciary point” recorded by the National Geospatial-Intel- ligence Agency (NGA). NGA produces an extensive Digital Point Positional Data Base (DPPDB) of reference photos with embedded and accurately determined reference points, or fiduciary points. The DPPDB enables mission planners to derive in minutes accurate geographic coordinates for a visually observable point—for example, from a recent photo taken in the field. The DPPDB today covers significant areas in known hot spots around the world. In the event that a target emerges in an area of the world not covered by the DPPDB, the process of reducing TLE to an acceptable level can take days or even weeks. In cases where the TLE is only somewhat degraded, CTM with KEP warheads could still be effective by pattern-targeting its RVs over the target to increase the area over which lethal effects are spread.19 Target Detection and Localization In addition to the problems of location and control described above, there remains the challenge of finding a target in the first place, especially in the case of a movable, but not continuously moving, target at rest. If U.S. forces are deployed in the area and maintain control of the airspace, a high-altitude system such as 19 “Pattern-targeting” as used here means that the individual shots would be directed at slightly different aimpoints in a pattern around the nominal aimpoint so as to increase the probability of a successful impact despite small residual bias errors in the target’s actual location.

OCR for page 27
5 U.S. CONVENTIONAL PROMPT GLOBAL STRIKE Global Hawk could detect and locate a target with synthetic aperture radar or light imaging—assuming suitable fiduciary points. However, where satellites are needed to detect the target, the detection will be episodic at best. And sometimes it will be necessary to rely on human intelligence and to translate information like “The ship is moored at Pier 9” into coordinates. Today, with the war in Iraq requiring maximum support by scarce ISR resources, the competition for satel- lite resources is severe, but it is difficult even in more normal times. There are significant limits on how quickly satellite tasking can be adjusted to accommodate even the highest intelligence priorities. Fortunately, when there is activity some- where in the world that might call for use of a CPGS weapon, there is likely to be heightened interest for other reasons as well, so that the satellite systems will be concentrating their coverage on the problem area. The National Reconnaissance Office (NRO) developed plans and programs in the 1990s to address the revisit problem by increasing the area of Earth observed per day by a factor of about 100. While there have been program cancellations and delays, about a factor-of-10 improvement should have been effected by the time the CTM is deployed. The planned Air Force Space Based Radar Program, which is in the initial stages of implementation, will provide relatively persistent coverage of chosen points on Earth, with maximum gaps between available obser- vation times measured in tens of minutes.20 This revisit time does not guarantee the ability to track a target with certainty to see where it stops, but it is effective where congestion is not too high, such as in areas where potential CTM targets would most likely exist. If this program is deployed, by 2020 its ability to find targets for CTM and its successors will change from being episodic to being relatively reli- able. In addition, these new systems (the NRO programs and Space Based Radar Program) will allow acceptable TLE determination after a few observations—after approximately 30 minutes with systems of the Space Based Radar Program and within a day or less for those of the new NRO programs, without using fiduciary points, if this capability is built into these systems. Only when sufficient work has been accomplished in advance to provide very accurate fiduciary points near a target will CTM and its successors be provided with information about fixed aimpoints with a TLE small enough to expect that the weapon will achieve the programmed damage. A patterned attack with multiple RVs could mitigate these problems. Adding more RVs to the CTM, as discussed elsewhere in this chapter, would increase the ability to do such patterned attacks. For fixed facilities, this is straightforward. For other potential target locations that have a time-dependent value, such as a pier that is a valuable target when it has WMD cargo ready to load but not when it is empty, the detection of the 20 Congressional funding for development of the Space Based Radar Program has been limited over the past few years. For an overview of the program, see General Accounting Office, 2004, Defense Acquisitions: Space-Based Radar Effort Needs Additional Knowledge Before Starting Development, GAO-04-759, Washington, D.C.

OCR for page 27
55 MILITARY ISSUES increased value will probably be provided by satellites or other sources episodi- cally. Because CTM and other envisioned CPGS systems will rely on ballistic or high-altitude delivery of the weapon, terrain masking and other protection capabilities will be relatively ineffective, and additional data concerning the surrounding area will most likely not be needed. However, when payloads must be delivered using maneuvering RVs with limited corner-turning capability, the requirement for information to support flight planning will increase. Such data should still be available from the NRO systems and would be available from the space-based radar when the payloads are delivered by canister into what is in effect an atmospheric-flight terminal engagement. The requirement for accurate position location is then reduced significantly, as is the constraint on effectiveness on mobile targets, because that engagement can use onboard sensors to control terminal maneuvers—for example, GPS plus fine imagery to move the aimpoint by a few meters. Because the initial CTM system will have so few system tests before deploy- ment, the tolerance for error in the enabling systems will be small. It will be very important to have tests by the NGA using precise measurement of errors in the location of known fiduciary points to detect and overcome bias errors. In addition, NGA will need to determine accurately and quickly fiduciary points in the vicinity of potential targets to use when a satellite happens to locate a target of interest. Later versions of CPGS could use a more robust system of enablers, with world-networked capabilities to share information being deployed, and signifi- cantly improved satellite reconnaissance capability to find targets and track them. Later CPGS versions themselves will have better terminal systems and will also require less external information in order to be effective. Exactly how those capa- bilities will be combined is unknown at this time, but the constraints on the CTM itself are well defined and subject to analysis, simulation, and, most importantly, testing of the enabling systems. Ensuring Safety and the Absence of Errors A strategic system such as CPGS must be designed with extremely high lev- els of protection to ensure safety and the absence of mistakes. As the world was reminded recently when nuclear-armed missiles were accidentally loaded onto a B-52, transported to another air base, and essentially left unattended for many hours, mistakes can happen—even with the nuclear weapons to which the United States has applied special precautions for decades.21 21 Inresponse to the incident of unauthorized movement of nuclear warheads from Minot Air Force Base, North Dakota, to Barksdale Air Force Base, Louisiana, Secretary of Defense Robert Gates authorized an independent investigation of the incident resulting in the following report: Defense Science Board, 2008, Report on the Unauthorized Movement of Nuclear Weapons, Permanent Task Force on Nuclear Weapons Surety, Washington, D.C. For additional reading, see Joby Warrick and Walter Pincus, 2007, “Missteps in the Bunker,” Washington Post, September 23.

OCR for page 27
56 U.S. CONVENTIONAL PROMPT GLOBAL STRIKE The proposed CTM program raises at least the possibility of an accidental launch of a nuclear weapon instead of the intended launch of a conventional weapon because (1) both kinds of weapons would be carried on the same subma- rine platform and (2) prompt global strikes may often allow little time for second checks. General Considerations The committee discussed at some length this matter of the need to ensure safety and the absence of errors with respect to CPGS, drawing in part on the knowledge of several members who have been involved for many years with strategic command and control. Some observations follow. 1. Ensuring safety and security is a system problem. Errors can occur in any of many parts of the overall system for handling and employing weapons: for example, during production, handover to the military customer, transport to a military base, warehousing, selection of ordnance to be loaded on an opera- tional platform, handling within the platform, the order to use a particular type of weapon for a particular target, the choice of a weapon consistent with the command order, and the delivery of the weapon to the correct target. And, when weapons are returned to storage, decommissioned, and destroyed, additional opportunities arise for problems. The events that led to the recent incident involv- ing the unintended transport of nuclear weapons on a B-52, followed by many hours during which the missiles sat unattended on the airplane at Minot Air Force Base, North Dakota, and Barksdale Air Force Base, Louisiana, included the errors made by support personnel in mistaking nuclear weapons for unarmed missiles to be destroyed (despite distinctive markings) and the failure of a series of support personnel and flight-crew members to fully check what weapons were being hung on a B-52 for transport and what weapons were on an arriving B-52. Ultimately, the responsibility lies with those who designed the system and those who failed to monitor the detailed performance of the redundant system and the many partial failures that must have preceded this major lapse. 2. Software problems at the “front end” are a special system issue. Those working to reduce safety and security risks to a minimum are typically associated with the weapon system program, such as the CTM program, but the possibility exists—at least in principle—for errors to occur between the time of decision, its interpretation, translation, encoding, and transmission, and the time at which the message to execute is received on the weapon system platform. Especially in a fast-track system such as the CTM is intended to be, the possibility always exists for software errors that are not detected early. A program to ensure safety, then, must include the full end-to-end process, including aspects of command and control over which the program office has no control.

OCR for page 27
57 MILITARY ISSUES 3. Errors are often correlated rather than independent. As illustrated by the recent B-52 fiasco, error or complacency by one person or one group of people can be “passed on,” so that a sequence of intended checks turns out to be an initial failure to check followed by a sequence of pro forma nods. That is, what should have been sequential independent checks are instead correlated. In addition, of course, a system of checks and controls may fail if it has been poorly designed, so that the failure of a particular critical component means the failure of the whole. A familiar example is that of a building or area secured with formidable checkpoints with well-disciplined guards, but to which a construction crew is allowed easy entry by merely flashing a generic badge. In military systems, a parallel example might involve a critical command-and-control node. 4. Errors of procedure are more likely to occur when equipment needed for different purposes is co-mingled and when people have more than one mission. All else being equal, it is unwise for mechanics to have similarly shaped parts for dif- ferent devices on their immediate worktable; for soldiers to be carrying weapons and ammunition that would be prohibited for their current mission; or, more to the point, for conventional and nuclear weapons to be co-mingled in ways that would make it relatively easy for the wrong weapon to be used—whether accidentally, in anger, or as the result of conspiracy. 5. Security designs that include a combination of technical and procedural safeguards can be much more effective than either class separately. The recent B-52 event occurred because of a sequence of human failures. It might have been readily avoided if, for example, an electronic system refused to permit mounting of the missiles because of a test indicating that the missiles were nuclear-armed. An electronic system might have refused to permit nuclear-armed missiles from being removed from the warehouse without a special electronic key available only to logistics personnel recently certified for such activities. However, comput- ers and electronic systems are by no means foolproof and by no means do they always work properly. The classes of error that beset them, however, are different from those that can defeat processes dependent on consistent human adherence to procedures. Planning for the CTM The reason for including this discussion of safety is that the committee believes that those expressing worries about the safety and security of strategic weapons such those for CPGS are quite right to have concerns, but that solutions to the problem should come from careful system engineering rather than simple heuristics and impressions. That the B-52 incident represents an “unimaginable” lapse in the handling of nuclear weapons does not mean that all systems for han- dling, controlling, and using nuclear weapons (or strategic conventional weapons) are equally likely to have such mishaps. Different processes and protections are used for different weapon systems (e.g., B-2s, B-52s, ballistic missiles); some are

OCR for page 27
58 U.S. CONVENTIONAL PROMPT GLOBAL STRIKE better than others, and the relative goodness (reliability and effectiveness) can be assessed by independent reviewers with in-depth knowledge. Recognizing the potential for error or loss of control of security or safety, the committee was very impressed by the preventive measures contained in the Navy’s plan for the CTM. The committee was briefed in detail on the proposed missile and shipboard modifications and the operational procedures that would provide CTM command-and-control and nuclear surety for SSBNs deployed with both nuclear and conventional Trident missiles. Among these measures are separate safes for the firing keys of the nuclear-armed and the conventionally armed missiles (one safe would be held by the commanding officer, the other by the executive officer), incompatible electrical interfaces, two physical blocking elements in the firing circuit in the missile, software interlocks, physical examina- tions on loading, and unique missile designators. Many standard SSBN nuclear surety measures apply. Standard nuclear surety procedures in system development will also mitigate the risk of an accidental nuclear launch in a CTM engagement. Personnel from the Naval Surface Warfare Center, Dahlgren Division, Virginia, who write the Trident fire-control software are closely monitored in a Personnel Reliability Program. Dahlgren personnel independently verify missile software developed by Lockheed Martin and vice versa. OSD periodically conducts an extensive 3-week review for all system changes. CTM-equipped Trident submarines would not be the first nuclear-capable mixed-load platforms. In the 1980s, attack submarines carried nuclear-armed land-attack Tomahawk cruise missiles in addition to conventionally armed anti- ship Tomahawks. Earlier in the Cold War, smaller nuclear weapons for antisub- marine warfare (antisubmarine rockets and submarine rockets) and antiair warfare (one of the family of Standard Missiles) were carried together with conventional weapons on surface ships and attack submarines. Furthermore, SSBNs routinely operate with mixed loads during fleet commander’s evaluation tests, during which nuclear payloads are switched out with test payloads before beginning a strategic patrol, and these reconfigured missiles are fired during the patrol under operational conditions. In the committee’s opinion, the Navy’s Strategic Systems Programs (SSP) presentations to the committee address the nuclear surety issue well. It is clear that SSP recognizes the additional risks related to CTM and is exercising due dili- gence in its plans to mitigate those risks. SSP’s track record gives one confidence. However, owing to the magnitude of the consequence of an accidental launch and the small amount of time that commanders in the field might sometimes have in prompt global strike engagements, the committee recommends the following: in addition to the risk mitigation measures proposed by SSP, OSD should appoint a red team to review thoroughly SSP’s plans, searching specifically for ways in which, perhaps due to a series of concurrent failures, an accidental launch of a

OCR for page 27
5 MILITARY ISSUES nuclear weapon might occur in an intended CTM engagement. SSP has used aggressive red teams in the past with excellent results. With the above proviso, the committee concludes that the risk of an accidental launch of a nuclear weapon in a CTM engagement can be made very low, and it believes that the risk of an accidental launch should not by itself be a reason to decide against developing the CTM. FINDINGS AND RECOMMENDATIONS The committee reached the following conclusions: • Importance. The CPGS mission is important and worthy of near-term priority and action. Indeed, extremely negative public reaction would be under- standable in a few years if the DOD were unable to accomplish crucial conven- tional strikes (e.g., against terrorist leaders or terrorists moving weapons of mass destruction, or against a rogue state that continued to fire missiles at an allied capi- tal for a period of hours) because of U.S. failure to have provided the option. • Near-term options. The only credible near-term (2 to 3 year) option is the Conventional Trident Modification and (with an additional year or 2 delay) a variant proposed by the committee for consideration, referred to as CTM-2. The Minuteman option, which might have been expected to be a contender for the near term, turned out not to be as attractive as might have been expected (see Appendix I). • Longer-term options. For the mid-term (e.g., around 2015), at least one additional option is available, the SLGSM launched from SSGNs. Thereafter, land-based options (CSM-1 and CSM-2) are proposed, with boost-glide capability if high-risk R&D proves successful over the next few years, as well as hypersonic cruise-missile options. These options are best seen as potential long-term options worthy of near-term R&D, but with IOCs not to be expected before about 2020. • The CTM. The near-term CTM would be effective for an important class of missions, it would be marginal for others, and ineffective for others. Describing the CTM as a niche weapon is appropriate. The committee emphasizes, however, that a niche weapon may be highly valuable if it provides operational capability for addressing critical situations, which would be the case with respect to CTM for several of the scenarios considered. The technical means for ensuring the nuclear safety of a CTM system ensure that the risk of an accidental launch of a nuclear weapon in a CTM engagement can be made very low. For the reasons explained in Chapter 3, the committee believes that it is highly unlikely that a launch of a conventionally armed Trident would be misinterpreted as a nuclear attack; that the chances that such a launch would trigger an immediate nuclear response is very low and could be reduced even further by means of established cooperative measures; and that the “ambiguity” issue is not a reason to forgo the capability that CTM would afford.

OCR for page 27
60 U.S. CONVENTIONAL PROMPT GLOBAL STRIKE • An alternative. The CTM-2 could be an immediate follow-on to the CTM, or possibly an alternative, albeit one that would add to delay in IOC and would have to build heavily on the development and testing of the CTM, as would the SLGSM. The CTM-2 would have greater payload volume than the CTM and, being a two-stage missile, under some circumstances could be distinguishable from the three-stage CTM to observers of the CTM-2 launch and trajectory. • Enablers. A number of enabling activities are essential to mission success for CPGS. These activities, such as the precise determination of aimpoints, will sometimes have been done in advance and if so will not be a limiting factor, but they could also need to be done in a matter of an hour or so. Regardless of how quickly they must be accomplished, the quality of the enabling activities is critical, and the demands on those activities will grow over time if the ambition is to be able to strike moving targets. Major efforts are needed with respect to the enabling activities, not just the CPGS platform and weapon system. • Avoiding excessive constraints on R&D. A congressional prohibition on CPGS R&D building on the Trident missile would have serious adverse conse- quences for almost any CPGS option, causing delays of at least 4 or 5 years. If Congress should allow Trident-related R&D but not authorize the CTM’s deploy- ment, the CTM-2 or SLGSM availabilities would not be much affected—perhaps by a year or so. • Special issues. The committee recommends the following: (1) Congress should require that all CPGS options be developed with concepts of operations meeting very high, strategic, levels of safety and security—comparable to or better than the concept for the CTM, and planning for accomplishing this should include a review of command-and-control software, not just of weapon system issues, as well as rigorous red teaming to detect problems; (2) OSD should appoint a red team to thoroughly review the Navy’s plans, searching specifically for ways in which, perhaps due to a series of concurrent failures, an accidental launch of a nuclear weapon might occur in an intended CTM engagement; (3) options for increasing the number of warheads per CTM, giving the CTM increased capabil- ity against terrain-protected targets, and perhaps improving the lethality of the flechette-based weapon used on the CTM (and CTM-2), should be examined immediately, since such options are potentially important and may be feasible even in the initial deployment; and (4) Congress and the DOD should resolve the pure-load versus mixed-load issue on its merits, adjusting the various options accordingly so that they are measured by the same assumptions.