Introduction
The armed military airplane made its formal debut on the battlefield in World War I and has subsequently proved a weapon of enormous significance, as increasing aviation technology has extended its range, speed, ordnance load, and general sophistication. These very characteristics that make the airplane such a potentially useful and destructive weapon have also made it a significant threat to friendly troops.
As the speed and range of military aircraft have increased, the difficulties of control and coordination have increased accordingly. Indeed, the present technological sophistication of military aircraft, both fixed and rotary wing, like that of modern artillery weapons, threatens to outstrip the capabilities of its human operators to control its employment adequately.
Among the 269 cases of amicicide identified in this
study, incidents of air amicicide exceeded artillery incidents
in frequency. Ninety-nine of 269 cases (37 percent) involved
the engagement of friendly troops by their own aircraft. More
significant, in terms of the number of friendly casualties
caused, air incidents clearly predominate, both as to total
casualties and casualties per incident. In one World War II
incident alone, 111 friendly troops lost their lives and 490
were wounded. [1]
The effects of air amicicide on friendly combat power
and the conduct of operations are the same as those of
artillery incidents, but they are magnified by the greater
destructive power of air-delivered ordnance. Death and wounds,
as well as confusion, result from the bombing and strafing of
friendly troops by their own air forces, and such incidents
can and do have adverse effects on combat power, the progress
of offensive operations, the viability of defensive positions,
and the morale and confidence of troops. Indeed it appears
that morale and confidence in supporting arms are more
seriously affected by incidents of air amicicide than by
artillery incidents. While ground troops may grudgingly
acknowledge that an artillery strike might have been
perpetrated by the enemy rather than by friendly guns, or even
that the friendly artillery could not, for whatever reason,
know their location, they are far less able to under
stand why friendly aircraft, which they could see perfectly
well, could not see and recognize them. [2]
Although as a result of previous study and analysis,
the available data on air incidents are somewhat more complete
and detailed than that for artillery amicicide, they too
remain sketchy and generally unreliable. They do suggest,
however, that the incidence of both types of amicicide respond
to the same three general factors.
Technological advances in aircraft design have
increased speed, range, and ordnance load. These improvements
have complicated the problem of adequately controlling the
placement of aerial fires, speed alone serving to make
enormously more difficult the correct identification of ground
troops. Concurrent improvements in communications and
electronic methods for locating both friendly troops and
friendly aircraft have offset somewhat the problems of control
brought on by advancing aviation technology.
The development of close air support doctrine in World
War II and the subsequent heavy and frequent use of armed
aircraft in support of ground combat forces made the
occurrence of air amicicide inevitable, given the available
identification and location technology. The addition of very
close support by armed rotary wing aircraft in Vietnam and the
use of extremely fast, high performance jet aircraft in the
close air support role only served to increase the hazard.
Continued use of tactical doctrines calling for close support
of ground forces by air forces will, of course, do nothing to
reduce the possibility that friendly troops will fall victim
to the fires of their supporting aircraft, just as they may be
subject to their own close artillery fires.
Just as in the case of artillery incidents, air
amicicide in the final analysis is usually the result of some
human error. The rapid advance of aviation technology, even
with a corresponding increase in communications and position
location technology, threatens to overwhelm the capability of
even the most competent and best trained human beings to cope.
It is too much to hope that a pilot, diving at 600 mph
through smoke while taking evasive action and attempting to
deliver area-type ordnance accurately, could instantaneously
and correctly identify camouflaged friendly ground troops
making maximum use of available cover and concealment. As
was noted in the "Lessons Learned" section of HQ, Advanced
Allied Expeditionary Air Force, letter of 20 July 1944:
"In all air support operations conducted extremely
close to the friendly troops, there will be the danger
of bombing and strafing behind the friendly front
lines. In case of air preparation close to friendly
troops, this hazard must be recognized and accepted by
the requesting agency." [3]
World War I
Although the armed military airplane became a
significant weapon of war in the course of World War I, its
slow speed and light armament, as well as the existence of
clearly defined front lines and the absence of a true close
air support doctrine, rendered incidents of air amicicide few
and relatively insignificant. Friendly troops on both sides in
World War I were bombed and strafed by their own aircraft, but
rarely did such incidents have any major impact on the course
of combat operations. Amid the tremendous casualties incurred
by ground weapons, the deaths and wounds attributable to
friendly air strikes went almost unnoticed.
One curious World War I incident of air amicicide,
however, did have some influence in stifling the development
of a particularly frightful method of aerial bombardment. At
the very beginning of the First World War, in August 1914,
the French invented a device consisting of a large metal can
filled with minuscule steel flechettes and designed to be
dispersed from aircraft. The use of this weapon was
discontinued by the French (although the Germans copied and
used it) when an aviator dropped some of them in error on a
detachment of Zouaves. [4]
It was found that the flechettes
were poisonous because their fall through the air was not
rapid enough to clear them of the oil in which they were
packed, and the resultant wounds became infected.
World War II: North Africa and Europe
The development of a doctrine of close air support and
the frequent employment of large numbers of friendly aircraft
in the battle area in World War II caused the problem of air
amicicide to grow to significant proportions. The danger posed
to friendly ground troops by supporting aircraft was
recognized early in the war, and various procedures, at first
visual and later increasingly electronic, were developed to
reduce the occurrence of such incidents. The slow but steady
progress achieved through better communications and improved air-ground
coordination procedures up to June 1944 was temporarily undone* by the greater intensity of tactical air operations in northern Europe and the repeated, and mostly unsuccessful, attempts to employ heavy bombers for close air support in the months immediately following the Normandy invasion. Eventually, technology and technique once more began to overcome the problem, but despite such improvements in all theaters, air amicicide incidents throughout the war continued to exact a heavy toll on friendly ground forces as well as friendly, or at least neutral, civilian
populations. The primary factors involved in World War II air amicicide, other than the adolescent state of the art of close air support and its technological handmaidens, were clearly human ones not amenable to technological solutions. Lack of coordination, the failure to identify ground troops as friendly, and direct pilot or navigator errors predominate as the causes of most World War II incidents.
American air operations in support of advancing ground forces faced their first real test during the campaign in North Africa in 1942-43 and in Sicily in 1943. Until adequate systems of air-ground coordination were worked out, US Army forces had to endure several attacks by friendly aircraft due to poor coordination or misidentification as the result of inadequate marking systems.
Although the British, for their part, faced severe problems in communications and in identification of friendly ground troops, by 1942 Air Vice Marshal Coningham's Western Desert Air Force (RAF) had developed usually effective methods for avoiding air amicicide. Tactical air support at the battle of El Alamein on 30 October 1942 is a case in point. Despite a restricted (nine-mile square) area in which the bomb line constantly shifted, none of the ninety-five tons of bombs delivered by more than 300 sorties fell on friendly forces. [5]
The results of three years' experience by the RAF could not be immediately absorbed by US air forces, however, and many hard lessons had to be relearned by the Americans on their own.
On 9 November 1942, the 1st Battalion, 60th Infantry, commanded by Maj.
Percy DeW. McCarley, Jr., failed to mark its position with the prescribed
identification panels during a pause before the final advance on French positions around the airfield at Mehdia-Port Lyautey, Morocco. A US Navy plane dropped two bombs among the troops, and the resultant disorganization delayed preparations for the attack until approaching darkness made it necessary to postpone the advance until the following day.
[6]
The 701st Tank Destroyer Battalion, allocated to the US
1st Armored Division, was among those units repeatedly hit by
friendly air strikes during the North African campaign. One
company of the 701st TD Battalion was attached to the British
11th Brigade in the vicinity of Medjez-el-Bab, 22-30 November
1942. The company operated successfully against German Mark IV
tanks on 23 November 1942, but the following day was attacked
by US P-38s and "practically all its vehicles were temporarily
put out of action." [7]
The disabled vehicles were cannibalized, and three tank
destroyers were soon returned to service with another three
subsequently repaired. Less than two weeks later on 2 December
1942, during the battle of the Zaid Pass, the 1st and
Reconnaissance Platoons of Company B, 701st TD Battalion, were
strafed at 1020 by one of a flight of four P-38s that flew
over their position. Three men were killed and two were
severely wounded. The author of the unit's North African
campaign diary commented:
"This case of mistaken identity was inexcusable and
highly demoralizing to us. The explanation was that our
vehicles were too far out, that is, outside the bomb
line. As a matter of fact, our vehicles had been
drawing enemy artillery fire from the vicinity of the
pass; had they been any closer in the pass they could
not have performed their assigned mission." [8]
Portions of the 701st TD Battalion were also among the 1st
Armored Division units bombed by US B-25s near Station de
Sened, Algeria, on the late afternoon of 3 February 1943. [9]
On a few occasions it was absolutely impossible from
the air to identify a target as friendly. A German ship was
loaded with Allied POWs in Tunis on 4 May 1943 and lay
anchored off Cape Bon for three days before the Germans
abandoned it. During that time at least forty Allied fighters
strafed the ship and aimed 100 bombs at it. Fortunately, the
fighter pilots proved somewhat unskilled, only one of the
bombs, a dud, hit the ship, and only one of the Allied POWs
was killed. [10]
Procedures for the positive identification of friendly
ground forces from the air had not noticeably improved by the
time of the Allied invasion of Sicily in July 1943. The
movement of a company of tanks from CCA, 2d Armored Divison,
toward Canicatti was briefly delayed on 11 July 1943 when its
march column was attacked by a P-38. [11]
The incident resulted in no casualties or equipment
loss, but it was only the first of several such attacks during
the week of 11 July that cost CCA fourteen vehicles and
seventy-five men, as friendly pilots, alert for the 15th
Panzer Grenadier Division,
repeatedly mistook the armored vehicles of CCA for enemy one.
The prominent display by CCA of yellow smoke, the agreed
identification signal for friendly troops, did little to
alleviate the problem.
In fact, one pilot, 1st Lt. R. F. Hood of the 86th Fighter-Bomber Group, who was shot down by CCA's antiaircraft fire,
admitted he had seen the yellow smoke, but did not know its
meaning. The 15th Army Group subsequently changed from smoke
to pennants as the means of identification, and thereby
reduced the problem.
Unfortunately, timely improvements in recognition
procedures did not entirely close the gap in air-ground
coordination. On 12 August 1943, Lt. Col. Lyle A. Bernard,
commanding the 3d Infantry Division's 2d Battalion, 30th
Infantry, ordered his Company F down from Monte Cipolla to
relieve Company E in defense of the Brolo River line. [12]
Delayed by German fire, Company F did not reach the flats and
move toward the river until almost 1600. Its arrival coincided
with a prescheduled air strike.
Seven A-36s arrived over Monte Cipolla about 1600 and,
probably as a result of pilot misorientation to the ground,
dropped two bombs on the battalion CP, causing nineteen
casualties. They dropped the remainder of the bombs on the
supporting howitzer positions of Battery A, 58th Armored Field
Artillery Battalion, destroying the remaining four howitzers.
Although Company F was unharmed, Lieutenant Colonel Bernard,
deprived of his remaining support for the two companies along
the river, was forced to withdraw his troops onto Monte
Cipolla in anticipation of a final stand against the opposing
elements of the 71th Panzer Grenadier Regiment.
Interviews conducted by Army Ground Forces observers in
Sicily during and after the campaign testify to frequent
bombing and strafing of friendly troops by Allied aircraft
despite the supposedly agreed upon use of panel markers and
yellow smoke as recognition signals. Although yellow smoke was
probably difficult for the pilots to distinguish from dust and
smoke from fires on the ground, the opinion of ground troops
was clearly that the pilots were unfamiliar with the meaning
of yellow smoke despite the understanding of the ground forces
that it was to be the primary recognition signal. [13]
Some improvements in air-ground coordination were made
during the Sicilian campaign, however. Noticeable advances in
the use of radio communication and mobile air control parties
would prove of tremendous value as the war moved onto the
continent. [14]
The destruction of one of the oldest centers of Western
Christian monasticism, the venerable Abbey of Monte Cassino,
by Allied aerial bombardment in March 1944 is still viewed by
many as not only a clear-cut case of amicicide but a crime
against civilization itself. Whether the abbey was occupied by
the Germans has little bearing on the fact that the 15 March
bombardment of Cassino resulted in the first major air
amicicide incident of World War II. Between 0830 and 1200 on
15 March 1944 some 435 Allied heavy and medium bombers dropped
more than 1,100 tons of 1,000-pound HE bombs in the Cassino
area. [15]
Some of the bombs from the heavies fell on Italian
civilians and Allied troops in the area, demolished the HQ
trailer of the British Eighth Army commander three miles away,
and among the Allied soldiers caused casualties of 28 killed
and 114 wounded. [16]
At the same time Allied aircraft bombed friendly troops
(including a French corps headquarters) in the town of
Venafro, ten miles from Cassino, killing 57 Allied soldiers
and friendly civilians and wounding 179. [17]
The mistaken bombings at Cassino and Venafro were due
to several causes. General Mark Clark attributed the tragic
errors to "poor training and inadequate briefing of
personnel," and the commander of the 6th New Zealand Brigade
later stated: "Heavy bombers operating from 14,000 feet are
not accurate enough for this class of close support. Medium
and light bombers are excellent. If the air force could have
used more medium bombers and still kept up the weight of the
attack, the results would have been better." [18]
The malfunction of a bomb rack on one lead aircraft had
resulted in the dropping of forty bombs on friendly positions,
and investigation revealed that poor air discipline,
obscuration of the target by smoke and dust, and the lack of
specific aiming points also had contributed to the fiasco. [19]
Regardless of the causes, the effects were noticeable.
Confusion reigned, friendly troops and civilians were killed
and injured, and the planned ground attack proceeded slowly
and unsuccessfully.
The greater scale and intensity of both air and ground
operations in Western Europe after the Normandy invasion in
June 1944 increased both the number and seriousness of air
amicicide incidents. And to the hazards of being strafed by
fighter aircraft or bombed by attack planes was added the even
greater hazard of serious heavy bombing by medium and heavy
bombardment aircraft employed in the close support role, as at
Monte Cassino.
As a result of detailed wartime and postwar analyses of
air operations in the European theater, the occurrence of air
amicicide, its causes and effects, and the measures taken to
prevent it are much better documented than are events in other
theaters. [20]
Thanks to the efforts of both official agencies and
civilian historians, only a brief review of the most
significant incidents need be made in this study.
It should be pointed out that during the course of the
campaign in Europe, constant attention was given to the
problem of properly coordinating air-ground operations.
Procedures for marking friendly positions, ground-to-air radio
communications, and radio/radar position location aids were
all improved. They reduced, but did not eliminate, the number
of amicicide incidents. Until the very end of the war in
Europe, pilot and navigator/bombardier errors continued to
result in friendly casualties, broken attacks, and the
degradation of friendly combat power.
Allied ground forces on the Continent were not the only
victims of Allied air power. Friendly civilians as well as
neutral populations were occasionally the unintended targets
of massive air operations. Before proceeding to describe
incidents of purely military significance, a brief, separate
account of representative incidents involving the bombing of
friendly or neutral territory may be instructive.
The US Army Air Forces bombed neutral Switzerland
several times between 1943 and the end of the war in 1945,
causing severe diplomatic repercussions, as well as death,
suffering, and heavy property loss among the presumably
unintended Victims. Helmreich has attributed these incidents
to a number of causes including bad weather, faulty equipment,
and incompetence or excess zeal on the part of air crews; but
he also suggests that the bombings may have been other than
accidental. [21]
On 1 April 1944, fifty American planes bombed the Swiss
city of Schaffhausen, causing serious fires and property
damage and killing or wounding more than 100 civilians. On 22
February 1945, even as presidential aide Laughlin Currie was
laying a wreath on the graves of the Swiss victims of the
Schaffhausen bombing, US planes took part in thirteen separate
attacks on Swiss territory, the most serious of which was at
Stein-am-Rhein, only twelve miles from Schaffhausen. The 22
February incidents killed or wounded more than thirty Swiss
citizens. [22]
On 4 March 1945, six B-24Hs dropped 12.5 tons of high
explosives and 12 tons of incendiaries on Zurich, and nine
others dropped 16.5 tons of high explosives and 5 tons of
incendiaries on Basel. [23]
The incidents, attributed to faulty equipment, bad weather
over France and haze over Switzerland, navigational errors,
and misplaced zeal, resulted in heavy damage to the main Basel
railway freight station, seven civilians injured in Basel, and
five civilians killed, twelve hospitalized, and twenty-two
families left homeless in Zurich. [24]
The pilot and the navigator of the lead plane in the Zurich
raid were subsequently tried and acquitted of violating the
96th Article of War. Their court-martial was held at HQ, 2d
Air Division, 8th Air Force, at Horsham St. Faith, England,
on 1 June 1945, and was presided over by Col. James M. (Jimmy)
Stewart. [25]
Friendly civilians in towns closer to the battlefront
also suffered severely from Allied air bombardment. As a
result of a "gross error due to poor navigation, poor headwork
and misidentification of target" one group of medium bombers
of the 9th Bombardment Division hit the Belgian town of Genck,
twenty-eight miles west of the assigned target, on the morning
of 2 October 1944, killing thirty-four civilians and wounding
forty-five. [26]
Two of the best documented incidents of this type
occurred at Malmedy, Belgium, on 23 and 25 December 1944. [27]
At 1526 on 23 December 1944, six B-26s of the 322d Bomb
Group (9th Bombardment Division, Ninth Tactical Air Force),
flying in support of the 30th Infantry Division, then heavily
engaged with Kampfgruppe Peiper (1st SS Panzer Division) near
La Gleize, dropped eighty-six 250-pound General Purpose bombs
on Malmedy. At least thirty-seven American soldiers from the
120th Infantry Regiment of the 30th Infantry Division and a
considerable number of civilians lost their lives, the town
was set on fire, and a mass panic of the civilian population
was averted only with great difficulty. [28]
The assigned primary target for the bombers had been
Zulpich, the railhead for the German Seventh Army, thirty-three miles away, which, unlike the hilly and forested terrain
around Malmedy, lay in the open. Both ceiling and visibility
during the mission were unlimited, and enemy counterair
activity was minimal. The pilots realized they had not hit
Zulpich, but believed they had bombed Hammersum, a town six
miles northeast of Zulpich, and reported excellent results.
Flight cameras had operated 100 percent during the mission,
and a photo-interpreter later identified the bombed town as
Malmedy. The reason given for the mistake was personnel error.
Two days later, about 1600 on 25 December, four B-26s
of the 387th Bomb Group dropped sixty-four 250-pound General
Purpose bombs on Malmedy. Again, the pilots realized they had
made an error and believed they had bombed the town of Born
rather than the assigned target of St. Vith, and again the
photo-interpreter disclosed the actual site of the bombing.
Plane-to-ground visibility in this case was three to four
miles, and again the apparent cause of the incident was listed
as personnel error. In both cases poor navigation and
inability to properly identify the target resulted in
significant friendly military and civilian casualties.
From the beginning of close air support operations on
the Continent, the Allied air forces received mixed reviews.
Although grateful for the assistance provided by Allied
fighter-bomber aircraft in the hard fighting against the
German defenders, many ground soldiers subscribed to the
opinion of the 1st Infantry Division staff officer who made
the following entry in the division G-3 journal for 7 July
1944: "Wish you would tell the Air Corps we don't want them
over here. Have them get out in front [and] let them take
pictures [but] no strafing or bombing." [29]
The attitude changed somewhat as air-ground coordination
improved, amicicide incidents declined, and close air support
proved its value in the battles across France and Belgium and
into Germany.
During the attack on Cherbourg on 22 June 1944, ground
forces (primarily the US 9th and 79th Infantry Divisions)
undertook to mark the front lines by yellow smoke and the bomb
line, 1,500 yards in advance, by white smoke fired from
mortars. [30]
The results were uniformly poor. No yellow smoke was
visible, and every few minutes a mortar would fire two or three
shells, the smoke from which quickly dissipated in the wind and
dust of battle. [31]
The result of poor marking was inevitable. Some planes
of the British 2d TAF (RAF) and the American Ninth Air Force
attacked the wrong area and caused friendly casualties. The
proportion of the aircraft so doing was relatively small
(perhaps only ten out of 700 fighter-bombers involved) but the
small proportion did little to alleviate the pain and
destruction in the units hit. Ironically, Generals Schlatter
and Nugent from HQ AAEAF were strafed by the RAF en route to
Lt. Gen. J. Lawton Collin's VII Corps HQ to witness the air
attack. [32]
An Army Ground Forces observer, Col. Charles H. Coates,
was present on 22 June in the CP of the 314th Infantry
Regiment (79th Infantry Division), which was strafed by P-47s
and which reported casualties in the regiment's front lines as
well. Colonel Coates commented:
"The artillery was called upon to mark the bombline
with WP but the WP shells were lost in the smoke of
bombing. There may be a need for air burst smoke for
marking bomblines or bombing observation so that the
marking can be seen above the smoke of previous
bombing." [33]
The executive officer of the 2d Battalion, 314th Infantry,
Maj. D. D. Hoggsdon, also commented to Colonel Coates:
"We have got to have a better way to identify our front
lines to the air. Before we jumped off on the high
ground before Cherbourg, the smoke of bombing drifted
on us and even though we were firing WP from artillery
the air people could not see it and dropped some bombs
on my battalion." [34]
Although in general the ground forces commanders in the
Cherbourg operation were satisfied that the use of smoke as a
marker was "moderately effective and should be continued in
air support operations," the air forces were not so well
satisfied, and they remarked:
"Smoke should be continued to mark targets, friendly
boundaries, and the bombline in air support operations
until a better marking method is devised. it is
believed, however, that yellow and white smoke are the
least desirable of all colors for the purpose. Both
tend to blend with the smoke and dust of battle. The
British are using brilliant colored smokes (red, blue,
purple) which can be distinguished both from smoke and
dust and from the natural ground colors. Smoke is not
entirely satisfactory, but it is the most practicable
solution to the problem of marking found to date." [35]
The other major alternative available at the time,
using display panels, was not much more effective. Six weeks
after the Cherbourg operation the G-3 of the 4th Infantry
Division noted:
"Once our own planes started strafing our CPs and
everybody started putting out panels, then the air
never did know where the front lines were." [36]
HQ AAEAF concluded that the great majority of gross
bombing errors (those causing heavy losses to friendly troops)
in the Cherbourg operations resulted from misidentification of
the target by the bombardier due to a lack of obvious
reference points and noted that the OBOE bombing method (using
radio beacons and radar positioning) would preclude such errors. [37]
Eventually, electronic marking systems would indeed
alleviate the problem of air amicicide, but before such aids
could be fully developed, Allied ground forces would suffer
the most severe friendly air attack of all time.
The St. Lo breakthrough of 24-25 July 1944, code-named
Operation COBRA, has been characterized as "a well-planned and
successfully executed attack by combined air and ground
forces," as indeed it was when viewed in general perspective.
COBRA was the most massive close air support effort ever
attempted and was properly exploited by ground forces that
subsequently destroyed German resistance, caused their
withdrawal behind the Seine, and paved the way for the rapid
advance of Allied forces to the German border. [38] COBRA also
resulted in the most devastating incident of amicicide ever to
occur.
The detailed story of the planning and execution of
Operation COBRA and the accompanying close air support
operations is set forth by Martin Blumenson in Breakout and
Pursuit (chapters 11 and 12). The details of the
supporting air operations are also recounted in part 1-C of
The Effectiveness of Third Phase Tactical Air Operations in
the European Theater, 5 May 1944--8 May 1945.
The essence of the breakthrough plan was a massive,
short, and violent attack by medium and heavy bombers on a
rectangular target 7,000 yards wide and 2,500 yards deep
immediately to the south of the Periers-St. Lo highway. This
was the prelude for the main attack on a narrow front between
Periers and St. Lo by the US VII Corps, supported by the US
VIII Corps on its right and the US V Corps on the left, with
the US XIX Corps prepared to exploit the breakthrough. The
operation was carefully planned, and presumably adequate
provisions were made to avoid bombing friendly troops in the
course of the preparatory air bombardment. To that end,
friendly troops were to be withdrawn 1,200 yards from the
target area, while the heavy bombers would bomb no closer to
the friendly troops than 1,450 yards, the 250-yard gap to be
covered by more accurate f ighter-bombers. The relatively
straight and well defined P6riers-St. L6 highway was to form
the no-bomb line; in the absence of direct ground-to-air
communications, artillery was to mark the northern limit of
the heavy bomber target with red smoke at two-minute
intervals; the ground troops were to mark their positions
after withdrawal with identification panels; and the Allied
white star insignia on all vehicles of participating units
were to be repainted. [39]
The air operations were to be divided into three main
phases. [40]
Three divisions of heavy bombers (1,500 aircraft
from the US Eighth Air Force) were to bomb the target for one
hour from H minus 60 minutes. Ten groups of medium bombers
(396 aircraft from the 9th Bombardment Division) would then
attack specified targets within the box from H-Hour to H plus
30 minutes. Finally, fifteen groups (700 planes from the IX
and XIX Tactical Air Commands) of fighter-bombers would attack
the 250-yard strip north of the heavy bombing area in two
twenty-minute phases, one immediately before and one
immediately after the heavy bomber attacks. Although
General Omar Bradley, the First Army commander, desired that
the bomb runs be made parallel to the front lines, the Eighth
Air Force insisted that the approach be at right angles to the
target (and consequently the run-in would be over friendly
positions) to minimize German counterair and to speed passage
of the aircraft over the target area. [41]
Operation COBRA was scheduled to begin on 18 July, but
poor weather caused several postponements. The attack was
subsequently rescheduled for 1300 on 24 July, and many of the
planes were already in the air when poor visibility over the
target again caused the cancellation of the mission. Many of
the aircraft could not be recalled in time, however, and 484
of the heavy bombers and 378 of the medium bombers, as well as
the first increment of fighter-bombers, in fact attacked the
target. [42]
Not only did the abortive air attack alert the Germans
to the coming ground attack, but the results of the partial
aerial bombardment were generally poor. Only 15 percent of the
bombs from the heavies landed in the target area, and only
twenty-one of the thirty attacking medium bomber units placed
their bombs in their target areas. [43]
Results achieved by the fighter-bombers were somewhat better.
More significant, however, the confused bombing in poor
visibility resulted in several instances in which friendly
troops and equipment were destroyed. one of these incidents
proved extremely costly. One fighter-bomber pilot made a
mistake in landmark identification and inadvertently bombed an
American ammunition dump. [44]
When one of the heavy bombers was hit by a packet of
chaff, the bombardier in a reflex action hit the bomb release
toggles and dropped his bombs on the American airfield at
Chippelle, destroying two manned planes on the ground and
damaging others. [45]
The lead bombardier of another heavy bomber had
mechanical difficulty with his bomb release mechanism and
prematurely released his bombs on 30th Infantry Division
positions 2,000 yards north of the Periers-St. Lo highway, the
other fifteen planes in his group also dropping on his lead. [46]
Five medium bombers of the 9th Bombardment Division
released their bombs seven miles north of the target, also on
troops of the 30th Infantry Division. [47]
The effect of the abortive air attack and short
bombings was disastrous. The 30th Infantry Division suffered
twenty-five men killed and 131 wounded. [48]
Most of the casualties were from the 2d Battalion,
120th Infantry Regiment, which had been in the open waiting to
lead the attack. [49]
Even men in foxholes were buried by near misses or
obliterated by direct hits. Confusion reigned as commanders at
various echelons attempted to determine whether the ground
operation was to continue as planned. It did not. Both the
full aerial bombardment and the ground attack were rescheduled
for the next day.
Improved weather on 25 July permitted the preparatory
bombing and ground breakthrough to take place as planned, but
short bombing took an even heavier toll and nearly wrecked the
offensive. In three instances on 25 July heavy bombers dropped
their loads on friendly positions. [50]
In the first case a lead bombardier made a visual
release after failing to synchronize his bomb sight, and
twelve B-24s thus dropped 470 100-pound HE bombs behind
friendly lines. Eleven B-24s dropped 352 260-pound
fragmentation bombs on friendly troops when another lead
bombardier failed to identify the target properly and dropped
at the point where the bombs of a previous strike, made in
error, were seen to explode. In the final instance, a command
pilot ordered bombs away while his bombardier was still
sighting for range, in the belief that the bombing was to be
by wing rather than by group. Forty-two medium bombers of the
IX Bomber Command also failed to identify their targets
properly through the thick smoke and dropped their bombs on
friendly positions.
Again, the results of these bombing errors were
disastrous. The leading battalion of the 47th Infantry (9th
Infantry Division) and the 30th Infantry Division's 120th
Infantry Regiment and 743d Tank Battalion were particularly
hard hit, and the 92d Chemical Battalion, attached to the 30th
Infantry Division, was completely knocked out of action. [51]
The quick substitution of less damaged combat units and
the grim determination on the part of ground force commanders
and the troops permitted the planned assault to take place
with only a minimum delay. The hard hit 120th Infantry jumped
off only thirty minutes behind schedule. [52]
The 957th Field Artillery Battalion, which had nearly
thirty casualties and lost its entire Fire Direction Center
when a B-17 dropped a string of bombs through the CP area,
transferred its fire direction functions to one of its
batteries and still fired all its planned fire missions for
the day. [53]
The 30th Division alone suffered 662 casualties from
friendly bombing on 25 July: 64 killed, 374 wounded,60
missing, and 164 cases of combat fatigue induced by the
stunning effects of the heavy bombardment (for a two-day total
of 814 casualties). Unknown to most of the participants, air
or ground, was the death of Lt. Gen. Lesley J. McNair, former
commanding general of Army Ground Forces, who was killed
instantly in the bombing while observing with the assault
elements of the 2/120th Infantry. [54]
The total friendly casualties from the second day of
erroneous bombing were III killed and 490 wounded. [55]
Aside from the human errors already mentioned,
contributory factors to this frightful toll included improper
briefing on the bombline and poor visibility due to dust and
smoke that obscured reference points and the Periers-St. Lo
road, causing a parallel road three miles to the northeast to
be mistaken for the bomb line.
Despite the devastation of friendly forces, Operation
COBRA proved a resounding success. Although perhaps as few as
3 percent of the enemy troops were killed by the bombing, the
massive air attack stunned and demoralized the German
survivors and severely disrupted their defense, making
possible the successful breakthrough of Allied forces and
precipitating what eventually developed into a general
withdrawal of German forces to the Rhine. [56]
The mistaken bombing in Operation COBRA did have some
positive effects. The demonstrated inadequacies of smoke and
panels as aids for target identification in close air support
operations caused great attention to be focused on the
development of good technical methods, a highly effective
marker system was subsequently developed, and air-ground
communications improved. [57]
As a result of Operation COBRA the concept of close air
support by heavy and medium bombardment aircraft was nearly
abandoned altogether by ground force personnel. General
Eisenhower in fact swore never to use heavy bombers in close
support operations again, but later relented. [58]
Until better procedures were developed for controlling
the heavies, however, ground force commanders generally
preferred to have their close air support provided by fighter-bombers, which were not only more accurate, but boosted ground
force morale by visibly delivering their ordnance on enemy
positions. [59]
The lessons learned at so high a price in Operation
COBRA did not, however, bring about even a temporary cessation
of the use of heavy and medium bombers in the close air
support role. Several additional serious amicicide incidents
unfortunately occurred before substantial improvements in
remote bombing, troop marking, and air-ground communications
technology were forthcoming. The use of-bombers in operation
TOTALIZE (Caen ID on 7-8 August 1944 resulted in Canadian and
Polish casualties of twenty-five killed and 131 wounded,
including a division commander, when they were mistakenly
bombed by two groups of bombers. Most of the casualties were
in the 1st Polish Armored Division. [60]
The errors were attributed to: heavy enemy antiaircraft
fire during runs parallel to the front lines when only light
antiaircraft fires were anticipated; inability to see the
colored smoke markers from the air and the obscuration of
landmarks by smoke and dust from bombing; inexperienced crews;
and the continued lack of adequate procedures and equipment
for certain identification of safety lines in the target
area. [61]
During the main effort of Canadian and Polish ground
units against Falaise on 14 August 1944, more than 800 RAF and
RCAF heavy and medium bombers supported the attack by dropping
3,700 tons of bombs in the target area. Canadian and Polish
units suffered almost 400 casualties when several bomb loads
fell short of target, but the attack nevertheless advanced to
within three miles of Falaise on the first day. [62]
The reduction of Fortress Brest between 25 August and
18 September 1944 proceeded much more satisfactorily in terms
of air amicicide losses. No Allied casualties were attributed
to friendly aircraft, but the Ninth Tactical Air Force
Operational Research Section nevertheless recommended on 28
September that medium bombers be used only on specialized
targets and not in direct support of ground troops and that
Petro-gell (napalm) be used sparingly in close support
missions by fighter-bombers because of its unpredictable
trajectory characteristics. [63]
After the disastrous COBRA, TOTALIZE, and Falaise
bombings of friendly troops, great emphasis was placed on
development of an effective system of close cooperation
between ground forces, particularly tanks, and the presumably
more accurate fighter-bombers. The system that evolved relied
on the exchange of liaison officers and more efficient air-ground communication equipment and proved a vital ingredient
in the rapid advance of the US Third Army across France. [64]
But the utilization of fighter-bombers, primarily the P-47 Thunderbolt, did not eliminate attacks on friendly troops
by their own supporting aircraft. Human errors continued to
result in significant damage to the bodies, equipment, and
morale of Allied ground forces.
Indeed, in its early versions tank-fighter coordination
left much to be desired. At sunset on 9 July 1944 six medium
tanks of CCB, 3d Armored Division, fresh from a serious
encounter with an American tank destroyer outfit, were strafed
by US aircraft at Hill 91 near Hauts-Vents (Normandy). The air
strike, which had been requested earlier in the day and was
delagd by bad weather, fortunately caused no friendly
casualties. [65]
About 1900 on 26 July 1944, Company A, 1st Battalion, 66th
Armor (CCA, 2d Armored Division), attached to the regiment's
2d Battalion, was moving through the town of Canisy in an
attack toward St. Martin de Bon Fosse (Normandy) when it was
dive-bombed by an Allied P-47 and lost one tank. The incident
occurred despite the use of identification panels on the
American tanks and presumably, the use of the standard yellow
smoke signals. [66]
The veteran tankers of CCA no doubt recalled the "bad
old days" in Sicily for their newly joined comrades.
Fighter-bombers also inadvertently bombed and strafed
several units of the already badly used 30th Infantry Division
near Troisgots on 29 July. [67]
During the same division's desperate defensive battle
at Mortain in early August, supporting US P-47s and rocket-
firing RAF Typhoons often hit friendly positions, one regiment
(the 120th Infantry) being hit by friendly aircraft ten times
on 7 August alone. On the same day CCB, 3d Armored Division,
operating in support of Company I, 3d Battalion, 119th
Infantry, in the same battle, lost two tanks to Allied
dive-bombing. [68]
Even the Army Air Forces' own advanced headquarters
were not immune from air attack by Allied planes. On 15 August
American fighters strafed the headquarters of the US Third
Army and XIX Tactical Air Command' near Laval, as well as
friendly troops southwest of Carrouges. Friendly antiaircraft
gunners around Laval proved resentful and downed one fighter
from the VIII Fighter Command. [69]
The concept of close air support by heavy and medium
bombers was revived in the fall of 1944 as Allied units
approached the borders of Germany. The 2d Armored and 30th
Infantry Divisions found themselves preparing to cooperate in
the breaching of the West Wall north of Aachen in late
September and early October 1944. The 30th Infantry Division
was to cross the Wurm River three miles above Aachen and
penetrate the German defensive line. The 2d Armored Division
was then to exploit the penetration.
The operation, codenamed CISCO, was originally to be
supported by "the greatest concentration of planes in close
support of American ground troops since the 'carpet' bombing
along the Periers-St. Lo road in Normandy." [70]
To the great consternation of the 30th Division planners, the air support was to involve more than 3,300 planes, including more than 1,000 heavy bombers.
Fortunately, the heavies did not become available, and the entire air operation was greatly reduced in scale. Eventually only 360 A-20 and B-26 medium bombers of the 9th Bombardment Division and 72 P-38 and P-47 fighter-bombers of the IX Tactical Air Command were scheduled to take part.
The concern of the 30th Division planners for troop safety, greatly
heightened by their memories of the COBRA fiasco, was further increased after a flight of P-38s dropped four napalm bombs within the division lines, destroying an ammunition dump and six vehicles and killing two men and wounding four during an abortive attack on the West Wall on 22 September.
The division commander, Maj. Gen. Leland S. Hobbs, and his staff insisted that the Operation CISCO bomb runs be made over enemy territory parallel to the Wurm River and that the target areas be marked with the traditional smoke. The air planners for their part insisted on a perpendicular approach to avoid enemy flak and refused the use of smoke for fear it might obscure the target and/or friendly lines as had occurred in Normandy.
Although no friendly troops were hit, the air strikes that began at 0900 on 2 October under a scattered overcast proved a failure, in part perhaps because of excessive caution in planning and execution. The bomb runs were made perpendicular to the front lines. Many of the bombers overshot the target, but none released their bombs over friendly positions. The fighter-bombers were more accurate but had little effect. Even the napalm dropped in the woods east of Runburg was not particularly effective because the woods were wet and failed to ignite. The attack went practically unnoticed by the Germans; when interrogated later, one German prisoner even went so far as to inquire, "What bombing?"
Yet another attempt to employ heavy and medium bombardment aircraft in
close support of ground troops was made in conjunction with efforts of the US First Army to breach the Roer River line on 16 November 1944. Operation QUEEN, as the breakthrough attempt was called, was carefully planned and demonstrated the recent advances made in marking devices and other safety and technical aids. [71]
The plan called for the employment of more than 4,500 aircraft, making
Operation QUEEN the largest air attack in direct support of ground troops in World War II. More than 1,200 heavy bombers of the US Eighth Air Force, a like number of RAF heavies, 600 medium bombers from the US Ninth Air Force, 750 fighter-bombers of the IX and XXIX Tactical Air Commands, and 800 escort fighters for the heavy bombers were scheduled to participate in opening
the way for the main breakthrough attempt by the US VII Corps
in the Eschweiler area.
The earlier catastrophic bombing errors in Normandy
were clearly in the minds of the Operation QUEEN planners, and
elaborate, detailed precautions were taken to avoid a
repetition of the previous disasters. Among the aids used to
designate the target areas and the position of friendly troops
were giant panel markers (one of which was located nineteen
miles to the rear of the front line), captive balloons flown
parallel to the front line, bright-colored panels in the
immediate vicinity of frontline troops, and marking fire by 90-mm
antiaircraft guns. The Eighth Air Force also employed a system
of radio beacons close to the front lines and a radio fan
marker transmitting a thin vertical signal over the row of
balloons in addition to prescribing that bomb bay doors would
be opened and locked over the English Channel to prevent any
damage should bombs be released accidentally in the process.
The multiple safety precautions indeed prevented any
major tragedy. The new marking systems proved satisfactory,
and pilots had no trouble in identifying the locations of
friendly troops. The new safety aids demonstrated that visual
bombing by heavy and medium bombers was possible with a
2,000-yard safety zone without damage to friendly troops.
And yet Operation QUEEN was not completely flawless. A faulty
bomb release mechanism caused an Eighth Air Force heavy bomber
to drop four bombs on the 391st Armored Field Artillery
Battalion of the 3d Armored Division, killing one man and
wounding two. The same unit was later dive-bombed by a P-38, but no casualties resulted. The Ist Infantry Division
reported five incidents of stray bombs falling near its
troops, but only one human casualty resulted, although one
bomb exploded within 150 yards of the division artillery CP,
and another knocked the wings off a liaison plane and
destroyed the division artillery airfield. Such minor events
were clearly no repetition of the St. Lo-Periers Road, Caen,
or Falaise disasters.
In any event, the safety precautions proved perhaps too
restrictive. The excessive withdrawal of ground troops before
the air strike delayed the subsequent ground attack, over-bombing caused few of the enemy frontline troops to be
affected, and the wide extent of the target area somewhat
dissipated the psychological shock of the heavy bombardment.
Operation QUEEN clearly demonstrated that the effectiveness of
close air support was contingent upon the confidence of ground troops in their air support and a willingness of
commanders to accept the necessary risk of short bombing in
order to capitalize on the effects of aerial bombardment of
the enemy front lines.
Although the new and more effective procedures and
technical aids were sufficient to reduce the number of
catastrophic incidents of air amicicide by bombers, they could
not totally prevent the degradation of friendly combat power
as the result of air strikes on friendly units. Two air
amicicide incidents during the ill-fated attack of the 28th
Infantry Division on Schmidt in early November 1944
demonstrated that the supposedly more accurate fighter-bombers
were also capable of gross errors with tragic consequences for
friendly ground forces. Five fighter-bomber groups of the IX
Tactical Air Command were scheduled to isolate the battlefield
from counterattacking enemy armor. [72]
On the first day of the battle, 2 November, only one
group actually participated directly in the operation and one
of its squadrons mistakenly bombed an American artillery
position near Roetgen, killing seven men and wounding another
seventeen. [73]
On 7 November just as elements of Company B, 707th Tank
Battalion, had successfully penetrated to the far edge of the
village of Vossenack, they came under direct fire from enemy
positions on the nearby Brandenberg-Bergstein ridge, and the
tank company commander requested an air strike against the
enemy positions. Most of the twelve P-47s of the 365th
Fighter-Bomber Group assigned to conduct the strike
bombed and strafed the enemy position, but two planes
mistakenly bombed and strafed the town of Vossenack, then
occupied by friendly troops. One plane dropped two bombs and
machine-gunned the town; the other also strafed and dropped
its two bombs, one of which hit a house sheltering an American
tank crew and some of the accompanying engineers, killing one
man, seriously wounding another, and inflicting slight wounds
on three others. [74]
The maneuverability and low-level attack
characteristics of the P-47 were not proof against the human
error of its pilot, who all too often was improperly briefed
on the target and location of friendly troops and who was more
likely than not to become disoriented during the conduct of a
mission.
Brigadier General Boudinot's CCB, 3d Armored Division,
also continued to have problems with its friendly air support.
CCB began its attack in the Huertgen forest in mid-December
with sixty-four medium tanks and, in less than three days of
heavy fighting (16-18 December 1944), took its four objectives
(Werth, Koettenich, Hastenrath, and Scherpenseel). The effort,
however, cost CCB 42 of its
medium tanks plus 7 light tanks: German antitank guns
destroyed 24; Panzerfausts, 6; mines, 12; artillery fire, 6;
and one tank (2 percent) was destroyed by American bombing. [75]
The following week, during the battle between the Salm
and Ourthe rivers, the same unit suffered even more heavily at
the hands of friendly fighter pilots. Task Force McGeorge,
consisting of a company of armored infantry and a company of
medium tanks from CCB, arrived west of Grandmdnil in the early
afternoon of 25 December 1944 with the mission of restoring
blocking positions in Grandmenil. [76]
Just as the task force was beginning its attack, eleven
P-38s from the 430th Fighter Squadron, being controlled by the
7th Armored Division, mistook the Americans for Germans and
bombed their wooded assembly area, killing three officers and
thirty-six men (and presumably wounding a proportionate number
of others). The attack had to be rescheduled for 2000 that
same day, and a reconstituted TF McGeorge entered Grandmenil
only to be successfully counterattacked by the Germans. The
incident was due as much to failure in coordination between
the 3d and 7th Armored Divisions as to confusion among the
pilots. The TF McGeorge assembly area was located almost
exactly on the no-bomb line established by the 7th Armored
Division, and the display of orange panel markers on TF
McGeorge vehicles did nothing to deter the attacking P-38s.
The result, as always, was death and disruption of the
operation.
The role played by Allied air power in the desperate
Battle of the Bulge is well known. The confusion caused by the
German drive in the Ardennes and consequent American efforts
to hold their positions and restore the integrity of the lines
involved amicicide incidents of all types, including air
strikes on friendly positions. One of the few good flying days
of the period, 24 December 1944, saw several incidents. One
American officer was killed and another was wounded when a
squadron of P-38s attacked the village of Buisonville, which
had just been taken by CCA, 2d Armored Division, as part of
VII Corps's effort to blunt the German salient.
At embattled Bastogne, P-47s of the XIX Tactical Air
Command's 512th, 513th, and 514th Fighter Squadrons made good
use of the fair weather on 24 December to work around the
Bastogne perimeter. The disputed village of Marvie was hit by
P-47s during the afternoon, and the fighters bombed so close
to American lines in the Noville sector that the 101st
Airborne Division frantically signaled VIII Corps to call off
the mission.
The difficulties of controlling close air support by
heavy and medium bombers were continually reduced in the
European Theater of Operations by better planning and the
development of technical aids for blind bombing and marking of
friendly positions. Experience, better training, and improved
air-ground communications and coordination procedures also
greatly improved the amicicide record of fighter-bombers
during the course of the campaign. Despite the tragic
incidents in Normandy and the numerous occasions on which
American units were bombed or strafed by their own fighter-
bombers during the campaign in Europe, ground commanders and
soldiers alike acknowledged the high value of close air
support and, as better procedures and technical aids were
developed and experience gained, were increasingly willing to
place confidence in their air support and follow it closely.
Units that seldom asked for air strikes closer than 1,000
yards from the front lines in the early days of the campaign
later requested air strikes on targets as close as 300 yards
from their positions. [77]
Lt. Gen. J. Lawton Collins, commander of the US VII
Corps, acknowledged, "We could not possibly have gotten as far
as we did, as fast as we did, with as few casualties, without
the wonderful air support that we have consistently had." [78]
World War II: The Pacific
As was the case with the employment of artillery, the
war in the Pacific posed different, and, in some ways, more
difficult challenges to the coordination and control of close
air support operations. Distant airfields, more difficult
terrain that reduced both visibility and communications, and
the requirement to coordinate Navy and Marine Corps aircraft,
as well as Army Air Corps planes, complicated and intensified
the problem of avoiding friendly air strikes on friendly
troops. Although the combat operations in the Pacific did not
witness any major disaster such as occurred in Operation
COBRA, serious incidents involving both medium bombers and
fighter-bomber aircraft were not uncommon and served, as in
Europe, to cause suffering, loss of combat power, delay or
halting of offensive operations, and lowered morale on the
part of ground troops.
The frequent use by both Allied and Japanese forces of
small boats and coastal vessels in the island campaigns of the
Pacific war posed special problems of identification for
Allied pilots, in that such craft could be extremely difficult
to identify properly from the air. In mid-October 1942
during the Papua Campaign on New Guinea a ferry-shuttle
service for men and supplies was established by Allied forces
between Wanigela and Pongani. On its second day of operation,
18 October 1942, two newly arrived luggers were mistaken for
enemy boats and bombed by a Fifth Air Force B-25 off Pongani.
[79]
Lt. A. B. Fahnestock, a well-known explorer and yachtsman
in charge of the small boat operations for the Combined
Operational Service Command, and Byron Darnton, a New York
Times correspondent, were killed, six men were wounded, and
one of the boats was so severely damaged that it had to be
withdrawn from the shuttle service. [80]
Later investigation revealed that the Army had failed to
notify the Air Corps about the ferry service, and the pilot of
the B-25 had not bothered to insure that his target was
Japanese before attacking. [81]
Under similar circumstances three months later, on 5
January 1943, four American P-47s strafed two American LCMs
off Saidor, New Guinea, but fortunately caused no
casualties. [82]
The dense vegetation and uncertain location of friendly
troops in the Papuan jungle during the Buna campaign posed
other problems. On at least six occasions during the campaign,
Fifth Air Force planes were responsible for friendly
casualties. [83]
The assault of the 32d Infantry Division's Task Force
URBANA on the Buna Mission on the morning of 19 December 1942
received close support by B-25 and A-20 aircraft. Their
accuracy left something to be desired, however, and four bombs
landed within fifty yards of the bivouac area of the 127th
Infantry. A chaplain visiting troops in Buna Village was also
wounded by machine gun fire intended for the Japanese at
Giropa Point. [84]