Purpose

This chapter describes highlights of recent analytical work assessing the prospective capabilities of various types of U.S. force structure (circa 2005-2015) to conduct and win major regional conflicts (MRCs). The work examines future versions of the force structure defined by the Bottom-Up Review (Aspin, 1993) and variants with a wide variety of postulated new features. Our emphasis is exploratory; that is, the analysis attempts to extract broad insights that can be used to construct alternative force postures for subsequent in-depth evaluation.[1]

This search for insights is appropriate because, since 1993, there have been many claims and counterclaims regarding U.S. capabilities in future theater conflicts. Our goals here are to clarify the issues, summarize insights, indicate where military problems are most and least serious with respect to MRCs, and suggest possible directions for program changes to mitigate the difficulties. Readers hoping to find a simple statement supporting or rejecting the claim that the United States can fight and win two concurrent MRCs will be disappointed because U.S. capabilities in this regard depend to a great degree on details of scenario, strategy, and even measures of effectiveness.

Approach

Our approach is as follows: First, we lay out a relatively generic campaign structure in terms of its major events and phases. Next, we discuss the analytic methodology, which varies with campaign phase. Then we discuss results from a very large set of simulations conducted in 1995 and 1996. After that, we provide reductionist explanations independent of simulation details. Finally, we review shortfalls in U.S. military capabilities that emerged from our analysis, and suggest corrective measures that those building the defense program might consider. Our emphasis is on "Achilles' heel" problems, which are more severe than any shortcomings of overall force structure, and on identifying high-leverage opportunities. We end with summary conclusions about how to discuss the subject of "MRC capability."

A CAMPAIGN STRUCTURE FOR ANALYSIS

Campaigns, Phases, and Operations

One of the difficulties in characterizing military capabilities is that war comes in many forms. Advocates of one or another point of view can talk past each other by having different implicit conceptions of what a future war would be like. To avoid this difficulty, it is useful to be more explicit by adopting a generic campaign structure.

This chapter considers the class of MRCs that can be characterized as in Figure 6.1 (see also Frostic and Bowie, 1994). As this figure suggests, a given MRC can be characterized by charting key events along a time line, i.e., when there is strong strategic warning of attack (W); when preliminary preparatory and deployment measures begin (e.g., movement to the crisis region of carrier battle groups [CVBGs] and perhaps maritime prepositioning ships)(C₁); when full-scale U.S. deployment commences (C); when combat begins (D); when the enemy's attack is halted (T_h); when the counteroffensive begins (T_co) and ends (T_v); and, finally, when the entire contingency is over and forces can come home (E). The figure also shows that a second MRC would have its own time line, with the D-Days of the two separated in time by days or weeks. There is no necessary relationship between events along the two time lines. For example, the halt phase in Theater A might take twice as long as the halt phase in Theater B.

Figure 6.1--Schematic Time Line for Two MRCs

The Classic Problem: Halting an Invading Army

Within this depiction, the central problem is usually considered to be halting the invading army quickly enough--before unacceptable losses of life and territory occur. This is not particularly abstract when one considers specific theaters. In the old days of the Cold War, NATO was resolute in planning to halt any invasion of West Germany near the border. Similarly, combined forces in South Korea plan to stop any invasion within a short distance of the border to avoid occupation of Seoul. Other theaters have more depth, but how much depends on perspective and strategy. If the objective in Southwest Asia is to defend Kuwait, one does not want to allow Kuwait City to be overrun. By contrast, if the objective is to defend Saudi Arabia, Kuwait can be considered "depth." Prior to the 1990 invasion of Kuwait, the latter interpretation was common among military planners because there was no commitment to defend Kuwait and, it was believed, little chance of success. The United States now maintains some forces in Kuwait much of the time and has prepositioned equipment for the use of rapidly deployable ground forces.

The "halt phase" of a campaign might last days or weeks. In the usual image, the enemy is halted by joint and combined operations that include ground forces and air forces in particular. There has been relatively little discussion of the naval campaign in the public literature, but the Navy might play a major role in securing sea lines of communication, establishing air and missile defenses, preventing the insertion by sea of special operations forces, and threatening the enemy's flanks and rear with Marines in "operational maneuver from the sea." If the Navy had a CVBG or two in the region by D-Day and especially if other deployments had been delayed because regional states denied access or warning was too ambiguous to justify a full deployment early, Navy operations would be critical in the early stages of the campaign.

Other Phases

Assuming success in a halt phase, the next mission--perhaps preceded by what some call a prolonged "build-up and pound phase"--is to restore the border and, quite possibly, to continue as necessary to defeat the aggressor's army thoroughly, even if that means going into his homeland. Finally, assuming victory in the overall war, one must anticipate a consolidation and stabilization phase that would include occupation.

EXPLORATORY ANALYSIS FOR MRCs

Dimensions of Uncertainty

Given this generic structure of two nearly simultaneous MRCs, the next task in analysis is setting up the conditions for simulated war. The results of simulated war depend on a large number of factors, all of them highly uncertain and resistant to anything like a best-estimate characterization. For example, how can one meaningfully establish a "best estimate" of the time between C-Day and D-Day for an abstract war 5 to 15 years in the future?

Even more troublesome is the delay between "warning" and deployment of U.S. forces. Warning is usually ambiguous, and there can be high political and economic costs associated with reacting, much less overreacting, to warning. After all, the enemy need not proceed with the attack, in which case the reaction can be seen as having been alarmist and disruptive.

These, however, are only a few of the many uncertainties. In our approach, we categorize uncertainties as falling into six categories:[2]

• Political-military context (e.g., timelines, allies, access rights)

• Strategies (the enemy's attack strategy and the friendly side's defensive strategy, which are functions of military objectives)

• Forces (size and general character of all relevant forces)

• Capabilities of forces and weapons (not only the nominal effectiveness of the various divisions, wings, and battle groups, down to the level of particular weapons, but also their real-world in-war capabilities, which might prove quite different)

• Environmental factors (e.g., weather and the particular terrain on which battle is joined)

• Algorithms and related parameters for describing military operations (e.g., the equations and parameter values used to estimate attrition rates and movement rates).

Context of Analysis

As computer power has improved, we and other RAND colleagues have sought to confront massive uncertainty directly by examining a much broader range of the potential scenario space than has heretofore been feasible. Nonetheless, many assumptions are necessary to contain the scope of analysis. The work described here takes the perspective of force-planning studies. That is, it focuses on relatively "macro" factors, such as force size, time lines, strategies, and the presence of high-leverage new capabilities--rather than other, more specific, factors that would be important to a particular commander worried about war tomorrow in his theater.

Experimental Design

Our experimental designs for analysis varied a great deal with the future adversaries and theaters considered, but Table 6.1 illustrates variables addressed, especially in work concerned with Southwest Asia and South Korea. We also considered other theaters, both specific and generic. However, considering a wide variety of operational circumstances for the "standard" name-level scenarios of Iraq versus Kuwait and North Korea versus South Korea provides a rich set of tests with which to assess U.S. MRC capabilities.

^aAn important feature of the modeling was rule-based "adaptive strategies" to meet the needs encountered in the particular run. For example, the U.S. forces associated with the POMCUS brigade in Kuwait might marry up with its equipment and fight in northern Kuwait, marry up with the equipment and fall back to fight later, or not be deployed at all (sacrificing the equipment), depending on how the war develops.
^bChemical effects were reflected largely through airfield and port availability and sortie rates.
^cThese were reflected by requiring initial deployments of infantry forces and counter-mine forces, thereby delaying deployment of main forces for halting the invasion.
^dA given combat unit may be much less effective than feasible given its equipment and manning, as the result of such diverse issues as training, doctrine, morale, poor leadership, and commitment to the war.
^eThese were inputs for ideal circumstances (e.g., an F-15E attacking a moving road formation). Simulation outputs for kills per sortie were typically less, by as much as a factor of two, because of variations in the type of battle, corrections for redundant attacks on the same vehicles, and other factors.
^fSortie-rate suppression could be due to many factors, ranging from chemical attacks on forward air fields to greater-than-expected surface-to-air-missile threats delaying suppression of air defenses and reducing the rate at which air-to-ground sorties could be safely flown.
^gBad weather was manifested through reductions on both kills per sortie and sorties per day.
^hIn baseline calculations, the maximum speed of an armored force is limited when it is under air attack, even if attrition has not yet reached high levels. This simulates indirect effects of confusion and attacks on support systems. However, this slowing effect is controversial within the modeling community, so we ran cases that turned the effect off, allowing the units to "keep on trucking" until substantially destroyed.

We considered nominal force structures for a period between 2000 and 2015. In addition, in one or both of the two studies underlying this chapter, we considered cases representing a wide variety of additional capabilities, as suggested in Table 6.2. In some cases, the extra capabilities were represented explicitly (e.g., additional B-2s or C-17s). In other cases, they were represented indirectly through changes in parameter values (e.g., increasing crew ratios would increase the sortie rates of aircraft). In still other cases, they were represented by a combination of force-employment strategy and parameter values. For example, one of the most important payoffs from a moderately high level of information dominance was assumed to be the ability to recognize quickly where the attacker's main thrusts were and to inform our ally's ground forces quickly so that both those ground forces and U.S. air forces and long-range fires could be appropriately counterconcentrated. We considered this to be a relatively straightforward and feasible way to use technology as leverage in improving the effectiveness of allied forces on the ground.

Although we examined a wide range of potential capability enhancements, many of them were motivated by what we considered to be generic challenges: (1) enhancing and complementing capabilities of allies; (2) getting forces to the theater quickly; (3) dealing with access problems, including those due to mines; (4) establishing secure lodgments quickly; (5) suppressing air defenses quickly to allow high-intensity interdiction attacks; (6) minimizing casualties; (7) defeating--not merely attempting to deter--WMD (especially chemical weapons) delivered by missiles; (8) fighting in difficult terrain; and (9) conducting counteroffensives. Our initial hypothesis was that all of these were nontrivial challenges. Subsequent analysis strongly confirmed this.

Varying Objectives, Strategies, and Measures of Effectiveness

Table 6.3 indicates some of the many changes in objectives, strategies, and measures of effectiveness considered. This reinforces the earlier point made that these variables matter greatly in assessing capabilities and the relative merit of different improvement measures. One example is that a constraint to minimize attrition to manned air forces increases the perceived value of stealth systems, missiles, and, in the longer run, armed unmanned aerial vehicles (UAVs).

A related issue in setting up the experimental design is recognizing that there are at least three distinct purposes to consider when examining alternative investments:

• Doing the same warfighting job using fewer of the available forces, thereby increasing strategic flexibility, reducing casualties (e.g., by having fewer manned sorties, but achieving more kills per sortie), and perhaps reducing costs

• Doing the same warfighting job nominally, but increasing the certainty of success--i.e., hedging against the effects of unex-pectedly severe air defenses with new tactics, poorer-than-expected performance from allied forces, or operational constraints

• Improving warfighting outcomes (e.g., shortening the war and decreasing casualties).

Figure 6.2--Illustrative Invasion Case for Southwest Asia

Results of Exploratory Analysis

The kinds of experimental design suggested in Tables 6.1 and 6.2 involve tens or hundreds of thousands of cases, even with shortcuts.[3] Much of the analysis is performed by "flying through the outcome space" after completing the computer simulations and storing the results (i.e., by moving slider bars on a computer display that varies the factors in Tables 6.1 and 6.2). What we show here are merely "slices" through that outcome space.[4]

Figure 6.3 shows results for baseline future forces (the result of a decade's buildup consistent with current programs) as a stoplight chart in which open squares represent favorable results (indicating a successful defense), black squares indicate very unfavorable results, and other shadings are intermediate. In normal work the display has colors ranging from red to green.

Figure 6.3--Illustrative Results for Defense of Saudi Arabia (circa 2010)
(shading of each cell indicates quality of one simulation's outcome)

Merely to illustrate the kinds of operational assumptions that underlie such simulations, let us quote from a portion of our informal documentation:

Our initial analyses assumed the goal of halting the advance as far north as possible. Kuwaiti armored/mechanized units, roughly the equivalent of two Saudi heavy brigades, defended in northern Kuwait, between Al Jahra and the Iraqi border. Of the five Saudi armor/mech brigades assumed available in the eastern region, four moved to defend along the route from Wariah to the Kuwaiti border, while the remaining brigade defended the Dhahran area. In addition, infantry units defended Kuwait City, Wariah, and Dhahran.

U.S. forces joined this defense as far forward as circumstances permitted (the war plan used in the simulation was adaptive). The amount of warning time was a key factor. If U.S. mobilization and deployment began early enough, the United States deployed forces to man the POMCUS heavy brigade set in southern Kuwait, which then joined the Kuwaiti defenders. However, if the U.S. deployment was late, offloading here was likely to be too dangerous, due to the threat of Iraqi ground advances, or possibly air attacks. In this case, the Kuwait POMCUS set was abandoned. As explained above, the hypothetical U.S. "Sea Cavalry" included a carrier-based attack helicopter regiment (e.g., 54 AH-64 helicopters or a comparable component of Comanches). These units, if available, deployed to wherever they could do the most damage to the Iraqi advance.

Other U.S. early entry ground forces initially deployed to the Dhahran region and advanced from there to join the GCC defending forces. This always included two heavy brigades: one to man a POMCUS set in Qatar and another in the northern United Arab Emirates. A U.S. Army Pre-positioning Afloat (APA) heavy brigade, if available for this contingency (which it was except in cases in which we assumed a prior contingency elsewhere), also deployed to Dhahran. These forces were supplemented by airlifted U.S. light forces, which either supported the advance against the Iraqis or, in the case of opposition elements in Saudi Arabia, supported Saudi government forces and secured air bases, ports, and other key facilities.

While the baseline results (Figure 6.3) were favorable, the results were much less good if the objective was to halt the imagined Iraqi advance close to the Iraq-Kuwait border--i.e., to prevent the overrun of Kuwait itself. Figure 6.4 indicates that the Iraqi advance could be stopped within roughly 60 km of the Iraq-Kuwait border if the United States had a week before D-Day to deploy. For a deployment time of about four days, results vary with the deployment rate and effectiveness assumed for air forces (both Air Force and Navy).[7]

Figure 6.4--Prospects for Defense of Kuwait with Baseline Forces (circa 2010)

Figure 6.5--Defense of Saudi Arabia (circa 2010), Considering Effects of Sortie Suppression

Figure 6.6--Defense of Saudi Arabia (circa 2010), with Entry and Access Problems

Again, then, we conducted an extensive set of analyses examining a broad range of assumptions. These included not only assumptions within a given theater but also assumptions about the time between the outbreak of conflict in two theaters, which war started first, how airlift was allocated among types of forces and between the two theaters, and so on. Some of the details are classified, but the broad conclusions can be understood in simple terms. And while the underlying simulation is sophisticated in many respects, the particular results of interest in the current analysis depended primarily on some very simple assumptions. Although intended as a screening analysis to identify key factors and hypotheses for subsequent study, the exploratory analysis yielded insights (not precise numbers) that we believe have a relatively firm foundation.

Understanding the Results in Simple Terms

Good Cases. In "good cases," which correspond to the enemy pursuing a straightforward armored invasion and the United States having enough deployment time to put air forces into the region and to deploy Army forces to their prepositioned equipment before it is overrun, the "halt" problem consists of destroying enough armored vehicles so that, even in a conservative analysis, one could be confident that the enemy forces would be stopped. Suppose that there are approximately 1,000 armored vehicles in an "equivalent division" (ED), which is perhaps 30 percent more than a current Iraqi division. Suppose further that there are roughly eight such EDs in the main attack. Next, suppose that the attack will stop if half of the armored vehicles are destroyed. If each sortie killed an average of one vehicle, then a total of 4,000 sorties would be required. Assuming three sorties a day for the critical period (higher than sustainable), that would correspond to 1,333 aircraft-days, counting only air-to-ground aircraft with high capability. If there were an average of 200 such aircraft in the theater during the first week of war (mostly Air Force and Navy fixed-wing aircraft, but with some attack helicopters), which is a function of when deployment began, of course, then the invasion should be halted in a week. This ignores the attrition due to ground combat of maneuver units or to long-range missiles, such as ATACMS, that might be launched from Army units or Navy ships.

Mechanized army units can move at high speeds for short periods (e.g., 30 km/hour). However, they cannot move rapidly over long distances, especially when in the formations required under combat conditions. Tanks must be refueled; traffic jams occur; units stop and must be cajoled or threatened into continuing; command and control is usually confused; and even moderate resistance on the ground can compel the invader to proceed cautiously with relatively concentrated forces. That concentration increases the effectiveness of air forces. Historically, average movement rates of 20 km/day have been typical of successful offensives. Movement rates of 80 km/day or so are usually assumed impossible except in the absence of opposition (e.g., as when U.S. forces swept into Iraq in the famous left hook). Returning to the hypothetical example and assuming a movement rate of 20 km/day, the halt would occur within 140 km of the border. Roughly speaking, then, this explains the kinds of results seen in the above figures.

Now, the numbers in the above paragraph were purely illustrative. Other assumptions are equally plausible: higher or lower kills per sortie or daily sortie rates, smaller or larger Iraqi forces, or a smaller or larger number of aircraft available in the theater during the first week. The notion that half the armored force must be destroyed before the invasion halts is also open to debate, especially for attacking forces without intense motivation. Thus, the precise halt time can vary drastically with the specifics. Giving only "best estimate" results is quite misleading, because they are not obviously any better than substantially different ones. The overarching conclusion, however, is that air and missile power, in enough quantity, should be able to halt a classic invasion. Many observers have argued with this conclusion. But it seems valid. As a point of reference, we quote a renowned German army general of World War II, when aircraft were not nearly so lethal:

A large-scale offensive by massed armor has no chance of success against an enemy who enjoys supreme command of the air (von Mellenthin, 1955, Ch. 22).

Figure 6.7--Penetration Versus Weight of Airpower (higher movement rates are possible if ground forces fail to slow invader)

Bad Cases. It is similarly easy to understand why results would be much worse in other, "bad" scenarios. Some of the worst involve WMD, discussed below, but there are bad cases even without WMD. Suppose that resistance at the key air bases and seaports required the U.S. commander to force entry with specialized Army and Marine infantry. As noted earlier, this might not even be possible without weeks of preparation, but even if it were, it would require dedicating airlift to the deployment of those forces, which would be of little use in halting the armored invasion. Thus, both ground forces and tactical air forces would arrive later, during which time the Iraqi advance might continue. Understanding this does not, as the clich goes, require rocket science. For similar reasons, temporary closure of the Strait of Hormuz at the outset of such a conflict could be quite troublesome. So too would be political denial of base access.

Multitheater Issues. In two-MRC cases with a short period separating the conflicts, airlift would be divided between theaters. Even though Air Force tactical aircraft can have a profound effect on an invading armored force and can self-deploy, they cannot operate without a substantial support structure. Some of this may be prepositioned, but a good deal of support equipment and personnel must be airlifted in parallel with the deployment of the aircraft themselves. Thus, in the event of a two-theater war, if strategic-mobility assets had to be divided between the two theaters, everything would slow down--not just the deployment of ground forces, but even the deployment of air forces. The principal exception here would be Naval forces in or near the region by D-Day. Although Navy aircraft would be supported directly from CVBGs, these aircraft would probably not be numerous enough to carry the burden.[11]

Again, then, it is straightforward to understand approximately what is going on in the much more complicated computer simulations as one considers a variety of two-theater cases. Roughly speaking, if the period between the outbreak of war in the two theaters is at least three weeks (which allows the United States and its allies to halt the invader and gain the initiative in the first theater), then--in some cases--both wars can be serviced almost as well as if there were only one war.

To be sure, this result does not hold up if we assume that details of today's force structure are extrapolated into the future. Today (in 1997), there are many practical problems that would make simultaneous conflicts very difficult. In particular, unless mobilization were prompt, there would be severe shortages in critical support units because so many supporting forces and assets are in the reserve component. Also, it is questionable whether our decision-support and logistics systems would be adaptive enough to deal well with the situation (see discussion of the former in Kassing, 1992). It is not surprising that so many senior officers are skeptical today about two-MRC capability: They should be. These problems, however, are more a matter of management than the result of overall resource or force-structure limitations. In this information era, the U.S. military should be able to greatly reduce such frictions in the system, especially if civilian authorities and Congress permit necessary changes in infrastructure, logistics, and tailoring of reserves.

Very Bad Cases. So far, we have only discussed problems associated with time line and the halt phase. There are many cases to be contemplated that could be even more problematic. Suppose, for example, that the U.S. response to ambiguous warning were slow enough so that the invasion was successful before U.S. forces could arrive in numbers (e.g., C-Day = D-Day + 6, to mention a case with recent precedent). In that case, the challenge would include a long-distance movement followed by a counteroffensive against dug-in enemy forces, which in some theaters might be distributed through urban sprawl rather than strung out in the desert. The enemy might deliberately intermingle with friendly civilians to deter preparing for the counteroffensive with massive bombing. In such a theater as Korea, the battle might be infantry intensive in rugged, mountainous terrain. Add to these complications the possibility that an enemy might use chemical, biological, or nuclear weapons against U.S. and allied forces or allied population centers. In all such cases, the United States might have its hands full to deal with even one MRC. And that MRC might take many months or, conceivably, years. Although we have done some simulations of "very bad cases," the most important points can be understood without bothering.

CONCLUSIONS FROM INITIAL EXPLORATORY ANALYSIS

With this overview as background, let us now summarize some of the more important insights derived from the simulation-based exploratory analysis. The observations that follow assume that U.S. forces are equipped with excellent command, control, communications, computers, intelligence, surveillance, and reconnaissance (C⁴/ISR) assets, capabilities that will be a cornerstone of future U.S. forces (see, e.g., Johnson and Libicki, 1996; Shalikashvili, 1996).

Theater-by-Theater Observations

Southwest Asia. Because of the substantial prepositioning programs under way,

• Deterring or defeating a future invasion of Kuwait and Saudi Arabia by Iraq will probably be quite feasible in "standard" scenarios, which assume unopposed deployment operations and good use of warning (i.e., "leaning forward" with preliminary deployments and other preparations, well before "C Day").

• In other cases results could be much worse, especially for defense of Kuwait.

• Potential corrective measures include expanding stocks of high-quality anti-armor air-delivered munitions in theater, prepositioning or forward deploying ground- or sea-based long-range-fire systems, forward deploying attack helicopters on land or on carriers (a possible focus for Marines), or use of other forces, such as an enhanced force of long-range bombers with advanced weapons and avionics for prompt attacks on the advancing Iraqi army.[12]

• South Korea should be capable of defending itself against a North Korean attack, although an attack with weapons of mass destruction might cause a great deal of chaos and trouble, and Seoul is extremely exposed. U.S. air forces provide a powerful hedge against various potential ground-force disasters, although weather can be a mitigating factor. Procuring and prepositioning sizable stocks of anti-armor and anti-personnel munitions that are effective in poor weather is warranted.

• A unified Korea should be able to mount a substantial deterrent against China, at least insofar as preventing a quick and easy invasion is concerned. This assumes that Korea would have a modern army, including long-range-fire systems comparable to the Multiple Launch Rocket System, (MLRS), the Army Tactical Missile System (ATACMS), and perhaps the Brilliant Anti-Tank (BAT) munition. Also, it assumes Korea would not insist on a forward defense of its border with China, but would instead exploit some of the mountainous terrain and depth. U.S. air forces would provide a substantial hedge. Thus, not only is there no current reason to see China as a threat to Korea, so also is there reason to believe that the military balance will be adequate.

• Extending NATO's security guarantees to Poland should not create serious challenges--so long as Poland maintains a substantial military capability herself (a critical assumption with implications for negotiations during the NATO enlargement discussions). A NATO force of 5-10 projectable divisions, plus 8-10 effective Polish divisions, plus ten wings of tactical air forces, plus a contingent of MLRS, ATACMS, and BAT would be sufficient to defeat attack under a wide range of circumstances.

Turning now to more generic conclusions, we should start by observing that the capabilities of U.S. general-purpose forces are very substantial. In straightforward engagements, U.S. forces will simply outclass any regional aggressor on the landscape. Indeed, our analysis of programmed future forces suggests that, with sufficient warning time and reasonably effective allies (big ifs, to be sure), the United States should be able to defeat a classic armored invasion handily. In many cases, the United States should be able to handle two such invasions concurrently--so long as they are some weeks apart or, in some cases, even if they are more nearly simultaneous.[13] Thus, classic armored invasions affecting U.S. interests should be or should soon become obsolete. The United States will need to prepare for them indefinitely to keep them obsolete. But they may not occur again, at least not in pure form as in the past and in our current plans.

Our adversaries, of course, can recognize all this as well. Thus, we must expect them to avoid classic armored invasions and instead adopt strategies involving fast, no-warning invasions with armored forces and/or various "asymmetric" tactics exploiting U.S. weaknesses.[14] Current U.S. Achilles' heels involve limited capability to

• Assure having some high-lethality forces in the theater on or very shortly after D-Day, even in short-warning, late-reaction cases

• Prevent (as opposed to simply deter) the use of WMD, especially if delivered by ballistic missiles or covert means

• Seize and secure ports and bases rapidly that are not adequately protected by allied forces

• Clear sea lanes of mines quickly

• Employ air-to-ground munitions against invading armies in difficult terrain (including urban sprawl) or bad weather[15]

• Conduct counteroffensives in such terrain

• Halt large dispersed threats (e.g., a broad-front infantry invasion in broken or wooded terrain).

• Using multilayer ballistic missile defenses

• Leveraging allied capabilities with C⁴/ISR

• Exploiting the potential lethality of long-range bombers

• Planning rapid deployment or forward deployment of long-range precision fires, possibly including tactical air forces with standoff weapons, ship-based missiles and advanced guns, and mobile missile ground-force units, such as MLRS, with extended-range ATACMS and the BAT munition.[16]

Dependence on Precision Strike and Information Dominance

An important caveat should attend all of this discussion. At the heart of most optimistic assessments of U.S. capabilities, including this one, is the assumption that long-range precision-strike systems, whether aircraft or missiles, will prove to be effective. It is possible, however, that in some future contingencies the United States will find its high-tech systems being an order of magnitude less effective than the "potential." There are many reasons, including the potential for countermeasures of many types, as suggested in Table 6.4.

Figure 6.8--First-Cut Estimates of Countermeasure Effects (relative to program-office values, not baseline of this study)

Other factors that have major effects on what types and numbers of forces are needed and on assessments of success are assumptions about objectives and constraints, notably the following:

• Notions of what constitutes "winning" (halt, restore border, destroying X percent of the enemy's army, invade and occupy, etc.)

• Attitudes about attrition (not only to U.S. forces, but also to friendly forces and civilians, and even enemy forces and civilians).

We can only speculate about what constraints would apply in a future MRC. Despite the currently widespread view that the United States will not accept casualties, empirical and logical analysis suggests that, if the stakes are high enough, high casualties would be tolerated (Larson, 1996a, 1996b). Going into Desert Storm, many U.S. decisionmakers were anticipating 1,000 to 10,000 casualties. On the other hand, sharply limiting casualties is likely to be a major objective in other cases where the stakes are less dramatic. Also, the low attrition in Desert Storm seems to have established optimistic expectations about war, which may constrain future presidents and commanders.

The desire to limit U.S. casualties tends to favor the use of long-range fires, whether from Air Force and Navy aircraft or from ground forces and ships. Survivable systems and related concepts of operation have a high premium.

Also, concerns about killing civilians (both friendly and enemy) argue for increased nonlethal and nondestructive capabilities, where feasible. Both "point" and wide-area capabilities are needed here.

Weapons of Mass Destruction

As mentioned above, WMD represent a major challenge, and the United States clearly needs the ability to neutralize such weapons early in a conflict. The WMD problem is critical because the threat to use WMD could deter our intervention, deter threatened states from asking for assistance, or deter allies from cooperating. And, of course, WMD could cause major casualties. This would probably precipitate an extraordinarily destructive U.S. retaliation, but we cannot rely upon deterrence alone--especially if the adversaries are desperate, as the North Koreans might be in invading the South, or as any adversary might be once the United States and its allies had begun a counteroffensive. Thus, theater missile defenses and counterforce capabilities are high on the priority list. But even with improved defenses, the WMD challenge tends to weigh in favor of a force posture and war plans that do not depend on dispatching large, densely packed U.S. forces into range of enemy WMD.

Important Improvement Options

For the middle term, our studies indicated substantial value to a wide range of special capabilities (these seem more important than more or less force structure on the margin). These involved missile defenses; airlift; long-range strike; enhanced allied capabilities (primarily ground forces); low-observable aircraft; battle management and C⁴/ISR, and--important across the board--leveraging existing platforms (ships and aircraft) with advanced weapons; and increased sortie rates for aircraft.

In the longer term, we concluded that UAVs have high potential for defense suppression, for improved C⁴/ISR, and even for direct attacks on ground forces (see also DSB, 1996). Anti-missile defensive systems, including directed-energy weapons, such as airborne lasers, have a high potential payoff. Nonlethal area munitions not requiring precise location data also have an important role, in both MRCs and lesser conflicts.[19]

In general, the most promising approaches to dealing with the challenges and shortfalls we have identified have little or nothing to do with adding more force structure. Rather, the shortfalls are best addressed by altering the posture of the force (e.g., through forward deployment) or by reorganizing and re-equipping forces with new capabilities.

Examples here include rapidly deployable and perhaps partially prepositioned MLRS, ATACMS/BAT units or something equivalent, forward-deployed arsenal ships, forward-deployed small-deck carriers with attack helicopters, and B-2s with enhanced capability for attacking armored columns.

The Cost Dimension

Obviously, it is much easier to identify desirable additional capabilities than to propose ways of funding those additions. Although we do not discuss such matters here, we are heavily engaged in efforts to identify priorities both for investment and for protecting programs in the event of cuts or general cost growth. Some of the methods being used are discussed elsewhere in the context of a different application (Hillestad et al., 1996a, b).[20]

Sizing the Force and the Two-MRC Issue

As more fully discussed in Chapter Five, the force-sizing debate currently revolves around the controversial two-MRC criterion. This focus is misdirected. Our analysis shows that the United States has more than ample capability for two MRCs in favorable or only moderately degraded situations. But it would find itself stressed in even one MRC in worst-case situations (e.g., a large-scale counteroffensive after a North Korean surprise attack with chemical weapons had shattered South Korea's defenses). Hence, the results of fighting two MRCs would depend on the nature of the challenge imposed by our adversaries and other variables. In favorable cases, we might be able to win both conflicts quickly and decisively. In others, we might have to hold and punish the aggressor in the second theater until, having defeated the first aggressor, we could redeploy certain forces to the second. Or we might be able to depend more on allies in one of the MRCs. In still other cases, we might find that fighting even one war would require months or years. It all depends. Thus, the two-MRC criterion, as currently defined, applies to only a small portion of the potential scenario space and is not, in and of itself, a sound basis for planning.

Nonetheless, if we must have a number, DoD has it right: "Two" is the right one. It would be folly for the United States to announce a one-MRC (or even 1-1/2 MRC) criterion, because such a strategy would give us pause before acting in crisis and would encourage aggressors to exploit the opportunity of our being engaged elsewhere. Our own view is that force sizing should be based on multiple criteria, including environment shaping. We believe the more important issues relate to modernization, reengineering the forces, and strategy.

BIBLIOGRAPHY

Air Force Scientific Advisory Board, New World Vistas: Air and Space Power for the 21st Century, multiple volumes, Washington, D.C., 1996.

Aspin, Les, Report on the Bottom-Up Review, Washington, D.C.: Department of Defense, October 1993.

Barnett, Jeffrey, Future War: An Assessment of Aerospace Campaigns in 2010, Maxwell Air Force Base, Ala.: Air University Press, 1996.

Bennett, Bruce, Two Alternative Views of War in Korea: the North and South Korean Revolutions in Military Affairs, Santa Monica, Calif.: RAND, MR-613-NA, 1995. For Official Use Only.

Bennett, Bruce, Sam Gardiner, and Daniel Fox, "Not Merely Planning for the Last War," in Davis (1994), pp. 477-514.

Biddle, Stephen, "Victory Misunderstood: What the Gulf War Tells Us About the Future of Conflict," International Security, Vol. 21, No. 2, 1996.

Birkler, John, Myron Hura, David A. Shlapak, David Frelinger, Gary McLeod, Glenn A. Kent, John Matsumura, James R. Chiesa, and Bruce Davis, A Framework for Precision Conventional Strike in Post-Cold War Military Strategy, Santa Monica, Calif.: RAND, MR-743-CRMAF, 1996.

Birkler, John L., David Perin, Christopher Bowie, David Shlapak, James A. Winnefeld, James R. Chiesa, Considerations in Balancing U.S. Land- and Sea-Based Air Forces, Santa Monica, Calif: RAND, MR-632-CRMAF, forthcoming.

Bonder, Seth, "Impact of the New Global Environment on U.S. National Security Planning: Challenges to the OR Community," International Transactions of Operational Research, Vol. 1, No. 1, 31-39, 1994.

Bowie, Christopher, Fred Frostic, Kevin Lewis, John Lund, David Ochmanek, and Philip Propper, The New Calculus: Analyzing Airpower's Changing Role in Joint Theater Campaigns, Santa Monica, Calif.: RAND, MR-149-AF, 1993.

Buchan, Glenn, "The Use of Bombers in a Changing World: A Classical Exercise in Systems Analysis," in Davis (1994), pp. 393-450.

Davis, Paul K., ed., New Challenges in Defense Planning: Rethinking How Much Is Enough, Santa Monica, Calif.: RAND, MR-400-RC, 1994a.

Davis, Paul K., "Institutionalizing Planning for Adaptiveness," in Davis (1994), pp. 73-100.

Davis, Paul, David Gompert, and Richard Kugler, Adaptiveness in National Defense: The Basis of a New Framework, Santa Monica, Calif.: RAND, IP-155, 1996.

Defense Science Board 1996 Summer Study Task Force, Tactics and Technology for the 21st Century: Vol. 1, Final Report, Washington, D.C.: Department of Defense, October 1996.

Defense Science Board 1996 Summer Study Task Force, Tactics and Technology for the 21st Century: Vol. 2, Supporting Analysis, Washington, D.C.: Department of Defense, forthcoming.

Frostic, Fred, Air Campaign Against the Iraqi Army in the Kuwaiti Theater of Operations, Santa Monica, Calif.: RAND, MR-357-AF, 1994.

Frostic, Fred and Christopher Bowie, "Conventional Campaign Analysis of Major Regional Conflicts," in Davis (1994), pp. 351-392.

General Accounting Office, Bottom-Up Review: Analysis of Key DoD Assumptions, Washington, D.C.: General Accounting Office, GAO/NSIAD-95-56, January 31, 1995.

Hillestad, Richard, Warren Walker, Manuel Carrillo, Joseph Bolten, Patricia Twaalfhoven, and Odette A.W.T. van de Riet, FORWARD--Freight Options for Road, Water, and Rail for the Dutch: Final Report, Santa Monica, Calif.: RAND, MR-736-EAC, 1996a.

Hillestad, Richard, Warren Walker, Manuel Carrillo, Joseph Bolten, Patricia Twaalfhoven, and Odette A.W.T. van de Riet, FORWARD--Freight Options for Road, Water, and Rail for the Dutch: Executive Summary, Santa Monica, Calif.: RAND, MR-739-EAC, 1996b.

Horrigan, Timothy, various publications, Chicago: Horrigan Analytics, including Configurations and the Effectiveness of Air Defense Systems in Simplified Idealized Combat Situations, 1995.

Huber, Reiner, and Paul K. Davis, "Lastenteilung und Streitkräfteumfang der Atlantischen Allianz," Europäische Sicherheit, November 1996.

Johnson, Stuart, and Martin Libicki (eds.), Dominant Battlefield Knowledge, National Defense University, Washington, D.C.: Institute for National Strategic Studies, 1996.

Kassing, David, Getting U.S. Military Power to the Desert: An Annotated Briefing, Santa Monica, Calif.: RAND, N-3508-AF/A/OSD, 1992.

Kaufmann, William W., Assessing the Base Force: How Much Is Too Much, Washington, D.C.: The Brookings Institution, 1992.

Keaney, Thomas A., and Eliot A. Cohen, Gulf War Air Power Survey: Summary Report, U.S. Air Force, Washington, D.C.: U.S. Government Printing Office, 1993.

Larson, Eric, Casualties and Consensus: The Historical Role of Casualties in Domestic Support for U.S. Military Operations, Santa Monica, Calif.: RAND, MR-726-RC, 1996a.

______, Ends and Means in the Democratic Conversation: Understanding the Role of Casualties in Support for U.S. Military Operations, dissertation, Santa Monica, Calif.: RAND Graduate School, RGSD-124, 1996b.

Matsumura, John, Randall Steeb, Tom Herbert, Mark Lees, Scot Eisenhard, Angela Stich, Analytic Support to the Defense Science Board: Tactics and Technology for 21st Century Military Superiority, Santa Monica, Calif.: RAND, DB-198-A, forthcoming.

National Defense Research Institute, Assessing the Structure and Mix of Future Active and Reserve Forces: Interim Report to the Secretary of Defense, Santa Monica, Calif.: RAND, N-3520-OSD, 1992.

National Defense University, Strategic Assessment 1997: Flashpoints and Force Structure, Washington, D.C., forthcoming.

O'Hanlon, Michael E., Defense Planning for the Late 1990s: Beyond the Desert Storm Framework, Washington, D.C.: The Brookings Institution, 1995.

Pape, Robert A., Bombing to Win: Air Power and Coercion in War, Ithaca, N.Y.: Cornell University Press, 1996.

Shalikashvili, John (GEN), Joint Vision 2010, Joint Chiefs of Staff, 1996.

Steeb, Randall, John Matsumura, Terrell Covington, Thomas Herbert, Scot Eisenhard, Laura Melody, Rapid Force Projection Technologies: A Quick-Look Analysis of Advanced Light Indirect Fire Systems, Santa Monica, Calif.: RAND, DB-169-A/OSD, 1996.

von Mellenthin, F. W. (Maj. Gen.), Panzer Battles: A Study of the Employment of Armour in the Second World War, London: Futura, MacDonald and Co., 1995.

Welch, Jasper (Maj Gen, USAF, ret.), "Analysis of Requirements for Conventionally Armed Bombers," Northrop-Grumman, July 1994.

Wilkening, Dean, and Kenneth Watman, Nuclear Deterrence in a Regional Context, Santa Monica, Calif.: RAND, MR-500-AF, 1995.

[2] This draws on work developed over a decade, early portions of which were associated with development of the RAND Strategy Assessment System (the predecessor to the JICM model) and "multiscenario analysis" of Central Region and global-warfare issues. See Davis (1994), Chapter 4, for a summary. Similar ideas about scenario-space analysis have been proposed independently by Bonder and Cherry (Bonder, 1994) and applied to challenges of the post-Cold War problems of the Supreme Allied Commander, Europe.

[3] Some of the work was accomplished on networked Sun workstations using the JICM model. Other work was accomplished on a Macintosh computer using a more simplified spreadsheet campaign model (START). Our colleagues Carl Jones, Barry Wilson, and Jeff Hagan were responsible for much of this work. Bruce Bennett provided valuable advice.

[4] In 1991, the technology for this type of exploratory analysis was developed by our colleagues Steven Bankes and James Gillogly.

[5] The overall effectiveness of anti-armor sorties was quite low, perhaps 0.2 or so. However, most such sorties were flown by F-16s dropping dumb bombs from relatively high altitudes to avoid air defenses. By contrast, the bulk of vehicle kills were achieved by A-10s, F-15Es, and F-111s carrying precision-guided munitions (PGMs). Most of the PGMs, moreover, were used against stationary dug-in targets, rather than moving targets, which are easier to kill. It is reasonable to assume that by the period of interest (roughly 2005-2015), both Air Force and Navy aircraft will be predominantly precision-weapon capable. Some will also have sensor-fuzed weapons, which can be more lethal than earlier PGMs. For discussion see Bowie et al. (1993).

[6] Clear and differentiated data on air-to-ground effectiveness in the Gulf War do not seem to exist in the unclassified literature. See Frostic (1994) for a short description of how the various aircraft operated and the effects the pilots reported (including those of F-15Es and those F-16s with LANTIRN pods). See Keaney and Cohen (1993) for the official history. One interesting point Frostic made is that, where hard data exist, the weapon systems performed much as peacetime testing predicted.

[7] Our baseline assumes more rapid deployment of Air Force fighter wings than in Desert Shield (for which the average was about one squadron every other day, according to Kassing, 1992, p. 24). Although the fast deployments should be feasible in the future, especially with appropriate prepositioning, many problems can arise and results should not be taken for granted. This is the kind of crucial issue on which there can be substantial differences between what force planners expect and what is in fact achieved. As noted in the Kassing paper, there were many troubling discrepancies between planning factors and the performance of U.S. strategic lift.

[8] The feasibility of prompt seizure operations would depend on good use of warning, even prior to "C-Day." The 82nd airborne, afloat Marines in the region, and Rangers would all be candidate units. Prompt mine clearing would depend on early deployment of specialized helicopters from the continental United States.

[9] See also Bowie et al. (1993), which was probably the first significant publication suggesting such impact.

[10] For discussion of disruption see Keaney and Cohen (1993), Frostic (1994), and Chapter 7 of Pape (1996), which draws on several of the earlier studies that had been conducted independently.

[11] See Birkler, Perin, et al. (forthcoming) for comparisons of Air Force and Navy effectiveness in tactical air missions.

[12] There has been much controversy about whether more B-2s should be procured. The assessed value of such additional B-2s depends on many factors, such as warning time, access problems of the sort described above, the fighting quality of allied ground forces, the air defense environment (which affects whether B-52s and B-1s can be used over hostile territory), munitions on the B-2s, their effectiveness when operated in a hostile air defense environment (B-2 radars could be used only intermittently, and B-2s may be visually detectable in the daytime by interceptors), the enemy's attack formation, and the ability of the formation to keep moving despite massive losses. Aside from our own analyses, see Bowie et al. (1993), Welch (1994), Buchan (1994), and O'Hanlon (1995, pp. 145-149). The principal factor favoring the B-2 is that--if properly equipped and able to deal with residual fighter aircraft in the day, with escorts or parallel missile strikes against air bases--it would provide a powerful hedge against plausible short-warning or delayed-access cases.

[13] Some caveats: This statement assumes no ongoing lesser regional conflicts or peacekeeping operations that materially interfere, wise allocation of military resources rather than a "piling on" against the first adversary, and a series of important adjustments in support forces and stocks.

[14] The problem of "asymmetric strategies" has been studied in RAND projects for OSD (Bennett et al., 1994, reviewing several years' worth of "future-of-warfare games" and analysis) and jointly for the Air Force and Army (led by Kenneth Watman). It was studied intensively during a Defense Science Board study as well (DSB, 1996).

[15] A number of important technical developments are under way that could mitigate the problems. These include improved sensors that can operate in imperfect weather (e.g., millimeter-wave radar and differential use of the Global Positioning System). See, e.g., the Attack Volume of AFSAB (1996). Urban sprawl is a different matter because of the potential for killing innocent and perhaps friendly civilians. There seems to be no alternative to infantry-intensive operations, although advanced technology can certainly help a great deal.

[16] It should be possible to deploy such a unit, and missiles, within days. Some prepositioning would probably be desirable, especially if mobility and self-protection required additional forces. Navy options, such as those involving the Arsenal ship, might use similar or identical missiles and munitions.

[17] Based on discussions with colleagues Randall Steeb and John Matsumura, who have done extensive entity-level analysis on the effectiveness of precision weapons in tactical-level combat. See Steeb et al. (1996), Matsumura, et al. (1996), and DSB (1996) for more discussion.

[18] See work on "configural problems" in mine warfare and air defense problems in Horrigan (1995).

[19] Something notably absent here is discussion of advanced options for strategic bombing. Methodologically, it is difficult to demonstrate adequately the considerable value of strategic bombing. While our work to date has not adequately measured its value, there are reasons to believe that, while substantial (e.g., in destroying major elements of the air defense system and disrupting logistics and command and control), the value of strategic bombing against fixed targets, as distinct from invading armies, has often been exaggerated. Air power had awesome consequences in Desert Storm, but the most evident and significant were related to direct attacks on the enemy's forces. For an extensive discussion of strategic bombing, see Pape (1996), as well as Keaney and Cohen (1993).

[20] Some of the important but subtle cost issues involve time, constraints, and the value of accumulated small efficiencies. Many valuable improvements would require near-term investment, with savings realized only some years in the future. Constraints are a ubiquitous problem. Often tradeoffs are made within artificial funding categories, such as munitions with "deep attack" systems, rather than across categories. This often limits the ability to make economically and militarily rational choices. On the good-news side of the problem, we find that there can be considerable payoff, in both performance and cost savings, as the result of numerous "small" efficiencies. Sometimes looking for a package of such efficiencies is more fruitful than looking for the proverbial "silver bullet."