The Simulation-Reality Mismatch

March 1, 2010

Historically, there has been a mismatch between the air combat attrition predictions of analysts’ computer models and the actual attrition experienced in major air campaigns. Simply put, the models have always erred on the pessimistic side.

Understanding the sources of this mismatch is important today because many Air Force and Pentagon experts are at it again. This time, they are claiming that even stealthy fifth generation fighters and B-2 bombers cannot penetrate future threat arrays without unacceptable losses. They conclude that the Air Force must depend almost solely on weapons launched from standoff distances to defeat future integrated air defense systems (IADS).

An Air Force B-52 bomber, centerpiece of Linebacker II. (USAF photo)

If this viewpoint takes hold, the consequences would be severe. Such a belief undermines the perceived value of airpower, weakens our ability to deter potential adversaries, and will raise the cost of equipping future air forces.

However, an accurate review of history reveals a far different picture of the ability of USAF aircraft to penetrate contemporary defenses.

In the Linebacker II campaign over heavily defended Hanoi and Haiphong in December 1972, analysts predicted we would lose one B-52 out of every three B-52 sorties flown. We lost 15 B-52s out of 729 B-52 sorties in the 11-day campaign, an attrition rate less than two percent, not 33 percent.

In the June 1982 Bekaa Valley War, the Israeli Air Force mounted a mass suppression of enemy air defenses (SEAD) campaign against a Syrian air defense force in Lebanon equipped with modern, state-of-the-art Soviet systems. There, analysts’ models of similar threat arrays predicted an attrition rate of 10 to 20 percent. In 1,100 fighter sorties, the IAF lost no aircraft—zero percent attrition, not 10 to 20 percent.

In the first four days of the January 1991 Desert Storm campaign over heavily defended Baghdad, Air Force leaders predicted a loss of 30 to 40 combat aircraft on the first night, 20 the second night, 10 the third night, and five on the fourth night, for a total of about 70 losses. Analysts’ models predicted far greater losses. Coalition forces flew 7,000 combat sorties (shooters) in those four days and lost 27 fixed wing aircraft, an attrition rate less than 0.4 percent, less than half what had been predicted by Air Force estimates, and way below analysts’ predictions.

The Faults of Modeling

Why do analysts’ models invariably predict high loss rates against state-of-the-art defenses when actual combat experience yields significantly lower attrition? The answer lies in understanding modeling limitations and the actual execution of operational strategies and tactics in air campaigns.

Models that simulate combat between friends and foes are useful to assist combat planners and tacticians in developing strategies and tactics, and to find gaps in capabilities for those who develop operational requirements. They also assist scientists and engineers in understanding the technical innovations and developments necessary to counter threat systems.

Models are also valuable for evaluating the relative impact of adding or subtracting threats and friendly forces, but not for absolute outcomes.

Good analysis identifies the variables of combat actions having the greatest leverage. Competent analysts then perform sensitivity studies to understand the impact of altering variables, but not to predict absolute answers. Good modeling is meant to inform decision-makers of the relative consequences of different courses of action.

North Vietnamese SAM battery, a key threat to US aircraft. B-52 attrition in the 11-day campaign was less than two percent, not the 33 percent predicted by analysts’ models. (USAF photo)

However, models lose their fidelity when they try to simulate large-scale air campaigns because they cannot faithfully replicate the enormous complexity, including the countermeasures, tactics, and on-scene decisions of actual combat. Many models simply take a single, one vs. one, encounter and extrapolate the results to an entire air campaign. That is grossly faulty modeling. They invariably grant to the adversary capabilities in both operator competence and system performance that exceed actual experience. These three examples from the real world illustrate the dangers of trying to use modeling and simulation beyond their useful purposes.

Modern air defense systems consist of surface-to-air missiles (SAMs), radar-directed anti-aircraft artillery (AAA), fighter interceptor aircraft, and early warning, target acquisition, and target tracking radars. These systems are joined together via voice and data links in an integrated command and control network. The entire system of systems is called an integrated air defense system. Every nation that we would consider a potential adversary deploys IADS around its high-value assets.

A National Asset To Retain

The first SAM-based IADS arrayed against US forces was in North Vietnam, and our Air Force was ill-equipped to deal with it. The North Vietnamese, with Soviet advisors, had pieced together a Russian-style IADS to protect high-value targets such as airfields, power plants, harbors, and command centers. The centerpiece was the SA-2 missile and Fan Song radar arrayed in batteries of six missiles each. USAF took heavy losses until it equipped its fighters with radar warning and self-protection electronic countermeasures pods, and provided defense suppression assets such as the F-105 Wild Weasels with anti-radiation missiles and support jamming EB-66s. However, political constraints on attacking targets in the Hanoi and Haiphong areas prevented a full-fledged suppression and destruction of the IADS, so airmen had to operate in the face of SAM and AAA defenses.

That changed with Linebacker II. The politicians removed all targeting constraints. The North Vietnamese had upgraded their IADS with an improved radar system and many more SAM batteries. However, the Air Force was able to plan and execute a rollback of the SAM, AAA, and radars systematically to allow the B-52s to bomb their primary targets effectively. The combination of support jamming with standoff EB-66s, lethal defense suppression with F-105 Wild Weasels, self-protection jamming and chaff on the B-52s, use of decoys and deception, and smart flying tactics on the part of the B-52 crews resulted in mission success with few losses.

What did the analysis miss? Well, it did not account adequately for the systematic destruction of the IADS components as a prelude to the bombing of primary targets. It did not account sufficiently for the effectiveness of support and self-protection jamming by all the players. It did not account for the decoys and deceptive tactics employed to confuse the enemy. It did not account for the changes in tactics of the B-52 crews as they varied their routes, call signs, attack parameters, and escape maneuvers in succeeding waves of attacks. It did not account for the aircrews’ on-scene, real-time situational awareness that allowed the crews of individual flights of fighters and bombers to make changes on the fly—audibles at the line of scrimmage, as it were—to succeed.

This latter factor, decentralized execution, the ability of USAF aircrews to operate relatively independently of the command and control system if necessary, and make smart decisions during execution, is the product of realistic training programs and the competence of Air Force people that remains the hallmark of America’s combat pilots and crews. Models cannot simulate this unique strength. This factor, alone, is a national asset that the Air Force must retain now and in the future.

In combination, these factors overwhelmed the North Vietnamese IADS, resulting in less than two percent attrition for the lumbering, nonmaneuverable B-52 force, a rate that the analysts had predicted would be at least 10 times higher.

392nd Air Expeditionary Group maintainers inspect damage to an A-10 hit by an Iraqi missile. (IDF photo)

In 1982, at the time of the Israeli Bekaa Valley attacks, the Syrians, with hundreds of Russian “advisors,” had deployed the most modern IADS available. It consisted of 19 batteries of SA-3 and SA-6 radar guided SAM batteries, SA-7 infrared guided SAMs, ZSU-23 radar guided AAA, and MiG-21 and MiG-23 fighters. The Soviets hailed this array as the showcase of air defenses. It was a model of the IADS they had arrayed against NATO in Eastern Europe. It was their state-of-the-art system to demonstrate to the world how effective an IADS could be. The analysts’ models also considered it nearly impenetrable without heavy losses. They projected loss rates of 10 to 20 percent for aircraft flying into this defense system.

The Israeli government considered this escalation of forces in southern Lebanon a hostile invasion of its right to protect and defend its citizens in northern Israel. Syrian provocations at the time also forced the Israelis to act. On June 9, 1982, the Israelis, flying about 600 F-15 and F-16 sorties and 100 support sorties, destroyed 17 of the 19 SA-6 batteries and downed 23 MiG fighters without a single loss. The initial wave of attackers took only about 10 minutes of concentrated, intense action over the Bekaa Valley to accomplish this victory. The Israelis followed this attack with a second day of intense bombing and air combat. Israeli pilots destroyed the remaining two SAM batteries and another 15 MiGs, again without a single loss.

So, why the mismatch between the models’ predictions and the actual outcomes? The Israelis were equipped with US-supplied fighters and electronic combat systems, and their strategy and tactics were a mirror image of Air Force training in Red Flag exercises over the Nevada ranges.

They perfected the combat principles of surprise, force saturation, deception, synchronization of attacks, and operating inside the opponent’s reaction time, all simultaneously. They were superbly trained, and their equipment, designed to counter just such a defensive array, worked as designed.

The attrition models, as detailed as they had become by 1982, were not able to account for the dynamics, on-scene decisions, and operational tactics of the actual battles. They also simulated a higher than warranted level of proficiency by the Syrian and Russian forces. All in all, the Bekaa Valley campaign demonstrated the severe limitations of modeling when used to predict absolute force-on-force attrition.

F-117 stealth fighter aircraft in the Gulf (USAF photo)

Greater Efficiency, Less Attrition

In the years between Linebacker II and Desert Storm, the Air Force recognized the continued huge investment in SAM, MiG, and radar-based defenses by the Soviets and their client states. Having learned its lessons well, the Air Force embarked on its own vast investment in electronic combat systems and electronic warfare training. The Air Force upgraded its Wild Weasels with the F-4G and the High-speed Anti-Radiation Missile, developed self-protection ECM jammers and chaff and flare dispensers for its combat aircraft, fielded the EF-111 for close-in support jamming, developed wide-area emitter location systems, and developed new standoff weapons like the Conventional Air Launched Cruise Missile. The Air Force also developed night attack systems such as LANTIRN, to deny the enemy the sanctuary of night operations. The Air Force also sharpened its operational strategies and tactics in large-force exercises in Nevada with Red and Green Flag programs. It deployed an array of simulated SAM missiles, radars, and MiG fighters, called Aggressors, in the Nevada desert to hone its capabilities against modern threat arrays.

Most importantly, however, the Air Force began its investment in stealth technology as the means to negate the massive investment in radar-based defenses by the Soviets and their client states. This investment has paid big dividends with the fleet of stealthy fighters and bombers, such as the F-117, F-22, F-35, and B-2, and JASSM cruise missiles. Now, the Air Force can execute air campaigns with smaller force packages, led by stealth fighters and bombers. This allows for greater efficiency, even less attrition, and overwhelming results.

If DOD and the Air Force continue to invest wisely in the technologies required to stay at least a generation ahead of any potential adversary, and train USAF crews to think smartly and with innovation, then US forces can win quickly, decisively, and with low attrition regardless of the analysts’ predictions to the contrary.

This investment and operational strategy paid big dividends when put to the test in Desert Storm. The Iraqis had netted together a sophisticated French-built IADS, called KARI. Analysts estimated that, outside Moscow, Baghdad had the highest concentration of air defenses of any city in the world. It consisted of the most modern, state-of-the-art Soviet SAMs and MiGs, including the MiG-29. Measured by the attrition models at the time, this threat array should have extracted an attrition rate of 10 to 20 percent during the first few days.

Downtown Baghdad target on Jan. 17, 1991, seen through an F-117 strike camera. In the first four days of the war, the coalition attrition rate was less than 0.4 percent, far below estimates. (USAF photo)

The air campaign, led by the stealthy F-117, succeeded because the coalition forces overwhelmed the IADS during three days and nights of intense air attacks. It was not easy and required a significant change in tactics after the second night of attacks. The attacking aircraft shifted from low-altitude attacks to medium-altitude attacks to avoid the intense, barrage, firing by small-caliber AAA and shoulder-fired SA-7 IR missiles that had downed several aircraft the first two days. Low altitude (under 200 feet) had been the preferred tactic, except for the stealthy F-117, in nearly all the training for defense suppression missions. Flying under 200 feet allowed the attackers to avoid radar detection and radar guided SAMs. In Desert Storm, however, the electronic suppression and lethal attacks by HARM missiles on SAM radars had negated the SAM threat early, allowing the Desert Storm air commanders, Lt. Gen. Charles A. Horner and Maj. Gen. Buster C. Glosson, to order attacks at medium altitudes and avoid the intense barrage fire at low altitude. This is another example of making changes during the course of the battle that are not accounted for in attrition modeling.

In Desert Storm, the coalition forces achieved complete control of the air within seven days with an attrition rate under 0.4 percent. For the entire war, the coalition lost one aircraft for every 1,800 combat sorties flown. However, the important lessons from Desert Storm were the execution of the SEAD campaign with smartly conceived force packages, operating inside the opponent’s reaction time, employing stealth, electronic combat, decoys, and intense attacks with manned fighters, and changes in tactics as the campaign progressed, to reduce attrition and increase effectiveness.

In all three of these campaigns against formidable, state-of-the-art air defenses, the actual attrition was well below that predicted. Today, analysts in DOD and contractors supporting them are once again predicting that manned fighters and bombers cannot fly in the teeth of modern defenses without unacceptable losses. They conclude that only standoff weapons and unmanned vehicles should penetrate these defenses until, and if, the threats are reduced to zero or low threat conditions allow other aircraft to perform their missions.

There are many problems with this conclusion. First, as these three examples have shown, smart planners with multiple attack systems can plan SEAD campaigns that penetrate and take down even the toughest defenses with a combination of a few standoff attacks and many force packages with penetrating attackers and support systems.

Infrared image of a nighttime target, seen on an F-15E head-up display. Use of such systems to deny enemies respite from attack, combined with innovative tactics, have helped USAF keep down its own attrition and raise mission success rates. (USAF photo)

The Human Touch

The objective of a SEAD campaign against IADS is to gain and maintain control of the air so as to allow unimpeded ground and air operations from that time forward. This requires the right forces, surprise, concentrated attacks to confuse the adversary, operating inside the opponent’s reaction time, and destroying the nervous system of the IADS quickly and decisively as the three examples have shown. A force cannot do that with standoff attacks alone, or with unmanned attacks whose links can be severed, whose situational awareness is deficient, and whose operators cannot respond quickly as conditions change during execution of the mission. Thus, it is imperative that the US Air Force retain the capability to conduct full-scale SEAD campaigns with manned, penetrating bombers and fighters that can accomplish the mission with acceptable attrition.

This means continued fielding of fifth generation fighters, development of a new stealthy, manned penetrating bomber, a new stealth air launched cruise missile, and significant investments in stealth, electronic combat, and cyber technologies. As potential adversaries continue to field more capable IADS, DOD must stay at least a generation ahead.

Manned bombers and fighters also have much greater flexibility to change missions and to adapt to different scenarios than do cruise missiles and unmanned aircraft. Bombers and fighters can launch standoff weapons as well as penetrate. On the other hand, large platforms designed to act as “trucks” launching cruise missiles have little value in other applications.

Can modeling and simulation be improved, shedding its limitations so as to allow accurate predictions of the outcome of large-scale campaigns? RAND researchers, under contract to DOD, have recognized the severe limitations of campaign models that are forced to operate at higher and higher levels of complexity. DOD is attempting to increase the fidelity of its models and use them only for limited applications, but researchers acknowledge that no model can simulate the thousands of decisions and interactions that occurred in real-world SEAD campaigns such as the Linebacker II, Bekaa Valley, and Desert Storm examples. Models may be good for measuring relative outcomes in small encounters but not absolute attrition for large air campaigns.

As the Air Force develops the operational requirements for long-range strike, electronic warfare, and future combat aircraft, it needs to recall the historical examples that demonstrate that integrated air campaigns, with SEAD as the first priority, work if planned and carried out with vision and air leadership.

An F-22 fighter goes through its paces over Guam. Stealthy fighters, bombers, and cruise missiles, as well as new electronic combat and cyber weapons, will be key to future attrition-denial schemes. (USAF photo by SrA. Christopher Bush)

The challenge for Air Force leaders today is to recognize that, however potent they appear on paper, IADS have exploitable limitations. Then, they need to invest in the right sets of systems—bombers, fighters, weapons, electronic combat, deception, cyber—and aircrew training programs to defeat them, underpinned by a strong science and technology investment. This will ensure that future air forces do what past air forces have always done, gain and maintain control of the air quickly and decisively, despite the predictions of the critics who argue otherwise. That is, after all, the primary mission of the Air Force.

Retired Gen. John Michael Loh was commander of Air Combat Command (1992-95), commander of Tactical Air Command (1991-92), and vice chief of staff of the Air Force (1990-91). He flew 204 combat missions in the Vietnam War and retired in 1995 after 35 years of active duty service. This is his first article for Air Force Magazine.