Document created: 3 September 03
Air University Review, July-August 1975
How accurately does a fighter pilot deliver a weapon on a target? What are his chances in a duel with an antiaircraft site? Does probability favor the destruction of a heavily defended railroad bridge attacked by a flight of four, or is the mission likely to be a failure? The questions seem so basic to military planning that it seems paradoxical that, in the early 1960s, almost fifty years after a German lieutenant first hand-dropped two four-pound bombs while flying over the outskirts of Paris, no one knew the answers.
Our German lieutenant, father of the concept of delivering weapons by air, was not too much concerned with measuring his accuracy. He did the best he could, as did his fellow World War I aviators. Their impact on the battle was minimal. Concern over accuracy during the Second World War varied from time to time depending on the nature of the target, its location, the weather, and the stage of the war. The air campaign in North Vietnam in the 1960s, in contrast, mostly involved missions where accuracy was of the utmost importance. Without an industrial base, North Vietnam had few significant targets. Even of these few targets that were of military importance, most were protected from attack by political considerations. During the first years of the air war in North Vietnam, or “Rolling Thunder” as it was called, targets for our fighter-bombers were essentially the lines of communication of the enemy. This meant roads, rail yards, lines and rolling stock, and bridges. North Vietnam, being a wet lowland cut by the Red and Black Rivers, was a land of bridges—fair game for the fighter pilot but among the most difficult of targets to hit. If the type of target required a high degree of accuracy, the constraints of the mission were demanding. The nature of the conflict, the way the war was conducted, and the criticism by the world press made large target misses, or “short rounds,” unacceptable. Rules of engagement followed by the pilot required not less than putting his weapons precisely where the mission planner dictated. A short round had to be accounted for in the court of world public opinion. Along with these demands for accuracy, there was a countervailing force militating against it.
The countervailing force was the enemy’s defenses. Inconceivable to those who have not encountered them, they were a powerful deterrent to the accurate placement of air-delivered weapons. The defense of North Vietnam was four-pronged. First, but not of greatest importance, was the enemy air threat. This amounted mostly to harassment but often caused F-105s to jettison their unarmed bombs to enter an air battle, from which the enemy then often tried to flee. With the advent of air protection by F-4s, this threat tended to diminish, enemy aircraft assiduously avoiding air-to-air engagement. What enemy attacks did come from the air were largely high-speed stern firings directed by ground-controlled interception (GCI), followed by rapid disengagement. Then there were the surface-to-air missiles (SAM). The kill ratio—number of aircraft downed to number of SAM’S fired—was misleadingly small. What the missiles did, before the introduction of fighter-carried electronic countermeasures (ECM) equipment, was to cause the tactical pilot to fly low, beneath the optimum altitude of the SAM guidance system and into the area of maximum ground-fire effectiveness. After the introduction of ECM equipment, the pilots flew above the ground fire but in a cumbersome “pod” formation, which made flying and bombing more difficult. The third prong of the enemy’s defenses was his high-altitude antiaircraft weapons. The 85- and 100-mm cannons were largely ineffectual and caused little distress to the fast-flying aircraft. It was the low-caliber, high-rate-of-fire weapons, 37- and 57-mm cannons and automatic guns, the fourth prong, that made fighter-bomber weapons delivery extremely hazardous and necessitated the standoff bombing techniques that were developed and used over North Vietnam.
Instead of delivering weapons from as low an altitude as possible, where accuracy was greatest, high release altitudes were chosen. Moreover, speeds were increased to minimize exposure time in the target area. So the defensive environment necessitated tactics that had an effect on accuracy that was beyond previous experience and, at that time, incalculable.
Shortly after our first employment of tactical aircraft in Southeast Asia, it became abundantly clear that very little was known about the accuracy of weapons delivery in a combat environment. Much was supposed, but beyond the optimism that our tactical aircrews could do the job expected of them, there were no firm accuracy figures on which to base war plans. What was known about accuracy of fighter-delivered weapons was learned from training bomb ranges. It is there that pilots developed their skills, going back day after day to their well-known, well-marked areas with highly defined targets (usually bull’s-eyes) to practice the techniques of ordnance delivery. Optimum conditions were used for each practice delivery: best airspeeds, minimum altitudes, turns at known geographic locations, and run-in lines—all this in usually good weather conditions and clear visibility of the Southwest United States, where most of the training ranges were located. In addition, there were ground observers to broadcast winds aloft, and, of course, there were no enemy defenses. Single-aircraft attacks, with each pilot concentrating on his own delivery, were the order of the day. Under these conditions, pilots calculated their circular error probable (CEP),* which gave us our only information about combat accuracy.
*The circular error probable is the diameter of a circle encompassing 50 percent of the weapons delivered.
Those who have not delivered weapons from an airplane have little or no conception of the problems involved or the requisite skills. There are so many variables in the accuracy equation and the chance for error is so great as to make one wonder how fighter pilots do as well as they do.
Dive bombing, for example, must take into account the ballistics of the weapon, the dive angle, airspeed, altitude, aircraft attitude, g (gravity) conditions, symmetricalness of flight, and wind. Of these, the only constant is the weapon ballistics, but even this is subject to errors due to manufacturing tolerances. Using tables, the pilot predetermines his release conditions—that is, airspeed, altitude, and dive angle—and computes a depression angle for his bombing sight. The reticle of the sight, if superimposed on the target when the pilot maneuvers to his preplanned angle, airspeed, and altitude, should provide an accurate release point. The difficulty, though, is in simultaneously achieving these three main variables as the reticle crosses the target. That is where the skill of the pilot comes in. If the airspeed is too fast or too slow, the dive angle too steep or shallow, the altitude too high or too low, the bomb will be long or short. Similarly, release at a g force other than the cosine of the dive angle (.866 g at a 30-degree angle, for example) will affect the bomb trajectory. Inability to hold the wings level will throw the bomb left or right of the target. Lack of a coordinated flight condition will do likewise. Wind, too, is a strong factor, drifting the aircraft during the prerelease run-in and affecting the bomb in free fall after release. The problem of a pilot, then, is not the same as that of a rifleman. A pilot may have the target centered under the reticle and still encounter gross errors. The chances for these errors, then, even under the ideal conditions of a training range, are significant.
Now take this same bombing problem of accuracy and place the pilot in a hostile situation where the chosen parameters are a much higher airspeed, steeper angles, and high weapon release altitudes; force him to fly a pod formation prior to bomb release; make it hard to find a target he probably has never seen before; make him keep one eye peeled for SAM’s; fill the sky with antiaircraft fire; and you will have some appreciation of the difficulty, in a combat environment, of putting a weapon precisely on target. It is because of these added difficulties of combat that the CEP’s scored in training, on which force planning was initially based in the Vietnam war, were erroneous.
New estimates of accuracy were urgently needed for a number of reasons. First, the mission planner must start with this statistical information to determine how many aircraft are necessary to achieve a certain probability of target destruction. Proceeding from this required number of aircraft, he can then apply known maintenance nondelivery rates, ground and air abort rates, and a percentage of weapons malfunctions to determine the total number of aircraft to schedule for a particular mission. The mission planner must also consider the probability of success, again based on expected accuracy, to determine the relative worth of attacking a particular target. A marginally important target, for example, that has a low probability of being destroyed might in fact not be worth attendant risks. Of crucial concern, too, is the matter of collateral damage, whether to surrounding targets that are important to avoid or, more critically, damage to friendly forces. It may perhaps be militarily feasible for an aircraft to attack enemy troops in contact, with a 99 percent chance of not missing by such a distance as to endanger friendly forces. A 70 percent probability, however, may not be acceptable because, on the average, three out of ten times the target will be missed by an unacceptable distance.
There were attempts during the Second World War and in Korea to measure bombing accuracy in combat, but none met with any degree of success. Methods generally involved using pre-and post-strike photography. The pre-strike photo was used to locate all existing bomb craters if the target had been struck before. Of course, a target that had not been attacked needed no such first step. After the mission, a reconnaissance aircraft would take photos of the target area, and photo interpreters would locate new bomb craters. Measuring the distance from the designated target to the bomb craters presumably yielded miss distances. The flaw in this method was that the target was not, in fact, precisely known. Although a rail yard, for example, might have been the designated target, there was no way to know where in the target complex the pilot was trying to center his bombfall; i.e., his desired mean point of impact (DMPI). It would have been highly inaccurate to assume that the center of the yard was the target because of the size of the target area. A crater at a rail choke point at one end, say, may have been a direct hit if that was where the pilot was trying to put his weapons. Otherwise it might have been a miss of several hundred feet. Even with more discrete targets such as buildings and bridges, the possible measuring error due to an unknown exact aim point allowed no better than gross estimates of accuracy. Further, there was no known method of determining which crater belonged to which pilot, an absolutely essential piece of information when individual pilots have different desired bomb impact points.
As a result of this inability to measure combat accuracy, the Joint Munitions Effectiveness Manual, the bible of the force planner, cited CEP’S developed on bombing ranges and suggested a degradation factor as a multiplier to estimate accuracy in combat. The factor, however, was conjecture, albeit an educated one, and it was subsequently found to be an inadequate predictor of success probability.
A new concept for measuring accuracy in combat came as a side product of the introduction of a new camera to the Vietnamese war theater. This was the Fairchild KA-71A and its follow-on model, the KB-18, a panoramic strike camera of high resolution designed to be carried aboard fighter-bomber type aircraft. The first cameras delivered to Southeast Asia were mounted in the nose of the F-l05, not for the purpose of measuring accuracy but to document what the pilot did; that is, what target he attacked and what damage ensued. The primary objective in using the camera was expressed in Air Force Regulation 95-13 as being the immediate evaluation of strike effectiveness. Such was the need for documentation that Lieutenant General William W. Momyer said, in July 1967, that we wanted a strike camera on every strike aircraft. Plans were therefore made to fit F-l05s, F-4s, F-l00s, and A-37s with the camera.
This camera, either housed in the body of the aircraft or carried in an external pod, was activated by the aircraft’s weapons release button and ran automatically for a preset length of time between 2 and 32 seconds. The lens aperture was controlled by an automatic internal light meter, and one, two, or four still pictures could be taken each second. With 250 feet of film, the capacity of the camera, approximately 300 exposures could be made. Designed for use in fighter or reconnaissance aircraft, the camera, with a rotating prism in front of the lens, was able to record, in one nine-inch frame, a scene encompassing 180 degrees vertically, fore and aft, and 40 degrees laterally. Such a side view of field generally recorded the complete flight path of the weapons from release to impact, no matter what aircraft maneuvers were performed after weapons release. Resolution of the resultant pictures was high and allowed precise location of weapon impact points.
Initial viewing of the product of this new camera, with its excellent portrayal of the battle scene, gave rise to the idea of measuring accuracy. To this end, a test program was started at the 388th Tactical Fighter Wing, Korat Royal Thai Air Base, in August 1967 to evaluate the F 105 on combat missions. The method of measurement developed was relatively simple. At the end of each daylight mission, the pilot of every strike aircraft was individually shown an intelligence prestrike photo of the target area and asked to indicate exactly where it was he was trying to center his bombfall. Such indication of the DMPI was made with an “X” mark. While this debriefing was going on, the film, which had been downloaded as soon as the aircraft landed, was being developed, an approximately 20-minute process. With prompt handling, the film was available for viewing at the completion of the pilot debriefing process. By careful analysis of the sequence of still photos, the weapons could actually be followed after their release from the aircraft to impact on the ground. Then by use of readily identifiable ground landmarks, bomb impact points were plotted on the original prestrike photo on which the pilot had indicated his target. (This was done because the scale of the panoramic photo of the strike camera was variable throughout the length of the photo and did not allow accurate measurements.) Distance was measured on the prestrike photo, which was a vertical image and had a constant scale. It was obvious in this initial study that bomb impacts could be plotted with a great deal of accuracy and that, if the pilot debriefing was handled correctly, the target could be pinpointed. The evaluation was direct and straightforward and led to a preliminary estimate of a combat CEP. Surprisingly, this preliminary estimate, which contained only 31 measurements, differed only eight percent from a more conclusive CEP derived some 500 samples later.
As a result of the brief but successful attempt to determine CEP’s, the decision was made at Headquarters Seventh Air Force to formalize the evaluation program and gather more data. Seventh Air Force Regulation 55-51 was written in September 1967 and directed procedures essentially the same as those developed in the test program. The postmission briefing was used to determine the pilot’s DMPI, and the strike film was studied to determine the point of weapons impact. Plotting of all data was done on a prestrike (vertical) photo of the target area. Since F-l05s were still the only aircraft with the KB-18 camera, the evaluation process was confined to Korat and Takhli Air Bases.
At about this time, forces were at work at Headquarter Pacific Air Forces (PACAF) to alter the program and, in effect, made serious changes in its concept. Despite the demurrers of Seventh Air Force, PACAF took over management of the evaluation, designated it Weapons Delivery Evaluation Program (WDEP), and expanded it to include peripheral data such as preliminary estimates of bombing accuracy, explanations of large miss distances, and maintenance reports. In addition, the governing directive, PACAFM 55-25, required that reports be identified by crew member and that wings keep a record of individual aircraft commanders’ CEP’s. The thrust of the program was clearly changed from one of analysis of accuracy to that of command and control.
The reaction from the wings was almost instantaneous. The program came to a virtual standstill, partly because of its new complexity but primarily because the finger was now being pointed at the pilot. Where reports had been anonymous in the past, records were now to be kept on individual accuracy. Cooperation, essential for the plotting of DMPI’S, died, and the data submitted dropped to very low levels, both unreliable and unacceptable.
As a result of the impending demise of the program and the importance attached to its original goal of determining accuracy, steps toward simplification were taken, and the reports once again were made anonymous. In addition, a new measuring technique was developed that obviated the need for a vertical photo of the target area. Designation of the DMPI and all measurements were made directly on the strike photos by use of a newly developed method of scaling. The changes brought a resurgence to the program.
In short order, more cameras were delivered to South Vietnam, and they were rapidly installed on the F-4, F-100, and A-37. The measurement program expanded, and the data being generated multiplied rapidly. By the end of the air campaign in Vietnam, the question originally posed, How accurate is weapons delivery by a fighter aircraft? was answered conclusively. Not only was there sufficient information to determine the accuracy of each of the camera-equipped aircraft, but a number of other parameters were addressed and quantified. Accuracy was determined by type of target, whether it was being struck for the first or a subsequent time, and by type of enemy defenses. It was possible to tell differences in accuracy when attacking in a SAM environment as opposed to just an AAA environment. Accuracy was plotted as a function of altitude, airspeed, and dive angle. The effect of weather was determinable. And, of great importance, accuracy with different types of weapons could be assessed. The figures were startling and gave rise to the questions of how valid they were.
The determination of bomb impact points was the least troublesome aspect of the measurement process. The estimate of photo interpreters was that they could pinpoint the location of weapon impact within 25 feet. Assuming no bias on their part, these measuring errors tended to cancel themselves out over an adequate sample size. Supporting this assessment of measuring accuracy was a study done by Joint Task Force Two at Sandia Base, New Mexico. In conjunction with an attempt to determine the feasibility of an aircraft recording system, they made a comparison between the measurement of miss distance on the basis of strike camera photographs and ground-scored impacts. Their investigation concluded that the CEP of error in weapon impact point measurement with strike cameras was 17 feet on photographs taken from the F-4D aircraft and 24 feet on those from the A-7A.
The determination of the DMPI was another question and was probably the greatest possible source of error within the measuring system. No matter how specific the target—a small bridge, for example—there was still a variety of DMPI’S the pilot could have chosen. The system depended on his accurately describing the target to the mission debriefer. To minimize this bias, the pilot had to indicate his DMPI before viewing weapon impact photography. There was the possibility, though, that he might have had a rough idea of where his weapons impacted. It would have been possible, therefore, if he so desired, for him to designate a DMPI that evidenced a greater degree of accuracy. There were mitigating factors, however. First and foremost was the PACAFM 55-25 prescription that if the pilot was unsure of his DMPI, no evaluation was to be made. Second, the report was anonymous, removing most motivation for indicating greater accuracy. Third, in a defended area, postrelease maneuvering often precluded observing impacts. Fourth, it was often difficult to adjust the DMPI rationally to any extent—if the target was a river ford, for example, a target downstream was not believable. Fifth, there was the integrity of the aircrews involved in the evaluation.
Another limitation of the measuring system was the possibility that a “short round” might have been out of view of the camera and that photographs were therefore taken only of bombs delivered close to the target. During operations in North Vietnam, however, there was a multiplicity of cameras photographing the same target at any one time, and the altitudes flown provided such wide coverage that the possibility of a weapon’s not being in at least one camera’s purview was minimal. On missions in South Vietnam, with only one or two cameras present and with lower weapons release altitudes, there was less photographic coverage in terms of square area—horizon-to-horizon and approximately 1500 feet wide when photos were taken at a 2000-foot altitude in level, nonturning flight. Compensating for this, though, was the greater pilot accuracy from at-tacking at lower levels.
At best, then, this measuring system was an imperfect tool. Yet the degree of imperfection was within reasonable limits, and the data were statistically valid and significant.
What was accomplished by this evaluation process that carried over the course of five years and involved all the Tactical Fighter Wings in Southeast Asia? Essentially, the program met its objectives. Only two questions were asked originally: “What was the pilot trying to hit?” and “Where did his weapons go?” The questions were answered definitively. For the first time, insight was gained into the effect of combat conditions on the delivery of ordnance. More was found out about bombing accuracy than was ever anticipated by those who conceived the idea of trying to measure it. Combat planning factors were revised to reflect the newly documented information. Beyond that, though, serious questions were raised regarding our avionics, weapons, and training. The bombsight being used in fighter aircraft was recognized as unsuitable for its task—an instrument designed for other conditions than those found in a heavily defended war zone. Improvements were made, and they were effective. The weapons, primarily unguided bombs, were supplemented by “smart bombs,” and the evaluation of their accuracy permitted the choice of the most effective weapon for a particular target. Training, too, was affected. Combat conditions were more closely simulated with practice attacks from higher altitudes, faster speeds, steeper angles, and from different attack formations. The evaluation program, in sum, had a profound effect on weapons employment and, by any criterion, was successful.
Maxwell Air Force Base, Alabama
Lieutenant Colonel Bernard Appel (M. S., University of Colorado) is Chief, Training Division, Headquarters AFROTC. He flew F-4s at Da Nang in 1967 and served at Headquarters Seventh Air Force and Headquarters PACAF in the field of tactical analysis. Other assignments have been primarily as a flight instructor in Air Training Command. Colonel Appel is a graduate of Air War College.
The conclusions and opinions expressed in this document are those of the author cultivated in the freedom of expression, academic environment of Air University. They do not reflect the official position of the U.S. Government, Department of Defense, the United States Air Force or the Air University.