[Back] [Index] [Next]

Operation Desert Storm:
Evaluation of the Air Campaign
(Letter Report, 06/12/97, GAO/NSIAD-97-134)


Appendix III

In this appendix, we respond to requester questions concerning the
effectiveness of the different types of aircraft and munitions, the
validity of manufacturer claims about weapon system performance, and
the extent to which the air campaign objectives for Desert Storm were
achieved.  We address aircraft and munition effectiveness by
answering nine questions, the first of which focuses on the quality
and scope of the weapon system performance data from the Gulf War. 
Questions 2 through 7 address the effectiveness of individual weapon
systems, and questions 8 and 9 address the combined effectiveness of
the air campaign in achieving various objectives.  The specific
questions are as follows. 

1.  Effectiveness Data Availability:  What data are available to
compare the effectiveness of the weapon systems used, and what are
the limitations of the data? 

2.  Associations Between Weapon Systems and Outcomes:  Did outcomes
achieved among strategic targets vary by type of aircraft and
munition used to attack targets? 

3.  Target Accuracy and Effectiveness as a Function of Aircraft and
Munition Type:  Did accuracy in hitting targets with LGBs vary by
type of delivery platform?  Similarly, did outcomes achieved among
strategic targets vary by platforms delivering unguided munitions? 

4.  LGB Accuracy:  Did laser-guided bombs achieve the accuracy
claimed to permit using only one per target? 

5.  F-117 Effectiveness Claims:  Did the F-117s actually achieve an
unprecedented 80-percent bomb hit rate?  Were the F-117s highly
effective against strategic air defense targets on the first night of
the campaign, thereby opening the way for more vulnerable nonstealthy
aircraft to attack? 

6.  TLAM Effectiveness Claims:  Do the data support claims for the
effectiveness of Tomahawk land-attack (cruise) missiles? 

7.  Weapon System Manufacturers' Claims:  What are the claims that
have been made by defense contractors for the effectiveness of the
weapons they produced, and do the data support these claims? 

8.  Air Campaign Effectiveness Against Mobile Targets:  What was the
effectiveness of the air campaign against small ground
targets--tanks, armored personnel carriers, and artillery? 

9.  Air Campaign Effectiveness in Achieving Strategic Objectives:  To
what extent were the overall military and political objectives of
Desert Storm met, and what was the contribution of air power? 

------------------------------------------------------- Appendix III:1

Our first subquestion is concerned with the reliability of the data
available to assess and compare the effects of the weapon systems
used in Desert Storm.  Under the best of circumstances, there would
be sufficient data on the use of aircraft, missiles, and munitions,
and on the damage inflicted on each target, to compare inputs and
outcomes comprehensively.  This would permit analysis, for example,
of whether or not an aircraft with unguided bombs is as effective as
one with LGBs or how different kinds of aircraft and munitions
performed against various targets under a range of threat and strike

However, Desert Storm was not planned, executed, or documented to
satisfy the information needs of operations analysts or program
evaluators.\1 As a result, there are sometimes significant gaps in
the data on weapon system performance and effectiveness, the latter
as a result of insufficient BDA, in particular.  For example, because
multiple aircraft of different types delivered multiple bombs, often
on the same aimpoint, and damage was often not assessed until after
multiple strikes, for most targets, it is not possible to determine
what target effects, if any, can be attributed to a particular
aircraft or particular munition. 

Making use of the best available data on both inputs and outcomes, we
compared the effectiveness of several air campaign systems both
quantitatively and qualitatively and also examined the extent to
which campaign goals were achieved.  Because specific aircraft and
munitions could not, for the most part, be identified with specific
damage to targets, we developed alternative measures of
effectiveness.  In particular, the Desert Storm data permitted us to
determine (1) the aircraft, munitions, and missiles that were
expended against the set of targets in each strategic category and
(2) the levels of damage achieved for many of the targets in most
target categories.  BDA reports indicating that restrikes were needed
provided a measure of inputs that had not fully achieved the required
results.  And when BDA reports indicated success, this was taken as
an upper-bound measure of what it took to achieve a successful

The total input measure can be compared with the prewar probability
of destruction (PD) estimates of the effectiveness of a given
munition, missile, or aircraft against a specific target type. 
Observed differences can potentially be explained by various factors
such as the effect of tactics on effectiveness, the uniqueness of
conditions encountered in Desert Storm, or the uncertainties and
risks to be considered when tasking aircraft and missiles against
specific target types. 

Our assumption is that under wartime conditions with imperfect field
information, delays in reporting BDA, communications breakdowns, and
other sources of friction, the inputs used on a target or class of
targets are likely to be the more accurate measure of future inputs
than PD calculations derived from less than fully realistic field
tests or earlier conflicts.\2 For example, the latter may indicate
that, under certain conditions, a 2,000-pound LGB has a 0.9 PD of
destroying a room inside a building with 2-feet-thick concrete walls. 
However, it may be more useful to know that, in an actual
contingency, six LGBs were used against such targets, because the
costs and risks of tasking additional pilots, aircraft, and munitions
against a target were less than the risk that the target objectives
had not been met. 

\1 While some may see this as solely a problem for postwar
evaluations, the frequent lack of timely data, such as BDA, was
repeatedly cited by Desert Storm pilots and planners as a problem
during the war. 

\2 Delivery accuracy data in the Joint Munitions Effectiveness Manual
are based in part on visual, manual system accuracies achieved in
prior combat dating as far back as World War II.  (JMEM, ch.  1, p. 
1-24, change 4.)

------------------------------------------------------- Appendix III:2

Our second subquestion is concerned with whether the degree to which
target objectives were met varied by type of aircraft or munition
used.  The available data reveal associations of greater and lesser
success against targets between types of aircraft and munitions over
the course of the campaign and with respect to individual target
categories.  However, data limitations inhibit direct comparisons
between weapon systems or generalizations about the effectiveness of
individual weapon systems. 

----------------------------------------------------- Appendix III:2.1

Data on the number of munitions, aircraft, and TLAMs used against
certain strategic targets were available, as were damage assessment
reports for 432 strategic targets with BE numbers that were attacked. 
By matching inputs to the targets for which damage assessments were
made, we examined whether any patterns could be ascertained between
the types of inputs and the outcomes. 

Using specific criteria, we rated outcomes on the strategic targets
with BE numbers for which there were sufficient phase III BDA data to
reach a judgment about whether attacks on a target had been either
"fully successful" or "not fully successful."\3 Out of 432 targets
with BDA reports, 357 could be matched with BE-numbered targets for
which campaign input data were also available.\4 For both the TLAMs
and eight air-to-ground aircraft reviewed here that delivered
ordnance against strategic targets, table III.1 shows a frequency
count, by platform, of the number of targets that we rated as damaged
to an FS or NFS level and the ratio of FS to NFS targets. 

                              Table III.1
                  Number of Targets Assessed as Fully
                 Successful and Not Fully Successful by

Platform                                        FS     NFS       ratio
------------------------------------------  ------  ------  ----------
A-6E                                            37      34       1.1:1
A-10                                            \a      \a          \a
B-52                                            25      35       0.7:1
F-111F                                          41      13       3.2:1
F-117                                          122      87       1.4:1
F-15E                                           28      29       1.0:1
F-16                                            67      45       1.5:1
F/A-18                                          36      47       0.8:1
GR-1                                            21      17       1.2:1
TLAM                                            18      16       1.1:1
Total\b                                        190     167       1.1:1
\a No data available. 

\b Individual platform data do not sum to the total because
individual targets were often attacked by multiple platforms. 

Table III.1 shows that, overall, there were more FS than NFS target
assessments and that, except for the B-52, F-15E, and F/A-18, all
platforms participated in more FS than NFS target outcomes.  The
ratio of FS to NFS target assessments was greatest for the F-111F,
indicating that it participated in proportionally more FS than NFS
target outcomes.  In addition, the ratios of FS to NFS outcomes for
the F-117 and F-16 were similar in magnitude. 

Another way in which to compare and contrast success rates among
platforms is to look at the number of FS and NFS targets with which
each delivery platform was associated across target categories. 
These comparisons are shown in table III.2. 

Table III.2 illustrates associations between individual types of
aircraft and outcomes (that is, number of FS and NFS assessments) in
various strategic target categories.  Two types of comparisons
evident in the data include the success of individual platforms
against individual target categories compared with (1) the success of
all platforms against individual target categories and (2) a
platform's success against all campaign targets. 

                                   Table III.2
                     Number of FS and NFS Targets by Platform
                                 and Target Type

Platform       FS    NFS     FS    NFS     FS    NFS     FS    NFS     FS    NFS
----------  -----  -----  -----  -----  -----  -----  -----  -----  -----  -----
A-6E            9      6      4      0     \a     \a      9      1      3      7
B-52            0      4      3      3     \a     \a      0      2      8     18
F-111F          4      0     \a     \a      0      0     11      3      5      3
F-117          49     36      0      1      9     11     21      4      9     17
F-15E           3      6      1      0     \a     \a      8      1      0      2
F-16           19     10      4      2      3      3      3      1     10     16
F/A-18          6      9      3      0     \a     \a      7      5      3      8
GR-1            0      0     \a     \a     \a     \a      7      3      2      3
TLAM            6      1      2      6      7      3      0      0      1      0
All\b          63     43     11     10     12     11     28     12     17     33

Platform             FS    NFS     FS    NFS     FS    NFS     FS    NFS     FS    NFS     FS    NFS
----------------  -----  -----  -----  -----  -----  -----  -----  -----  -----  -----  -----  -----
A-6E                  3      9      1      1      7      4      0      2      1      0      0      4
B-52                 \a     \a      1      1     10      2      2      3     \a     \a      1      2
F-111F               \a     \a      5      1     15      6     \a     \a      0      0      1      0
F-117                 0      1     14      5     13      6      0      1      5      4      2      1
F-15E                \a     \a      1      0     12      6      0      1      0      1      3     12
F-16                 \a     \a      4      2     16      5      2      1      3      0      3      5
F/A-18                1      9      1      1     10      6      1      4      3      0      1      5
GR-1                  0      0     \a     \a     10      6      1      5     \a     \a      1      0
TLAM                  0      0      0      3      1      1      1      2      0      0      0      0
All\b                 3     10     15      5     22     12      4     12     10      4      5     15
\a No records of platform tasked against target type in Missions

\b Individual platform data do not sum to category total because
individual targets were often attacked by multiple platforms. 

Success rates for individual platforms against individual categories
did not necessarily mirror the overall campaign's rate of success
against individual categories.  For example, while the overall ratio
of FS to NFS C\3 targets showed more FS relative to NFS assessments
(63:43), the ratios for the B-52, F-15E, and F/A-18 (0:4, 3:6, and
6:9, respectively) indicate that these platforms were less successful
against these types of targets than the campaign as a whole. 
However, some platforms are associated with higher rates of success
against individual categories than were achieved by the overall
campaign.  For example, the number of FS:NFS LOC targets associated
with the A-6E (9:1), F-111F (11:3), F-117 (21:4), and F-15E (8:1)
indicate higher rates of success than were achieved by the campaign
in the aggregate (28:12). 

While most platforms participated in more FS than NFS outcomes during
the campaign as a whole, some platforms participated against selected
target categories in more NFS than FS outcomes.  For example, TLAMs
participated in strikes against more NFS than FS targets in the ELE
and NBC categories, while F-117s and F-16s participated in more NFS
than FS outcomes in the MIB targets.  In contrast, while the B-52s
and the F/A-18s had more NFS relative to FS overall against OCA
targets, both platforms participated in more FS than NFS outcomes. 
In addition, the F/A-18s participated in more FS than NFS outcomes
against ELE, LOC, and SAM targets. 

The success rates for individual platforms over the course of the
campaign did not necessarily mirror the pattern of success achieved
by a platform against targets in specific categories.  For example,
while the ratio of FS:NFS for targets struck by the F-15E during the
campaign was 28:29, its association with success in the LOC and OCA
categories was proportionately far better (8:1 and 12:6,
respectively), yet its association with success in the SCU category
was worse (3:12).  In another example, the ratio of FS to NFS for
targets struck by B-52s over the course of the campaign was
relatively unfavorable (25:35); its association with success in the
OCA category was much better (10:2). 

In sum, while these data do not allow direct effectiveness
comparisons between aircraft types, they do indicate that
effectiveness did vary by type of aircraft and by type of target
category attacked.  Subsequent subquestions address more direct
aircraft effectiveness comparisons where the data permit. 

\3 An FS assessment means that the target objective had been met
sufficiently to preclude the need for a restrike.  An NFS assessment
does not equate with failure--rather, it means that despite the
damage that may have been inflicted at the time of the BDA, the
target objective had not been fully achieved and, in the opinion of
the BDA analysts, a restrike was necessary to fully achieve the
target objective.  For a more complete explanation of the strengths
and limitations of our methodology for assessing target outcomes, see
appendix I. 

\4 The Missions database contained input data on 862 BE-numbered

----------------------------------------------------- Appendix III:2.2

Another way in which the Desert Storm databases permit comparison of
inputs and outcomes is by type of munition used in each target
category.  Table III.3 shows the average amount, in tons, of guided
and unguided munitions used per BE, by target category, for both FS
and NFS targets and the ratio of unguided-to-guided bomb tonnage

Table III.3 shows that, on average, FS targets received more guided
munition tonnage (11.2 tons versus 9.4) and less unguided munition
tonnage (44.1 tons versus 53.7) per BE than NFS targets.  However,
this pattern did not hold across all target categories.  For example,
the opposite pattern occurred in the ELE, NAV, NBC, and SAM target
categories, where NFS targets generally received more guided munition
tonnage than targets rated FS, and the ratio of unguided to guided
munition tonnage was lower than for targets rated FS, as well. 

                                   Table III.3
                     Average Guided and Unguided Tonnage Per
                            BE by Outcome by Category

                                     Unguided-                         Unguided-
Target          Unguided  Guided     to-guided    Unguided  Guided     to-guided
------------  ----------  ------  ------------  ----------  ------  ------------
C\3                  7.2     3.9         1.9:1        14.7     4.0         3.6:1
ELE                 49.8     5.4         9.2:1        36.8     7.5         4.9:1
GVC                  6.7    11.2         0.6:1         4.4     9.5         0.5:1
LOC                  8.5     7.6         1.1:1        18.4     6.1         3.0:1
MIB                120.2    10.0        12.0:1       119.8     5.2        23.1:1
NAV                 17.5     1.2        14.2:1        29.0     5.2         5.6:1
NBC                 41.1    19.3         2.1:1       125.7    73.7         1.7:1
OCA                152.6    43.9         3.5:1       106.7    36.0         3.0:1
OIL                110.8     2.3        49.3:1        45.6     1.5        31.4:1
SAM                  7.2     0.8         8.8:1         1.1     4.8         0.2:1
SCU                 94.2     7.1        13.3:1        66.3     5.0        13.3:1
Total               44.1    11.2         3.9:1        53.7     9.4         5.7:1

----------------------------------------------------- Appendix III:2.3

A widespread image from Desert Storm was that of a single target
being destroyed by a single munition.  However, the data show that an
average of 55.3 tons (110,600 pounds) of bombs were expended against
each BE rated FS.  The average for BEs rated NFS was 63 tons of bombs
(126,000 pounds).\5 If the tonnage in each case was composed solely
of 2,000-pound bombs, this would have meant using, at a minimum,
nearly 56 bombs against every BE rated FS and about 63 on every NFS
target.  If the mix of munitions included smaller sizes as well, more
than 56 munitions would have been dropped on each FS target.  While
some of this tonnage almost surely reflects the fact that many
BE-numbered targets had more than one DMPI (or aimpoint), the fact
remains that the amount of tonnage used per BE (whether FS or NFS),
as well as the number of bombs that were dropped, was substantial. 

Since the exact number of DMPIs per BE is not known, we are unable to
determine whether the differences between the average tonnages
dropped on FS versus NFS targets are meaningful.  The fact that NFS
targets received more tonnage, on average, than FS targets, may
simply reflect restrikes directed at targets insufficiently damaged
by initial attacks. 

The data also show that FS targets received, on average, more tonnage
per BE of guided munitions than NFS targets (11.2 tons versus 9.4)
and less unguided tonnage per BE (approximately 44 versus 54 tons). 
Since most of the LGBs weighed from 500 to 2,000 pounds, the average
difference of 3,600 pounds of munitions is equivalent to about one
2,000-pound LGB and three 500-pound LGBs or to about seven 500-pound

\5 These data represent the total weight of bombs dropped on targets
according to the Missions database.  The database does not
consistently provide information on whether the bombs actually hit
the intended aimpoints.  Nor do these data include munitions dropped
by coalition members other than the United States and the United

------------------------------------------------------- Appendix III:3

Although the Desert Storm input and BDA data do not permit a
comprehensive aircraft-by-aircraft or munition-by-munition comparison
of effectiveness, it is possible to compare and examine the effects
of selected types of munitions and aircraft where they were used in
similar ways.  This is because the data on some systems--such as the
F-117 and F-111F--are more complete, better documented, and more
reliable than data collected on other systems.  Thus, our third
subquestion addresses the relationship between the (1) type of
delivery platform and target accuracy using LGBs and (2) type of
delivery platform and bombing effectiveness using unguided munitions. 

A major issue raised during and after Desert Storm concerns the bomb
delivery accuracy of stealthy versus conventional aircraft.  The Air
Force states that the F-117 was more accurate than any other
LGB-capable platform because its stealthiness negated the necessity
to engage in evasive defensive maneuvers in the target area, making
it easier to hold the laser spot on the target and reducing the
distance between the target and the aircraft.  In contrast,
nonstealthy aircraft are more likely to engage in defensive maneuvers
after the bombs are released--increasing the chance of losing the
laser spot, as the aircraft seeks to avoid air defense threats and
speeds away from the target.  Therefore, in LGB delivery against
fixed targets, it was argued that the type of platform did make a
difference in accuracy. 

Of all the Desert Storm strike aircraft, there were sufficient data
to compare only the F-117 to the F-111F on this dimension.\6 We
compared the reported target hit rates of the F-117 and F-111F
against 49 Desert Storm targets struck by both aircraft.\7 The 49
targets comprised primarily airfields; bridges; large military
industrial bases; and nuclear, biological, and chemical facilities. 
Table III.4 shows summary LGB strike data on the 49 targets for the
F-117 and F-111F. 

                                   Table III.4
                      F-117 and F-111F Strike Results on 49
                                 Common Targets\a

             Laser-                                        bombs
             guided bombs   Number of     Total      dropped per          Percen
Aircraft     dropped          strikes    dropped          strike  Number       t
-----------  ------------  ----------  -----------  ------------  ------  ------
F-111F       GBU-10               422      93                2.1     357      85
F-117        GBU-10               456      517               1.1     363      80
\a For this table, a strike is defined as one aircraft attacking one
target where one or more bombs were dropped.  More than one bomb can
be delivered on the same target.  More than one strike can occur on
the same sortie, which is one flight by one aircraft. 

The F-111Fs and the F-117s flew comparable numbers of bombing strikes
against the same 49 targets--422 and 456, respectively.  However, the
F-111Fs dropped more bombs than the F-117s (893 versus 517); thus,
the F-117s averaged only slightly more than 1 bomb per strike while
the F-111Fs averaged over 2 bombs.  For the F-111F, the reported
target hit rate was 85 percent, for the F-117s, 80 percent.  Thus,
despite the advantages of stealth in LGB-deliveries--for the 49
common targets for which we have data--the reported target hit rate
for the nonstealthy F-111F was greater than for the stealthy F-117. 

As noted above, the total number of F-111F bomb hits on a given
target was not recorded; a "hit" was counted if at least one bomb of
four released hit the target.  Therefore, it cannot be determined
from these data whether perhaps (1) the F-111Fs achieved a higher
reported target hit rate because they could drop more bombs on a
target than the F-117s, and therefore, the F-111Fs had a greater
number of chances of hitting the target with at least one bomb, or
(2) the F-111Fs achieved more bomb hits per target than the F-117s,
causing more damage per strike than the F-117s.\8

\6 The 48th TFW operations summary reported the outcome of each
F-111F strike mission as a hit ("Yes") or miss ("No").  The F-111Fs
dropped from one to four bombs per target, per mission.  A hit was
reported when at least one bomb struck the target.  It was not
possible to determine from the database the number of bombs that
impacted on a target reported as hit.  The F-117 database, in
contrast, reported outcome data for each bomb dropped. 

\7 Even though there are some data and methodological limitations to
this comparison (that is, aimpoints may differ; over time, the
intensity of the defenses could vary), the results on these
49 targets compare LGB results on the same targets, albeit with
limitations to the conclusions that can be drawn. 

\8 In Desert Storm, the F-111F typically carried four LGBs per
mission; the F-117 can carry a maximum of only two. 

      F-117 VERSUS F-111F TARGET
----------------------------------------------------- Appendix III:3.1

We compared the F-117 and F-111F target hit rates when using
precisely the same munitions on the same targets by analyzing only
strikes for which the same types of munitions were dropped (that is,
GBU-10 or GBU-12).\9 Table III.5 shows the number and percent of
strikes by F-117s and F-111Fs on 22 targets where only GBU-10 and
GBU-12 LGBs were dropped. 

                                   Table III.5
                      F-117 and F-111F Strike Results on 22
                      Common Targets With GBU-10 and GBU-12

          Laser-                                   Average
          guided                                     bombs
          bombs          Number of   Total     dropped per
Aircraft  dropped          strikes  dropped         strike   Number    Percent
--------  ------------  ----------  --------  ------------  --------  ----------
F-111F    GBU-10               130    285              2.2    123         95
F-117     GBU-10               212    271              1.3    167         79
The F-117s flew almost twice as many strikes with GBU-10s and GBU-12s
as the F-111F; however, the total number of GBU-10s and GBU-12s
dropped was almost identical.  Thus, the F-111Fs dropped more bombs
per strike (2.2) than the F-117s (1.3).  As with the set of 49 common
targets, the percentage of strikes where the target was reported hit
was higher for the F-111F than for the F-117, and the differential in
target accuracy was greater. 

\9 Reliability and generalizability constraints on this comparison
include the fact that the F-111F target hit data could not be
verified; a significant portion of the reported F-117 hits lacked
corroborating support or was inconsistent with other available data;
and the calculated target hit rates per mission do not necessarily
equate with bomb hit rate.  Moreover, the results apply only to
targets struck by both types of aircraft and thereby do not address
other target types where one aircraft may have performed better than
the other, such as F-111F conducting "tank-plinking" or F-117s
striking hardened bunkers in Baghdad. 

----------------------------------------------------- Appendix III:3.2

To examine whether the type of aircraft used was related to the
effectiveness of unguided bombs, we compared damage to targets
attacked with only a single type of unguided bomb.  Sixty-eight
strategic targets were attacked with the 2,000-pound MK-84 unguided
bomb and no other munition.  The available data indicate that the
platform of delivery may affect the effectiveness of the munition. 
Table III.6 shows the number of targets attacked by aircraft type and
the number and percent that were assessed as successfully destroyed. 

                              Table III.6
                Outcomes for Targets Attacked With Only
                          MK-84 Unguided Bombs

                 Targets                                  Categories
Aircraft        attacked     Number       Percent         struck\
------------  ----------  ------------  ------------  --  ------------
F-111E                 1       0             0            MIB
F-15E                  3       1             33           C\3, LOC
F-16                  34       18            53           C\3, ELE,
                                                           GVC, LOC,
                                                           MIB, NBC,
                                                           OIL, SCU
F/A-18                 7       3             43           C\3, LOC,
                                                           MIB, OIL
A-6E                   1       1            100           ELE
The two types of aircraft with the highest representation were the
F-16 and the F/A-18.\10 Of the 34 targets attacked by the F-16, 53
percent were successfully destroyed.  Forty-three percent of the
seven targets struck by the F/A-18 were fully destroyed.  However,
the differences in percentage of targets where the objectives were
successfully achieved were not statistically significant.\11

The number of target categories struck by the F-16 with MK-84s was
considerably larger than those struck by the F/A-18.  To eliminate
any bias from the range of categories struck, table III.7 presents
F-16 and F/A-18 strike results only for targets in categories common
to both. 

                              Table III.7
                Outcomes for Targets Attacked With Only
                 MK-84s Delivered by F-16s and F/A-18s

                 Targets                                  Categories
Aircraft        attacked     Number       Percent         struck\
------------  ----------  ------------  ------------  --  ------------
F-16                  23       12            52           C\3, LOC,
                                                           MIB, OIL
F/A-18                 7       3             43           C\3, LOC,
                                                           MIB, OIL
Table III.7 reveals that the F-16s appear to have been somewhat more
effective than the F/A-18s.\12 As in table III.6, the difference in
success rates was not statistically significant.  However, the ratios
of FS to NFS targets for each aircraft (12:11 for the F-16s; 3:4 for
the F/A-18s) are consistent with the ratios of FS to NFS targets
associated with these aircraft in the campaign.  (See table III.1.)
In each case, the FS to NFS ratio for the F-16s is greater than 1:1;
the ratio for the F/A-18s is less than 1:1. 

\10 With only 2 exceptions, each of the 44 targets was attacked
exclusively by a single type of aircraft.  One target was struck by
both the F-16s and F/A-18s, and a second target was struck by both
the F-16s and F-111Es. 

\11 We tested the direct comparisons between the F/A-18 and the F-16
statistically using the chi-square procedure, and we found them not
to be significant at the 0.05 level. 

\12 As noted in the discussion of table III.6, several data
limitations limit the reliability of conclusions.  These limitations
include the fact that data on Air Force aircraft in the Missions
database are more reliable than on Navy aircraft; some phase III
reports on targets may have been produced before the final strikes
occurred (with the result that damage that came after the last BDA
report would not be credited); and not all of the 68 common targets
were assessed by DIA. 

------------------------------------------------------- Appendix III:4

Videotapes of LGBs precisely traveling down ventilator shafts and
destroying targets with one strike, like those televised during and
after Desert Storm, can easily create impressions about the effect of
a single LGB on a single target, which was summed up by an LGB
manufacturer's claim for effectiveness:  "one target, one bomb."\13

The implicit assumption in this claim is that a target is
sufficiently damaged or destroyed to avoid needing to hit it again
with a second bomb, thus obviating the need to risk pilots or
aircraft in restrikes.  However, evidence from our analysis and from
DIA's does not support the claim for LGB effectiveness summarized by
"one target, one bomb."

To examine the validity of the claim, we used data from attacks on
bridges, aircraft shelters, radar sites, and bunkers of various types
with the most advanced LGBs used in Desert Storm, those with the
"Paveway III" guidance system.\14 (See table III.8.)

                              Table III.8
                     List of DMPIs and Identifying

Number      Target name         DMPI 1                   ATODAY\a
----------  ------------------  ------------------  ------------------
1           North Taji command  Fac 2                       3

2           Karbala depot,      E bnkr (1) N.               17
            ammo storage

3           Samarra CW          Bnk 1                       20

4           Samarra CW          Bnk 4                       20

5           Tallil airfield     Bnk 38 D116                 23

6           Iraqi AF hdq,       Bnk 5 OSP4                  33

7           Iraqi intel hdq,    Entrance                    36
            Ku bks

8           Al Fahud            Bridge                      38

9           Suq Ash Shuyukh     Bridge                      38

10          Pontoon bridge      None indicated              42

11          Taji bunker         Bunker                      42

12          Highway bridge      32 08 90 N                  2

13          Al Amarah           Command bunker              3

14          6 Corp Army hdq     Command bunker              14

15          Al Taqaddum         Shelter #2                  8

16          Kuwait City         Radar Site                  29

17          Al Qaim Mine        Mine entrance               32

18          Az Zubayr Radcom    Antenna                     33

19          Al Qaim phosphate   Earth covered bnkr          33

20          Ar Rumaylah Afld    Bridge S. end               36
\a ATODAY is the air tasking order day, the day of the war on which
the strike occurred. 

Source:  Missions database, January 1993. 

Each of these targets had a single, identifiable DMPI.  If the
"one-target, one-bomb" claim is accurate, there should have been a
one-to-one relationship between the number of targets and the number
of LGBs delivered to those targets.  Our data did not allow us to
determine whether one bomb typically caused sufficient damage to
preclude a restrike, and campaign managers evidently did not assume
this was the case, for the average number of LGBs dropped per target
was four.  Figure III.1 depicts the number of Paveway III LGBs that
were delivered against 20 DMPIs. 

   Figure III.1:  Paveway III LGBs
   Delivered Against Selected
   Point Targets

   (See figure in printed

Figure III.1 shows that the "one-target, one-bomb" claim for Paveway
III LGBs was not validated in a single case in this sample from
Desert Storm.  No fewer than two LGBs were dropped on each target;
six or more were dropped on 20 percent of the targets; eight or more
were dropped on 15 percent of the targets.  The average dropped was
four LGBs per target.\15

Similarly, a DIA analysis of the effectiveness of 2,000-pound
BLU-109/B (I-2000) LGBs dropped by F-117s and F-111Fs on Iraqi
hardened aircraft shelters and bunkers found that many shelters were
hit by more than one LGB, often as a result of insufficient BDA data
prior to restrike.\16 At Tallil airfield, for example, many bunkers
"were targeted with two or more weapons." (DIA, p.  28.) One bunker
was hit by at least seven LGBs, although aircraft video showed that
the required damage had been inflicted by the third and fourth bombs. 
As DIA noted, this meant that "two unnecessary restrikes using three
more weapons were apparently conducted because complete information
was not available, utilized, or properly understood/relayed." (DIA,
p.  49.) The DIA analysis also shows that one bomb was insufficient;
four bombs were required to achieve the necessary damage. 

The DIA analysis noted that the "penetration capability of a warhead
is determined by many factors:  impact velocity, impact angle, angle
of attack, target materials, and weapon design." (DIA, p.  7.) The
DIA data are consistent with our finding that targets were hit by
more than one LGB in part because more than one LGB was needed to
reach the desired damage level.  They also demonstrate that
insufficient BDA sometimes prevented knowing at what point a target
had been destroyed, thereby putting pilots and aircraft at risk in
conducting additional strikes.  Moreover, planners were apparently
ordering the delivery of multiple bombs because either BDA revealed
that one bomb did not achieve target objectives or they did not
believe the presumption that "one target, one bomb" was being

\13 This phraseology has been used by Texas Instruments, a
manufacturer of LGBs, in its public advertising. 

\14 LGBs have three component parts:  a guidance and control
mechanism, a warhead or bomb body, and airfoil or wings.  Three
generations of Paveway LGB technology exist, each successive
generation representing a change or modification in the guidance

\15 DOD commented that the types of targets in table III.8 are
primarily hardened shelters and bunkers or bridges where
probabilities of kill typically, require more than one bomb--even
with a direct hit.  We concur.  A single advanced 2,000-pound LGB was
often insufficient to achieve the desired level of damage against
high-value single-DMPI targets.  Thus, "one target, one bomb" was not
routinely achieved. 

\16 DIA, Vulnerability of Hardened Aircraft Bunkers and Shelters to
Precision-Guided Munitions (Secret), April 1994. 

------------------------------------------------------- Appendix III:5

The Air Force has written that

     "The Gulf War illustrated that the precision of modern air
     attack revolutionized warfare. 
     .  .  .  In particular, the natural partnership of smart weapons
     and stealth working together gives the attacker unprecedented
     military leverage."\17

According to a former Secretary of the Air Force, "In World War II it
could take 9,000 bombs to hit a target the size of an aircraft
shelter.  In Vietnam, 300.  Today [May 1991] we can do it with one
laser-guided munition from an F-117."\18

According to DOD's title V report, the F-117 proved to be a highly
accurate bomber with a bomb hit rate of 80 percent against its
targets--accuracy characterized by its primary contractor, Lockheed,
as "unprecedented."\19 In addition, DOD emphasized in post-Desert
Storm assessments that the F-117's stealth attributes and capability
to deliver LGBs were instrumental on the first night of the war when
the aircraft struck over 30 percent of all strategic targets,
including components of the Iraqi IADS, thereby opening major gaps in
Iraqi air defenses for conventional nonstealthy aircraft.  The Air
Force also contends that no other aircraft struck IADS and other
targets in downtown Baghdad on the first night of the campaign and
throughout the war because of the intensity of air defenses. 

It may well be that the F-117 was the most accurate platform in
Desert Storm.  However, the Desert Storm data do not fully support
claims for the F-117's accuracy against IADS-related targets, targets
on the first night of the campaign, or targets throughout the war. 
As discussed in detail below, we estimate that the bomb hit rate for
the F-117 was between 55 and 80 percent, the rate of weapon release
was 75 percent.  Thus, Desert Storm demonstrated that even in an
environment with historically favorable weather conditions, the bomb
release rate for the F-117 may be lower than for other aircraft.\20
Finally, the F-117 was not the only aircraft tasked to targets in
downtown Baghdad, but after the third day, planners concluded that
for the types of targets and defenses found in Baghdad, the F-117 was
more effective.\21

\17 USAF, Reaching Globally, Reaching Powerfully:  The United States
Air Force in the Gulf War
(Sept.  1991), p.  55. 

\18 Statement contained in a summary of public quotes and comments
about performance of the F-117A Stealth Fighter in Operation Desert
Storm provided to us by Lockheed Corporation on March 19, 1993. 

\19 In a briefing to us in September 1993, Lockheed also concluded
about the F-117 in Desert Storm "stealth, combined with precision
weapons, demonstrated a change in aerial warfare .  .  .  one bomb =
one kill."

\20 For example, historically over Baghdad, the average percentage of
time that the cloud ceiling is less than or equal to 3,000 feet is
only 9 percent; comparable percentages over Beirut, Lebanon; Osan AB,
Korea; and St.  Petersburg, Russia; are 17, 33, and 64, respectively. 
Thus, while the weather over Iraq was less favorable than average for
that location, the conditions encountered in Desert Storm may well
have been better than likely conditions in other likely contingency

\21 As discussed in appendix II, we also found that based on Air
Force intelligence analysis and other data, the defenses of the
greater Baghdad metropolitan area were as intense as those of
"downtown" Baghdad.  Multiple aircraft types were tasked to the large
area without experiencing casualties. 

----------------------------------------------------- Appendix III:5.1

Various components of DOD and GWAPS reported similar bomb hit rates
based on slightly different numbers of bomb drops and hits.  DOD's
title V report to the Congress stated that F-117s dropped 2,040 bombs
during the campaign, of which 1,634 "hit the target," achieving a
bomb hit rate of 80 percent.  (DOD, p.  T-85.) The Air Force Studies
and Analysis Group reported that the F-117s achieved an 80-percent
hit rate based on 1,659 hits.  The Air Force Office of History
reported that "Statistically, the 37th Tactical Fighter Wing compiled
a record that is unparalleled in the chronicles of air warfare:  the
Nighthawks [F-117s] achieved a 75 percent hit rate on pinpoint
targets .  .  .  recording 1,669 direct hits .  .  .  ."\22

The GWAPS report stated that "They [F-117s] scored 1,664 direct hits
.  .  .  ." and achieved a bomb hit rate of 80 percent.\23 We sought
to verify the data supporting these statements. 

\22 Office of History, Headquarters 37th Fighter Wing, Special Study: 
37FW/HO-91-1 (Jan.  9, 1992). 

\23 GWAPS, vol.  IV, pt.  I (Secret), p.  44; vol.  II, pt.  II
(Secret), p.  392. 

      F-117 HITS
----------------------------------------------------- Appendix III:5.2

During the war, mission videos of F-117 bomb releases were reviewed
after each night's strikes by analysts at the 37th TFW (and often by
planners in the Black Hole) to determine hits and misses and the need
for restrikes.  The analysts at the 37th TFW were able to determine
whether a bomb hit its intended target, or if the bomb missed, why
and by what distance.  This information was recorded on the 37th TFW
Desert Storm database, which summarized the disposition of each F-117
strike mission.  Our review of the database and interviews with F-117
pilots and the analysts who compiled the database show that some
reported hits (1) were accompanied by data indicating the "miss
distance" between the DMPI and point of bomb impact, (2) were not
based on mission video, (3) were credited when the available video
failed to record bomb impact, and (4) were accompanied by conflicting
remarks.  Our finding is that approximately one-third of the bomb
drops assessed to be hits either lacked corroborating video
documentation or were in conflict with other information in the
database.  (See table III.9.)

                              Table III.9
                      Reported F-117 Hits Lacking
                  Corroborating Support or in Conflict
                       With Other Available Data

F-117 hits                      Number                 Percent
----------------------  ----------------------  ----------------------
Total reported                  1,677                   100.0
Hits with miss                   360                     21.5
 distance data
Hits with no video                96                     5.7
Hits with video tape              69                     4.1
 recorder problems or
 impact not recorded
Hits with conflicting             49                     2.9
Total reports of hits            574
 lacking corroborating
 support or in
 conflict with other
 available data
Reported F-117 hits             535\a                    31.9
 without corroborating
 video or in conflict
 with other available
Reported F-117 hits             1,142                    68.1
 with corroborating
\a This total is less than the sum of the first four rows because, in
several instances, a reported hit was accompanied by more than one
piece of missing or incompatible data. 

--------------------------------------------------- Appendix III:5.2.1

The distance by which the bomb missed the aimpoint was recorded in
the TFW database.  For 360 of the 1,677 hits reported, the miss
distances ranged from 1.6 meters (approximately 5 feet) up to 164.5
meters (approximately 540 feet).  This range was comparable to the
range of miss distances recorded for the 70 reported misses, which
ranged from 3.2 to 178.1 meters.\24 However, while the ranges of miss
distances for hits and misses were equivalent, the distribution of
miss distances was clearly skewed toward larger values for reported
misses.  The mean miss distance for the hits was 13.1 meters (43
feet), while the mean miss distance for the misses was 69.2 meters
(226.9 feet)--five times the mean for hits.\25

\24 Paradoxically, the database contains more miss distances for
reported hits (360) than for reported misses (70).  This may be
because miss distances for misses occurring outside the field of view
of the F-117 DLIR could not be determined. 

\25 The median miss distance for the hits was 4.98 meters (16.33
feet), while the median miss distance for the misses was 66.75 meters
(218.94 feet)--13 times the average for hits. 

--------------------------------------------------- Appendix III:5.2.2

In 96 instances, hits were credited despite the absence of a video
record of the mission and in contrast to 37th TFW peacetime training
policy and the policies of other LGB-capable aircraft in Desert
Storm.  In peacetime training, bomb drops by F-117s without video
documentation are considered misses.  In Desert Storm, the 37th TFW
credited hits solely on the basis of pilot accounts; in contrast,
pilot reports were substantially discounted by Air Force analysts of
air campaign hits or kills by other types of air-to-ground aircraft
employing guided munitions but with inconclusive video.  For example,
for every three tanks claimed as kills by A-10 pilots, only one was
credited, for a 33-percent kill rate; F-111F pilots were credited
with a 50-percent tank kill rate for pilot-only claims.  The 37th TFW
justified crediting hits based solely on pilot reports on the grounds
that the F-117 demonstrated superior accuracy in Desert Storm. 

--------------------------------------------------- Appendix III:5.2.3

In 69 instances, the video recorded during a mission--from which hits
and misses are determined--was of poor quality or failed to record
bomb impact.  Poor quality video and video that did not record bomb
impact within its field-of-view pose unique BDA problems for the
F-117s.  F-117s are unique in that all missions are flown at night. 
A lone pilot must concentrate on the cockpit display to aim the laser
designator on the aimpoint until bomb impact, and the impact
typically occurs directly beneath the aircraft as it passes over the
target.  The aircraft's video records the image seen by the pilot
during the mission.  There is no other means for the pilot or BDA
analysts to view bomb impacts.  The intelligence chief for the 37th
TFW during Desert Storm told us that while to claim hits when miss
distances were small could be justified, hit claims made when
available video did not record bomb impact could not be justified. 
Table III.10 illustrates examples of remarks indicating nonsupporting

                              Table III.10
                     Examples of Remarks Indicating
                          Nonsupporting Video

Day                       BE        hits      Remarks
--------------------  ------  ----------  --  ------------------------
022                        A           2      No release on tape
006                        B           2      No impact seen, bad tape
001                        C           1      Gimbal, no impact seen
034                        D           1      Tape bad . . . , can't
                                               see impact
019                        E           1      Not on tape
Source:  37th TFW Desert Storm database. 

--------------------------------------------------- Appendix III:5.2.4

In 49 cases, credited hits were accompanied by remarks indicating
that the bombs missed the aimpoint or malfunctioned.  There was no
standing requirement that remarks be entered in the database, but the
analysts who reviewed mission video entered explanatory or clarifying
comments at their discretion.  Examples of remarks that are in
conflict with reported hits include references to dud bombs, bombs
that struck objects other than the DMPI, and bombs that did not
guide.  Table III.11 illustrates examples of remarks indicating
nonsupporting video. 

                              Table III.11
                  Examples of Remarks in Conflict With
                             Reported Hits

Day                   Target        hits      Remarks
--------------------  ------  ----------  --  ------------------------
025                        F           2      2nd bomb hit short and
011                        G           1      Dud wpn
040                        H           2      One bomb no guide
023                        I           1      Hit on wrong bunker
004                        J           1      Bomb long
Source:  37th TFW Desert Storm database. 

----------------------------------------------------- Appendix III:5.3

One of the primary reasons that reported hits are apparently in
conflict with other information recorded on the 37th TFW database is
that during Desert Storm, specific objective peacetime bomb hit
criteria were replaced with subjective wartime criteria.  According
to former 37th TFW officials, bombs making impact more than 3 feet
from a DMPI in peacetime training were considered "gross errors."
(And as noted previously, bomb drops without video were classified as
misses.) However, these officials told us that in wartime, they
deemed these criteria no longer appropriate.  In the words of one
former wing intelligence officer, "A GBU-10 striking 4 feet from a
radar will accomplish the objective of the mission." Thus, a bomb was
judged to be a hit when 37th TFW officials concluded that it probably
had an adverse effect on the enemy.  For example, if the intended
target was a specific bunker in a large ammunition storage facility
and the bomb missed the intended bunker but hit a bunker nearby, the
bomb was counted as a hit. 

In its Desert Storm white paper, the Air Force reported that campaign
planners' faith in the F-117 targeting system was so great that
pilots were tasked to hit not merely a particular building or shelter
"but a particular corner, a vent, or a door.  In fact, if they hit
the building, but not the particular spot, their sortie counted as a
miss, not a hit."\26 We conclude that the 80-percent "direct" bomb
hit rate claim is not fully justified.  The level of bomb accuracy
was clearly less than the characterization in the Air Force white
paper.  However, the subjective criteria and other data problems
prohibit us from recalculating a fully documented rate.\27

Therefore, we estimate that the F-117 bomb hit rate is likely to have
been somewhere in the interval between the upper bound asserted by
the Air Force of 80 percent and a worst-case, lower bound of
approximately 55 percent.  The lower bound assumes that all the
reported hits lacking corroborating support or in conflict with other
available data are discounted.\28 Whatever the actual bomb hit rate
for the F-117, it may well have been "unprecedented," "unparalleled,"
and higher than the rates achieved by any other aircraft in Desert
Storm; however, the data on the F-117 as well as other aircraft are
insufficient to make such characterizations. 

\26 Reaching Globally, Reaching Powerfully (1991), p.  24. 

\27 We reviewed a selective sample of mission videos in which
reported hits contained contradictory information to determine the
feasibility of verifying hit data.  We determined that hit data could
not be comprehensively verified because of (1) missing video, (2)
video records lost when tape was reused during the campaign, (3)
video images that were poor, (4) mislabeling of video, and (5) video
in which the impact image was inconclusive. 

\28 Clearly, some of the data in conflict with reported hits are more
convincing than others; we believe that it is likely that some of
these cases can be justified as functional hits.  However, some of
the evidence is equally convincing that some of the reported hits
should not have been credited (such as miss distances as great as 540
feet and hits credited when bomb impact was outside DLIR FOV).  The
data do not permit a bomb-by-bomb reassessment. 

----------------------------------------------------- Appendix III:5.4

An aircraft's bombing accuracy or bomb hit rate is one of two
essential variables that operational planners use in estimating the
probability that a given target will be damaged to the desired level
when a specific number of aircraft attack it.\29 The second variable
required by planners is the probability of weapon release.  Planners
need to know not only the accuracy of a weapon system but also the
likelihood that on a given sortie the aircraft will be able to
release its weapons.  The 37th TFW database allowed the calculation
of the probability of weapon release for the F-117 in Desert Storm. 

The probability of weapon release is a function of multiple
probabilities of potential failures during a mission that would
prevent an aircraft from arriving over a target and releasing its
weapons.  The potential aircraft failures include (1) mechanical
failure; (2) mission kill by enemy aircraft, SAM, or AAA; (3)
diversion in reaction to enemy air defenses; (4) inability to locate
the intended target; (5) inability to acquire the target in time to
effectively launch weapons; (6) inability to complete attack
coordination, and (7) inability to release weapons after arriving at
the target.  The F-117 proved more prone to some of these failures
than others.  In Desert Storm, no F-117 failed to release because of
enemy aircraft, SAMs, or AAA or because of reactions to enemy air
defenses.\30 However, F-117s did experience mechanical problems and
adverse weather.  Table III.12 presents the number of each type of
failure that resulted in aborts and prevented bombs from being
dropped on tasked F-117 strikes. 

                              Table III.12
                Failures That Prevented Bombs From Being
                   Dropped on F-117 Primary Strikes\a

Final disposition               Number                 Percent
----------------------  ----------------------  ----------------------
Total primary strikes           2,271                   100.0
Weather aborts                   412                     18.1
Air aborts                       140                     6.2
Ground aborts                     17                     0.8
Total primary strikes            569                     25.1
 where no bombs were
Total primary strikes           1,702                    74.9
 where bombs were
\a A primary strike is defined as one aircraft tasked to deliver one
or more bombs on a specific DMPI during a single sortie. 

Source:  37th TFW Desert Storm database. 

As table III.12 shows, one-quarter of all F-117 primary strikes
tasked were aborted, principally because of bad weather.\31 (As
explained in app.  II, poor weather made it difficult for F-117s to
identify and acquire targets and could prevent lasers from
illuminating targets for the bombs.) Thus, based on the Desert Storm
experience, operational planners considering the use of the F-117 in
a comparable scenario and environment would anticipate that the
expected probability of a target's being damaged to the desired level
would be based on the number of bombs tasked, reduced by the proven
probability of bomb release (75 percent), and reduced further by the
demonstrated hit rate (between 55 and 80 percent).  Therefore, in
Desert Storm, the probability of a target's receiving damage from a
scheduled F-117 strike (that is, the probability of bomb release
times the demonstrated hit rate) was between 41 and 60 percent.\32

\29 The Joint Munitions Effectiveness Manual states that damage
expectancy is determined by the probability of damage to a target
(that is, bomb hit rate) times the probability of release.  A
complete assessment of the probability that a target will receive the
desired level of damage would also need to consider the number of
aircraft sorties tasked and the appropriate selection of munition
type given the characteristics of the target. 

\30 We discussed F-117 survivability in Desert Storm in appendix II. 

\31 In contrast, according to GWAPS, 3,154 Air Force sorties were
canceled and 2,280 were aborted during Desert Storm and 69,406
sorties were flown, for a combined sortie cancellation and abort rate
of approximately 8 percent.  The GWAPS data include the range of
deployed Air Force aircraft performing the full range of service
missions.  Thus, while data are not available to compare mission
cancellation and abortion rates by strike aircraft, the available
data do indicate that the F-117 was more vulnerable to poor weather
in performing its mission than was the average Air Force aircraft. 
GWAPS, vol.  V, pt.  I (Secret) tables 76 and 174, pp.  267, 408. 

\32 DOD provided the following comment in response to this finding in
our draft report, "This statement corrects exaggerated information
(80 percent hit rate) supplied in the DOD title V report.  The
difference in the report represents confirmed and corroborated hits. 
Although statistically different, the important point is that two out
of every five bombs delivered were on target.  This represents a
quantum leap in bombing accuracy, especially when considering that
the CEP for laser guided munitions are measured in feet, not hundreds
of feet.  Aircraft without a precision guided munition (PGM)
capability could not repeatedly duplicate these results."

----------------------------------------------------- Appendix III:5.5

Lockheed, the primary contractor for the F-117, claimed after the war

     "During the first 24 hours [of the air campaign], 30 F-117s
     struck 37 high value targets, inflicting damage that collapsed
     Saddam Hussein's air defense system and all but eliminated
     Iraq's ability to wage coordinated war.  The concept of modern
     air warfare had been changed forever."\33

In April 1991, Lt.  Gen.  Horner, the Joint Force Component Commander
in Desert Storm, testified before the Congress that

     "The F-117 allowed us to do things that we could have only
     dreamed about in past conflicts.  Stealth enabled us to gain
     surprise each and every day of the war.  For example, on the
     first night of the air campaign the F-117s delivered the first
     bombs of the war against a wide array of targets, paralyzing the
     Iraqi air defense network."\34

This claim is useful in assessing F-117 performance because the first
night's missions exemplified the design mission of the aircraft:  to
strike selected high-value, well-defended targets with LGBs.  In
Desert Storm, these included the strategic air defense targets
referred to--comprising primarily SOCs, IOCs, and key C\3 elements of
the IADS. 

To assess whether the F-117s were as effective as claimed on the
first night, and specifically in contributing to the collapse of the
IADs, we addressed the following questions:  (1) What were the
reported F-117 bomb hit rates on the first night of the campaign
against all targets, and IADS-related targets in particular?  (2) Can
the damage done to IADS targets by the F-117s on the first night be
separated out from damage done by other aircraft? 

We found that the claim that the F-117s alone were crucial in
collapsing the IADS on the first night of the campaign is not fully
supported by strike, BDA, and other intelligence data.  These data
indicate that the F-117s achieved only partial strike success on the
first night; many other coalition aircraft attacked IADS-targets at
the onset of the campaign; and IADS capabilities were diminished but
continued to operate and remain viable past the first night. 

\33 Lockheed Corporation, "We Own the Night," Lockheed Horizons,
Issue 30 (May 1992), p.  57. 

\34 DOD 1992 appropriations hearings (Apr.  30, 1991). 

         F-117 HIT RATE ON PLANNED
--------------------------------------------------- Appendix III:5.5.1

We examined the F-117 database to evaluate whether it supported the
claim that the F-117s had hit all 37 targets to which they had been
tasked during the first night of the air campaign.  These data show
that only 57 percent of the targets were hit on the first night.\35

Further, approximately half of the reported bomb hits (16 of 31) did
not have corroborating documentation or were in conflict with other
available data.  (See table III.13.)

                                   Table III.13
                       37th TFW Data on Bombs Dropped by F-
                          117s During the First 24 Hours

                                                                       Hits with
                               Bombs      AC                      No        data
Target    Category     DMPIs  tasked  tasked    Hits  Misses   drops  problems\a
--------  ----------  ------  ------  ------  ------  ------  ------  ----------
A         [DELETED]        1       2       1       0       0       2           0
B         [DELETED]        1       1       1       0       0       1           0
C         [DELETED]        2       2       2       1       1       0           1
D         [DELETED]        2       2       2       1       1       0           1
E         [DELETED]        2       2       2       1       0       1           0
F         [DELETED]        1       1       1       0       1       0           0
G         [DELETED]        1       1       1       1       0       0           1
H         [DELETED]        1       1       1       1       0       0           1
I         [DELETED]        1       1       1       1       0       0           0
J         [DELETED]        2       2       2       2       0       0           1
K         [DELETED]        1       1       1       0       1       0           0
L         [DELETED]        1       1       1       0       0       1           0
M         [DELETED]        2       3       2       3       0       0           0
N         [DELETED]        1       1       1       0       0       1           0
O         [DELETED]        1       1       1       0       0       1           0
P         [DELETED]        1       1       1       0       0       1           0
Q         [DELETED]        3       4       4       4       0       0           2
R         [DELETED]        1       1       1       1       0       0           1
S         [DELETED]        1       1       1       0       1       0           0
T         [DELETED]        2       2       2       0       2       0           0
U         [DELETED]        1       1       1       0       1       0           0
V         [DELETED]        2       2       2       0       2       0           0
W         [DELETED]        1       1       1       1       0       0           0
X         [DELETED]        1       1       1       1       0       0           1
Y         [DELETED]        4       4       2       2       1       1           1
Z         [DELETED]        1       1       1       1       0       0           0
AA        [DELETED]        1       1       1       1       0       0           0
BB        [DELETED]        2       2       2       1       0       1           0
CC        [DELETED]        1       1       1       1       0       0           1
DD        [DELETED]        1       1       1       0       0       1           0
EE        [DELETED]        3       3       3       2       1       0           2
FF        [DELETED]        4       4       4       1       3       0           1
GG        [DELETED]        2       2       2       2       0       0           1
HH        [DELETED]        1       1       1       0       0       1           0
II        [DELETED]        1       1       1       0       0       1           0
JJ        [DELETED]        2       2       1       2       0       0           1
KK        [DELETED]        1       1       1       0       1       0           0
Total                     57      60      \b      31      16      13          16
\a Reported hits that lack corroborating support or are in conflict
with other available data. 

\b Column total would not equal sum of aircraft tasked because some
aircraft were tasked to more than one DMPI. 

Source:  37th TFW Desert Storm and Missions databases. 

\35 Fifty-nine percent of the tasked targets were hit on the second
night, for a two-night average of 58 percent.  Although the claim was
based only on the first night's 37 targets, we examined the data on
the second night as well, to determine if the first night's
performance was an anomaly. 

----------------------------------------------------- Appendix III:5.6

A key claim made for the F-117s is that their effectiveness in
destroying IADS targets on the first night opened up holes that
nonstealthy aircraft then used to successfully attack other targets. 
Fifteen of the 37 F-117 first-night targets were IADS-related. 
Because of weather aborts and misses, only 9 of these 15 F-117
targets (60 percent) were reported hit by the F-117s on the first
night of the campaign.  Table III.14 shows our analysis of the 37th
TFW database and DIA BDA reports. 

                                    Table III.14
                       F-117 Hit Rate on Strategic Integrated
                       Air Defense Targets on the First Night

Ta                Aircra                         Hits with
rg  DMPI   Bombs      ft         Misse      No        data                    Day\
et     s  tasked  tasked   Hits      s   drops  problems\b   Yes    No   I\c     d
--  ----  ------  ------  -----  -----  ------  ----------  ----  ----  ----  ----
A      1       2       1      0      0       2           0
C      2       2       2      1      1       0           1           X          28
H      1       1       1      1      0       0           1     X                 5
I      1       1       1      1      0       0           0           X           7
J      2       2       2      2      0       0           1           X           6
L      1       1       1      0      0       1           0
M      2       3       2      3      0       0           0                 X     2
Q      3       4       4      4      0       0           2           X           2
T      2       2       2      0      2       0           0           X           3
V      2       2       2      0      2       0           0           X           2
W      1       1       1      1      0       0           0     X                 2
GG     2       2       2      2      0       0           1           X           3
HH     1       1       1      0      0       1           0
II     1       1       1      0      0       1           0
JJ     2       2       1      2      0       0           1           X           2
To    24      27    17\e     17      5       5           7     2     8     1
\a Assessment of first phase III report issued on target. 

\b Reported hits that lack corroborating support or are in conflict
with other available data. 

\c Phase III assessment inconclusive. 

\d Day of Desert Storm on which first DIA BDA report on target was

\e Total does not equal sum of aircraft tasked; some aircraft were
assigned more than one target. 

Source:  37th TFW Desert Storm and Missions databases. 

The table shows that 17 F-117s were tasked to deliver 27 LGBs on 15
IADS-related targets with a total of 24 DMPIs.  According to the 37th
TFW database, 5 of the scheduled 27 LGBs (19 percent) were not
dropped, another 5 (19 percent) were misses; and 17 (63 percent) were
hits.  Of the 17 claimed hits, however, 7 (41 percent) either lacked
supporting video or were in conflict with other available data.  This
means that there are unambiguous data supporting hits by 10 of the 22
LGBs (45 percent) that were dropped on IADS targets.  The F-117s did
not hit 6 of the 15 (40 percent) IADS targets to which they were
tasked, 1 of which was the Air Defense Operations Center in Baghdad. 

During Desert Storm, DIA produced phase III BDA assessments on 11 of
the 15 IADS targets to which the F-117s were tasked on the first
night.  According to initial DIA BDA assessments of the IADS targets
(most of which were made by the end of day 3 of the campaign), 2 of
the 11 targets assessed were damaged sufficiently to preclude
restrikes, 8 targets remained functional and were recommended for
restrikes, and 1 could not be conclusively assessed. 

In sum, the claim that the F-117s were responsible for collapsing the
IADS on the first night appears open to question because (1) the
F-117s did not hit 40 percent of their tasked targets on the first
night and (2) of the 11 IADS-related targets attacked by F-117s and
assessed by DIA, 8 were assessed as needing additional strikes.  In
addition, the Missions database shows that 167 other platforms (such
as A-10s, F-4Gs, and F/A-18s) also struck 18 air defense-related
targets (IOCs, SOCs, and radars) on the first night. 

The lack of data on the exact degree to which most targets were
damaged, and how that might have affected total integrated
capabilities, precludes attributing greater effectiveness to the
F-117s than to other systems.  Thus, while, overall, the coalition
was able to neutralize the IADS in the early days of the war, the
data are insufficient to validate the claim that the F-117s alone
were the critical element, above all on the first night of the air

Moreover, Air Force intelligence assessments of the extent to which
the IADS was operating in the first few days of the war do not
support the assertion that the system was "collapsed" during the
first few hours of the first night.  Daily intelligence summaries
prepared during the war, called DAISUMs, characterized the IADS on
the third day of the campaign as "crippled but information is still
being passed" and "evidence of degradation of the Iraqi C\2 network
is beginning to show." The DAISUMs also described overall Iraqi
electronic warfare activity as low but radar and SAM activity in
Baghdad and KTO as heavy.  By the fifth day of the air war, the
DAISUMs described the situation as, "In general, the Iraqi IADS is
down but not out."

----------------------------------------------------- Appendix III:5.7

Related to the claim for F-117 effectiveness against IADS targets is
a broader claim made by the Air Force concerning the overall value or
survivability of stealth aircraft.  The Air Force stated in its
Desert Storm white paper that "the F-117 was the only airplane that
the planners dared risk over downtown Baghdad." The Air Force further
stated that "so dangerous was downtown Baghdad that the air campaign
planners excluded all other attackers, except F-117s and cruise
missiles, from striking it."\36 Similarly, in joint testimony to the
Congress on stealth and Desert Storm, Gens.  Horner and Glosson
stated "F-117s were the only aircraft that attacked downtown Baghdad
targets--by most accounts more heavily defended than any Eastern
Europe target at the height of the Cold War."\37 A virtually
identical claim was made by Air Combat Command's Gen.  Loh, also in
congressional testimony.\38 Contrary to these statements, however, we
found that strikes by other aircraft were not only planned but also
executed against key targets in downtown Baghdad. 

A CENTAF-prepared Master Attack Plan (MAP) identified all planned air
campaign strikes for the first 72 hours of the air war.  For the
third day of Desert Storm, the MAP called for three large-package
F-16 strikes against targets both in downtown Baghdad and against the
nearby Baghdad Nuclear Research Facility.  Forty F-16s in package G
were assigned to strike 5 leadership targets in the heart of the
city--the headquarters of Iraqi intelligence service, directorate of
internal security, military intelligence, national air force, and
Baath Party.  Another 16 F-16s in package N were assigned to restrike
military intelligence headquarters;
8 more were tasked to a sixth central city target, the Ministry of
Information and Culture.  Although planned, these attacks were
canceled because of poor weather. 

On day 3 of the campaign, the third and largest package (package Q)
included 72 F-16s; 56 were tasked against the Baghdad Nuclear
Research Facility, on the edge of the city and just 10 miles from the
presidential palace.  Eight F-16s were tasked against the Baghdad
Petroleum Refinery, across the Euphrates River from central Baghdad
and barely 2 miles from the presidential palace.  Four each were
tasked to restrike the air force and Baath Party headquarters.  These
attacks were carried out, and two F-16s in this package were lost. 

Thus, the MAP for day 3 called for a total of 152 F-16s to strike
targets within a radius of 10 miles of the presidential palace; 96
were specifically tasked to targets in the heart of the city. 
Moreover, those tasked to the nuclear research center were well
within the threat ranges of SAM and AAA sites that defended Baghdad
area targets, whether core or suburban.  And as explained in appendix
II, many types of aircraft struck targets in metropolitan Baghdad,
which was heavily defended throughout, thus making the distinction
about taskings over downtown Baghdad versus the metropolitan area
somewhat moot. 

While aircraft other than F-117s were not subsequently tasked against
downtown targets after package Q on day 3 of the campaign, many types
of bombers struck targets in the Baghdad metropolitan area repeatedly
throughout the air campaign.  And those attacks carried out at night
resulted in either zero or minimal casualties for nonstealthy,
conventional aircraft. 

\36 USAF, Reaching Globally, Reaching Powerfully (1991), p.  19. 

\37 DOD 1992 appropriations hearings (Apr.  30, 1991), p.  468. 

\38 Gen.  Loh, the "Value of Stealth," DOD 1992 appropriations
hearings (Apr.  30, 1991), p.  2.  Figure II.4 is an Air Force
depiction of the use of F-117s and F-16s against the Baghdad Nuclear
Research Facility to demonstrate the "value of stealth." Appendix XI
addresses the claim that the comparative advantage of stealth
aircraft delivering LGBs over conventional aircraft delivering
unguided bombs was demonstrated in Desert Storm when both types of
aircraft attacked the same Baghdad target. 

------------------------------------------------------- Appendix III:6

Extensive analysis of BDA imagery and other data on the effectiveness
of Tomahawk land-attack missiles by the Center for Naval Analyses has
found that TLAM performance in Desert Storm was well below the
impression conveyed in DOD's title V report to the Congress, as well
as in internal DOD estimates. 

The title V report, while essentially silent about the missile's
actual accuracy and effectiveness, notes that the "launching system
success rate was 98 percent." (DOD, p.  T-203.) CNA and DIA reported
that the Joint Chiefs of Staff estimated in April 1991 (just a couple
months after the conflict ended) that 85 percent of the TLAMs had hit
their intended targets.\39 Three variants of TLAMs were used in
Desert Storm:  TLAM Cs, with conventional unitary warheads; and TLAM
D-Is; and TLAM D-IIs, which dispense different types of conventional

\39 Joint CNA/DIA Research Memorandum 93-49, TLAM Performance During
Operation Desert Storm:  Assessment of Physical and Functional Damage
to the TLAM Aimpoints, Vol.  I:  Overview and Methodology (Secret),
March 1994, p.  21.  CNA/DIA noted that JCS assumed that TLAMs were
always responsible for all the damage at the aimpoint, even when it
had been targeted by other U.S.  weapons. 

\40 This report and the CNA/DIA reports cited do not assess the
performance of the TLAM D-IIs because of classification issues. 

----------------------------------------------------- Appendix III:6.1

During Desert Storm, a TLAM mission was loaded 307 times into a
particular missile for launch from a Navy ship or submarine.\41 Of
those 307, 19 experienced prelaunch problems.  Ten of the 19 problems
were only temporary, thus these missile were either launched at a
later time or returned to inventory.  Of the 288 actual launches, 6
suffered boost failures and did not transition to cruise.  Of the 282
missiles that transitioned to cruise, 22 were TLAM D-IIs and 260 were
TLAM Cs and D-Is. 

Of the 38 targets attacked by TLAMs, 37 were attacked by the 260 TLAM
Cs and D-Is.  The 37 targets had a total of 173 individual aimpoints;
they were aimed at 10 leadership targets:  6 C\3 targets, 3 air
defense targets, 8 electric power targets, 4 oil-related targets, 4
chemical and missile targets, and
2 airfield targets.  (The 38th target was targeted by TLAM D-IIs
alone.) However, TLAMs were limited in the type of target to which
they could be aimed, since they did not have anywhere near the "hard
target" capability of a 2,000-pound bomb.  CNA/DIA reported that
although two TLAMs hit the Baghdad air defense operations center,
they made only "small craters on the roof" of the 11-feet-thick
reinforced concrete bunker. 

\41 Some analysts may be more familiar with a lower figure of
intended launches.  However, as CNA/DIA stated, "a TLAM mission was
loaded 307 times into a particular missile for launch (i.e., there
were missile/mission pairs)." Of these, 10 missiles experienced
"temporary problems" preventing launch when intended (some were
launched later and some returned to inventory), and 9 had prelaunch
failures.  Subtracting these 19 missiles, there were 288 TLAM Desert
Storm launches at the time intended.  Since 307 missiles were
originally matched to a mission, we used that number as the universe
of TLAM launches.  (For further discussion, see CNA/DIA, vol.  I
(Mar.  1994), pp.  70-72.)

----------------------------------------------------- Appendix III:6.2

TLAM launches occurred overwhelmingly in the first 3 days of the war. 
Of the 260 TLAM Cs and D-Is that transitioned to cruise phase, more
than 39 percent were fired in the first 24 hours; 62 percent were
launched during the first 48 hours; just over 73 percent in the first
72 hours; and no TLAMs of any kind were launched after February 1,
1991, just 2 weeks after the war started.  CNA/DIA offered no
explanation for why there were no launches after February 1. 
However, CNA/DIA noted that on February 1, six TLAM Cs were fired in
a "stream raid," all aimed at the Rasheed airfield; they arrived in
the Baghdad area about 11 a.m., they were fired upon, and only two of
the six arrived at the target.  GWAPS reported that Gen.  Schwarzkopf
did not approve any additional TLAM strikes either because (1)
television coverage of daylight strikes in downtown Baghdad proved
unacceptable in Washington or (2) their use was deemed too expensive
given its relatively small warhead and high cost. 

----------------------------------------------------- Appendix III:6.3

Despite initial strong positive claims made for TLAM performance in
Desert Storm, analysis of TLAM effectiveness was complicated by
problematic BDA data.  Multiple TLAMs were targeted to the same
targets, and attacks by U.S.  Air Force bombers with other weapons
were also made against some TLAM targets before the targets could be
assessed for BDA purposes.  Thus, for many TLAMs, it was difficult to
identify the damage a particular missile may have done, or to know
whether it actually even reached the target, if the target was
scheduled for attack by other weapons before BDA collection. 

However, using BDA imagery and analysis, CNA/DIA's postwar analyses
have shown that about as many TLAM Cs and D-Is failed to arrive at
their intended targets--termed "no shows"--as are estimated to have
hit their targets.  Others arrived at the designated target area, but
impacted so far away from the aimpoint as to only create a crater. 
Of the 260 TLAM Cs and D-Is that transitioned to cruise flight, 30
were TLAM Cs with "programmed warhead detonation"--airburst
mode--that created damage effects that CNA/DIA stated could not be
evaluated adequately by existing BDA imagery.  Therefore, these 30
are excluded from CNA/DIA's assessment of the percentage of TLAMs
that arrived at the target area and that hit their intended target. 
(Since there was no way to reliably ascertain any damage caused by
the airburst mode TLAMS, it could not be determined how many arrived
over the targets either.) Ranges in the estimates for arrival and
hits reflect BDA uncertainties. 

Table III.15 shows the number of TLAMs launched and the number of
TLAM Cs and D-1s estimated by CNA/DIA to have arrived at their
targets and to have caused some damage. 

For those TLAMs for which CNA/DIA were able to interpret BDA data, an
estimated [DELETED] percent hit their intended aimpoint.  These
[DELETED] missiles represented [DELETED] percent of all 307 attempted
launchings.  If the [DELETED]-percent hit rate for the 230 detectable
TLAM Cs and D-Is was assumed to have been the case also for the 30
PWD TLAM Cs and the [DELETED] TLAM D-IIs that transitioned to cruise,
then a total of [DELETED] TLAMs would have hit their intended
targets, or [DELETED] percent of the 307 attempted launches.\42

However, actual damage to targets may well have been even less than
the [DELETED]-percent hit rate appears to imply, given that, as
CNA/DIA noted, the methodology used to define a TLAM hit was "in some
ways generous." CNA/DIA stated that a hit was defined as "damage of
any kind to the aimpoint or element containing the aimpoint."
(CNA/DIA, p.  67.) This meant, CNA/DIA explained, that "if a TLAM
impacts the dirt some distance from the target but causes even minor
fragment or blast damage to its aimpoint element, it is counted as a
hit." (CNA/DIA, p.  67.) CNA/DIA reported that there were [DELETED]
such marginal hits; if they are excluded, the TLAM hit rate was
[DELETED]-percent for nonairburst TLAMs. 

                              Table III.15
                    TLAM Performance in Desert Storm

    of                              C and D-I                 C and D-
 TLAMs  f TLAM use             All  only           All        I only
------   --------------------  ----  -------------  ---------  ---------
Missile/mission pairs         307  [DELETED]\a    [DELETED]  [DELETED]

Successful launches           282  [DELETED]      [DELETED]  [DELETED]

Transition to cruise flight    \b  [DELETED]      \b         [DELETED]

Arrived in target area\c       \b  [DELETED]      \b         [DELETED]

No shows at target\d           \b  [DELETED]      \b         [DELETED]

Hit or damaged target          \b  [DELETED]      \b         [DELETED]
\a Excludes 10 TLAMs with "temporary problems" from base used to
calculate percentages. 

\b Data not available. 

\c Excludes 30 TLAMs with programmed warhead detonation or airburst
mode that could not be assessed.  Therefore, numbers and percentages
at this line and below are based on a set of 230 non-airburst mode
TLAMs.  For further details, see CNA/DIA, TLAM Performance During
Desert Storm (Secret), March 1994, pp.  2-3. 

\d An additional [DELETED] TLAMs that arrived in their target areas
impacted at distances at least five times greater than their
predicted CEP (circular error probable)--that is, from [DELETED] from
their aimpoints.  These [DELETED] were not counted as "no shows" or
as hits. 

Source:  CNA/DIA, vol.  1 (Secret), March 1994, pp.  71-72. 

Beyond TLAM's [DELETED]-percent miss rate against intended targets,
it demonstrated additional problems.  The relatively flat,
featureless, desert terrain in the theater made it difficult for the
Defense Mapping Agency to produce usable TERCOM ingress routes, and
TLAM demonstrated limitations in range, mission planning, lethality,
and effectiveness against hard targets and targets capable of
mobility.  Specifically, CNA/DIA reported that mission failures
resulted from three issues independent of the missile and were
problems that existed before the missile was launched.  First,
mission guidance was not always clear and specific (12 TLAMs were
expended against 12 aimpoints where objectives were vague).  Second,
supporting intelligence was not always accurate (five TLAM aimpoints
were misidentified with respect to their function).  And third,
targets were not always within the capabilities of the TLAM warhead
(five aimpoints were either mobile or too hardened for the TLAM

Since the war, the Navy has developed a Block III variant of the
TLAM.  Its improvements include the use of Global Positioning System
in TLAM's guidance system.  With GPS, TLAM route planning is not
constrained by terrain features, and mission planning time is
reduced.  Some experts have expressed the concern that GPS guidance
may be vulnerable to jamming.  Thus, until system testing and
possible modifications demonstrate TLAM Block III resistance to
electronic countermeasures, it is possible that the solution to the
TERCOM limitations--GPS--may lead to a new potential
vulnerability--jamming.  Moreover, the Block III variant continues to
use the optical Digital Scene Matching Area Correlator, which has
various limitations.  [DELETED]

In sum, TLAMs were initially believed to be extremely successful in
hitting--and therefore damaging--their targets; however, subsequent
intensive analysis shows that the hit rate for 230 TLAM Cs and D-Is
was [DELETED] percent.  Moreover, a stricter definition of a "hit"
indicates a slightly lower rate of [DELETED] percent.  TLAMs were
aimed at just 38 targets, perhaps based on their limited capabilities
against reinforced targets.  While TLAMs offered a distinct
alternative to having to deliver weapons from a manned aircraft, the
data from Desert Storm suggest that there are important limitations
to their effectiveness in terms of hit rate and capability of
damaging a wide range of targets. 

\42 There were [DELETED] PWD TLAM Cs and D-IIs that transitioned to
cruise.  The range is [DELETED] percent, which is [DELETED].  Adding

------------------------------------------------------- Appendix III:7

We assessed the accuracy of statements made by various U.S. 
manufacturers about the performance of their products that played a
major role in the air campaign.  Table III.16 presents manufacturers'
statements and summarizes our finding on each product.\43

                                   Table III.16
                     Manufacturers' Statements About Product
                       Performance Compared to GAO Findings

Manufacturer  Product   Statement                    Finding
------------  --------  ---------------------------  ---------------------------
General       F-16      "No matter what the          The F-16's delivery of
Dynamics                mission, air-to-air, air-    precision air-to-ground
                        to-ground. No matter what    munitions, such as
                        the weather, day or night.   Maverick, was impaired, and
                        The F-16 is the premier      sometimes made impossible,
                        dogfighter."\a               by clouds, haze, humidity,
                                                     smoke, and dust. Only less
                                                     accurate unguided munitions
                                                     could be employed in
                                                     adverse weather using

Grumman       A-6E      "A-6s . . . [were]           The A-6E FLIR's ability to
                        detecting, identifying,      detect and identify targets
                        tracking, and destroying     was limited by clouds,
                        targets in any weather, day  haze, humidity, smoke, and
                        or night."\b                 dust; the laser
                                                     designator's ability to
                                                     track targets was similarly
                                                     limited. Only less accurate
                                                     unguided munitions could be
                                                     employed in adverse weather
                                                     using radar.

Lockheed      F-117     Achieved "80 percent direct  The hit rate was between 55
                        hits."\c                     and 80 percent; the
                                                     probability of bomb release
                                                     was only 75 percent; thus,
                                                     the probability of a hit
                                                     during a scheduled F-117
                                                     mission was between 41 and
                                                     60 percent.

                        The "only aircraft to        Other types of aircraft
                        attack heavily defended      frequently attacked targets
                        downtown Baghdad."\c         in the equally heavily
                                                     defended metropolitan area;
                                                     the Baghdad region was as
                                                     heavily defended as

                        "During the first night, 30  On the first night, 21 of
                        F-117s struck 37 high-       the 37 high-value targets
                        value targets, inflicting    to which F-117s were tasked
                        damage that collapsed        were reported hit; of
                        Saddam Hussein's air         these, the F-117s missed 40
                        defense system and all but   percent of their strategic
                        eliminated Iraq's ability    air defense targets. BDA on
                        to wage coordinated war."\d  11 of the F-117 SAD targets
                                                     confirmed only 2 complete
                                                     kills. Numerous aircraft,
                                                     other than the F-117, were
                                                     involved in suppressing the
                                                     Iraqi IADS, which did not
                                                     show a marked falloff in
                                                     aircraft kills until day 5.

                        "On Day 1 of the war, only   The 2.5-percent claim is
                        36 Stealth Fighters (less    based on a comparison of
                        than 2.5% of the             the F-117s to all deployed
                        coalition's tactical         aircraft, including those
                        assets) were in the Gulf     incapable of dropping
                        theater, yet they attacked   bombs. The F-117s
                        31% of the 17 January        represented 32 percent of
                        targets."\d                  U.S. aircraft capable of
                                                     delivering LGBs with
                                                     warheads designed to
                                                     penetrate hardened targets.
                                                     F-117s were tasked against
                                                     35 percent of the first-
                                                     day strategic targets.

                        "The F-117 reinstated the    Other nonstealthy aircraft
                        element of surprise."\c      also achieved surprise.
                                                     Stealth characteristics did
                                                     not ensure surprise for all
                                                     F-117 strikes;
                                                     modifications in tactics in
                                                     the use of support aircraft
                                                     were required.

Martin        LANTIRN   Can "locate and attack       LANTIRN can be employed
Marietta                targets at night and under   below clouds and weather;
                        other conditions of poor     however, its ability to
                        visibility using low-        find and designate targets
                        level, high speed            through clouds, haze,
                        tactics."\e                  smoke, dust, and humidity
                                                     ranges from limited to no
                                                     capacity at all.

McDonnell     F-15E     An "all weather" attack      The ability of the F-15E
Douglas                 aircraft.\f                  using LANTIRN to detect and
                                                     identify targets through
                                                     clouds, haze, humidity,
                                                     smoke, and dust was very
                                                     limited; the laser
                                                     designator's ability to
                                                     track targets was similarly
                                                     limited. Only less accurate
                                                     unguided munitions could be
                                                     employed in adverse weather
                                                     using radar.

              TLAM C/   "Can be launched . . . in    TLAM's weather limitation
              D cruise  any weather."\g              occurs not so much at the
              missile                                launch point but in the
                                                     target area where the
                                                     optical [DELETED].

                        "Incredible accuracy"; "one  From [DELETED] percent of
                        of the most accurate         the TLAMs reached their
                        weapons in the world         intended aimpoints, with
                        today."\g                    only [DELETED] percent
                                                     actually hitting the
                                                     target. It is impossible to
                                                     assess actual damage
                                                     incurred only by TLAMs.

Northrop      ALQ-135   "Proved itself by jamming    [DELETED]
              jammer    enemy threat radars"; was
              for F-    able "to function in
              15E       virtually any hostile

Texas         Paveway   "Employable" in "poor        Clouds, smoke, dust, and
Instruments   guidance  weather/visibility"          haze impose serious
              for LGBs  conditions.\h                limitations on laser
                                                     guidance by disrupting
                                                     laser beam.

                        "TI Paveway III: one         Our analysis of a selected
                        target, one bomb."\a         sample of targets found
                                                     that no single aimpoint was
                                                     struck by one LGB--the
                                                     average was 4, the maximum
                                                     was 10.

                        "LGBs accounted for only 5%  Data were not compiled that
                        of the total ordnance. But   would permit a
                        Paveway accounted for        determination of what
                        nearly 50%" of targets       percentage of targets were
                        destroyed.\a                 destroyed by any munition
\a From a company advertisement in Aviation Week and Space
Technology, (1991). 

\b Grumman Annual Report, 1991, p.  12. 

\c Lockheed briefing for GAO. 

\d From Lockheed Horizons, "We Own the Night," Issue 30 (1992), p. 
55, 57. 

\e Martin Marietta, 10-K Report to the Securities and Exchange
Commission, 1992, p.  14. 

\f McDonnell-Douglas, "Performance of MCAIR Combat Aircraft in
Operation Desert Storm," brochure. 

\g McDonnell-Douglas, "Tomahawk:  A Total Weapon System," brochure. 

\h Texas Instruments, "Paveway III:  Laser-Guided Weapons," brochure,

Table III.16 shows that each of the manufacturers made public
statements about the performance of their products in Desert Storm
that are not fully supported.  We also found that although some
manufacturers told us that they had only limited information
available to them--to the point of relying on hearsay--this did not
inhibit them from making unfounded assertions about system
performance, attempting to create favorable impressions of their
products.  Finally, while the manufacturers' claims were often
inaccurate, their assertions were not significantly different from,
nor appreciably less accurate than, many of the statements of DOD
officials and DOD reports about the same weapon systems. 

\43 We culled statements from annual reports to stockholders, "10-K"
annual reports to the federal government, and public advertisements
appearing in a major weekly publication (Aviation Week and Space

------------------------------------------------------- Appendix III:8

Over the 38 days preceding the ground campaign, approximately 37,500
strikes were conducted against Iraqi forces in kill box areas,
targeting tanks, armored personnel carriers (APC), and other tactical
vehicles.  Because there are few data on the precise number of
munitions expended or sorties flown against tanks and other vehicles,
and because it was impossible to systematically collect and compare
BDA data to assess munition hit rates, it is also impossible to know
what level of effectiveness was achieved in Desert Storm for the
various munition types used. 

Pilots reported that they had been able to destroy large numbers of
vehicles on the ground--tanks, APCs, and trucks--as well as artillery
pieces, before and during the ground campaign, especially with guided
munitions such as LGBs and Maverick missiles.  While much pilot
frustration stemmed from the use of unguided bombs from medium to
high altitudes, a number of limitations were also revealed in the use
of guided munitions. 

The Desert Storm databases do not provide data on attacks against
specific vehicles; many such attacks are subsumed as strikes against
kill boxes in the KTO.  Interviews with pilots revealed that the
effectiveness of munitions against small ground targets was
constrained both by Desert Storm altitude delivery restrictions and
the combined technical limitations of the aircraft, sensors, and
munitions used, whether guided or unguided.  At the same time,
because Iraqi KTO forces tended to remain in place through the 38
days preceding the ground campaign--and often put tanks in
recognizable formations--they were comparatively easy to identify. 

As noted in appendix II, after day 2, aircraft delivery tactics were
designed to maximize survivability--by dropping ordnance from medium
to high altitudes--rather than to maximize weapon effectiveness. 
Most pre-Desert Storm training occurred at low altitudes where bombs
are not subject to the high winds found in the gulf at high
altitudes.  It was the consensus of the Desert Storm veteran pilots
we interviewed that unguided munitions were much less accurate from
high altitude than from low. 

Pilots reported that guided munition effectiveness also decreased
somewhat from higher altitudes because (1) targets were more
difficult to designate with lasers, (2) some computer software did
not allow high-altitude bombing, and (3) the LGBs were also subject
to the effects of wind.  Depending on the missile sensors, guided
munition delivery was also degraded, if not altogether prevented at
times, by clouds, smoke, dust, haze, and even humidity. 

The difficulty in identifying and targeting vehicles and other small
ground targets, whether with guided or unguided munitions, was
reflected in the findings of postwar studies by the Army's Foreign
Science and Technology Center (FSTC) and the CIA that sought to
distinguish the relative effectiveness of the air and ground
campaigns in destroying Iraqi armor. 

FSTC and CIA both found that the attrition of armored vehicles from
guided munitions was probably less than was initially claimed for air
power.  FSTC personnel examined tanks that the Iraqis had left behind
in the KTO.\44 Of 163 tanks analyzed, 78 (48 percent) were abandoned
intact by the Iraqis or were destroyed by Iraqi demolition,
presumably to deny them to the coalition, while 85 (52 percent) had
sustained 145 hits.  Of these hits, only 28 (17 percent) were
assessed as having come from air-to-ground munitions. 

Using aerial photography, the CIA identified the number of Iraqi
tanks and APCs that did not move from areas where they were deployed
during the entire air campaign to areas where ground fighting
occurred and were therefore "destroyed or damaged during the air
campaign .  .  .  inoperable because of poor maintenance, or .  .  . 

The CIA study examined the damage done to armored vehicles of 12
Iraqi divisions, 3 of them Republican Guard divisions.  Of the 2,665
tanks deployed to those 12 divisions, the CIA estimated that 1,135
(43 percent) were destroyed by aircraft before the ground war and
1,530 (57 percent) were undamaged.  Of 2,624 APCs, 827 (32 percent)
were assessed as destroyed by aircraft; 1,797 escaped damage.  The
levels of attrition among divisions varied greatly, with the RG units
experiencing the lightest attrition, although there was substantial
variation among them as well--from 13 to 30 percent of tanks
destroyed before the ground campaign.\46

In sum, although the CIA and FSTC studies each had methodological
shortcomings, taken together, their findings suggest that while the
air campaign may have been less effective than first estimated
against these targets, it still destroyed (or rendered unusable) less
than half the Iraqi armor in the KTO. 

\44 The sample of tanks studied was not scientifically selected; it
consisted simply of those that the study participants were able to
locate and inspect. 

\45 CIA, Operation Desert Storm:  A Snapshot (Sept.  1993), last
page.  Even though some number of the vehicles were possibly
abandoned or broken down because of lack of maintenance, the study's
methodology credited all vehicles that did not move as vehicles
killed by air attack; thus, the study may have overcounted the
percentage of tanks, APCs, and artillery destroyed by air-to-ground

\46 The Hammurabi, Madinah, and Tawakalna RG divisions experienced
13, 23, and 30 percent attrition of their tanks, respectively (for an
average attrition of 21 percent).  Nine regular army armored and
mechanized divisions experienced an average tank attrition rate of 48

------------------------------------------------------- Appendix III:9

To what extent were each of the strategic objectives of the air
campaign met?  We addressed this subquestion in two parts.  First, we
reviewed the available outcome data for each category of strategic
targets as possible indicators of the campaign's effectiveness in
destroying different categories of targets.  Second, we reviewed the
available data and literature on the aggregate effectiveness of the
campaign in meeting each of the strategic objectives. 

----------------------------------------------------- Appendix III:9.1

The effectiveness of aircraft and munitions in the aggregate varied
among the strategic target sets.\47 While the attainment of strategic
objectives is determined by more than the achievement of individual
target objectives, the compilation of individual target objectives
achieved was one tool used by commanders during the war to direct the
campaign.  Table III.17 illustrates that just over half (53 percent)
of the final DIA phase III reports concluded that the target had been
destroyed or the objective had been met and no additional strikes
were required.  The percentage of targets assessed as fully destroyed
in each category ranged from a low of 25 percent in the SCU category
to a high of 76 percent in the NBC category. 

                                   Table III.17
                           Targets Categorized as Fully
                       Successfully Destroyed and Not Fully
                              Successfully Destroyed

                                          Number  Percen  Number  Percen
Target category\                              FS    t FS     NFS   t NFS   Total
----------------------------------------  ------  ------  ------  ------  ------
C\3                                           73      57      55      43     128
ELE                                           13      57      10      43      23
GVC                                           13      52      12      48      25
KBX                                           \a      \a      \a      \a      \a
LOC                                           35      67      17      33      52
MIB                                           18      31      40      69      58
NAV                                            4      29      10      71      14
NBC                                           16      76       5      24      21
OCA                                           24      65      13      35      37
OIL                                            9      38      15      62      24
SAM                                           18      69       8      31      26
SCU                                            6      25      18      75      24
Total                                        229      53     203      47     432
\a Data were not available. 

Although the rate of success varies across target categories, for
several reasons these rates do not necessarily reflect the relative
degree to which individual campaign objectives--as operationalized
through the formation of target categories--were achieved.  Desert
Storm campaign goals were not necessarily achieved through the
cumulative destruction of individual targets.  For example,
destroying x percent of all bridges does not automatically equate to
reducing the capacity of the lines of communication by x percent, for
several reasons:  the bridges destroyed may not be the most crucial
to the flow of supplies, intelligence may not have identified all of
the bridges, and the enemy may effectively respond with
countermeasures (such as pontoon bridges).  In addition, not all
targets are of equal importance.  The value in destroying a key
bridge over the Euphrates may well be higher than destroying a bridge
in Baghdad with its numerous alternative bridges. 

Another reason why the data in table III.17 must be interpreted with
caution is that the partial damage to the majority of targets
assessed as not fully successful could have contributed toward the
attainment of the overall campaign objectives.  Moreover, no
criteria, and no data, exist to determine the absolute or relative
effect of partially (or fully) damaged targets on the attainment of
campaign objectives. 

Further, table III.17 presents data only on targets for which BDA
data exist.  These targets constitute less than half of the targets
in the Missions database, and they do not necessarily represent all
of the targets in each category.  In addition, relevant targets that
should have been struck but were not on the list of strategic targets
(such as unknown Iraqi NBC targets) are not represented among the
targets in the table. 

\47 The number of targets in each strategic target set where the
target objectives had been successfully met was used as a measure of
the effectiveness of aircraft and munitions in the aggregate.  The
determination of whether the target objective had been met was based
on the final DIA phase III BDA report written on a target during the

----------------------------------------------------- Appendix III:9.2

The Desert Storm air campaign had larger goals than simply damaging
individual target.  For example, it is one thing to destroy a dozen
bridges; it is another to achieve the objective of effectively
cutting supply lines.  In this section, we examine the effectiveness
of the air campaign with regard to several broad objectives that
account for nearly all 12 of the strategic target categories shown in
table III.17.\48 Because of their limitations, the data shown in
table III.17 should be used only as supporting or partial evidence. 

We augment those success rates with information from pilots,
planners, and analysts summarized in table III.18, which compares the
Desert Storm results as reported in DOD's title V report to our

                                   Table III.18
                         Desert Storm Achievement of Key

Target set        DOD title V result              Our finding
----------------  ------------------------------  ------------------------------
IADS and          Air supremacy "attained."       Coalition rapidly achieved
airfields                                         complete control of Iraqi and
                  IADS "fragmented" within        KTO airspace, almost
                  hours; medium-and high-         uncontested by Iraqi
                  altitude sanctuary created;     aircraft.
                  however, AAA and IR SAMs
                  remained a threat to the end.   IADS fragmented over first few
                                                  days, but autonomous SAM and
                  Iraqi air force "decimated."    AAA sites and IR SAMs remained
                                                  serious threats.

                                                  Integrated threat overstated;
                                                  autonomous threat

                                                  290 of 724 fixed-wing Iraqi
                                                  aircraft destroyed, 121
                                                  escaped to Iran, and remainder
                                                  not hit; 43 percent of air
                                                  force intact and in Iraq at
                                                  end of war.

Leadership and    Leadership forced to "move      52 percent of leadership and
command,          often," reducing C\3;           57 percent of C\3 targets were
control, and      telecommunications facilities   successfully destroyed or
communications    destroyed but were often        damaged.
                                                  Despite hits on C\3 nodes,
                  Redundant and alternative       Saddam was able to communicate
                  communication facilities "were  with and direct Iraqi forces.
                  difficult to destroy."

                  Much of command structure was

Oil and           80 percent of oil-refining      Data support title V report's
electricity       capacity "damaged."             assessment.

                  National electric power grid
                  "eventually collapsed."

                  Early disruption of primary
                  sources negatively affected
                  entire war industry

Scuds             Scud facility damage "less      No known destruction of mobile
                  than previously thought."       Scud launcher.

                  Launches reduced after day 11,  Scud launches seemingly
                  with some increase in last      temporarily suppressed but
                  week and occasional large       end-of-war launches suggest
                  salvos.                         large reserve may still
                  No destruction of mobile
                  launchers confirmed; they were  Scud hunt level of effort
                  difficult to find.              overstated.

                                                  No correlation between rate of
                                                  launches and anti-Scud

Nuclear,          Nuclear facility destruction    76 percent of known NBC
biological, and   "was incomplete"; damage to     targets fully successfully
chemical          "known" nuclear facilities was  destroyed.
                  "substantial"; however,
                  nuclear program "did not        While known nuclear sites were
                  suffer as serious a setback as  severely or moderately
                  desired."                       damaged, overall program was
                                                  virtually intact because only
                  Chemical warfare program was    less than 15 percent of the
                  "seriously damaged; 75 percent  facilities were known and,
                  of production capability        therefore, attacked.

                  NBC destruction estimates
                  "suffered from incomplete
                  target set information."

                  Nuclear program virtually
                  intact; only less than 15
                  percent of the facilities hit
                  because of lack of knowledge
                  about the program.

Railroads and     Three-quarters of bridges to    67 percent of LOC targets
bridges (lines-   KTO destroyed; major food       fully successfully destroyed.
of-               shortages for frontline
communication)    forces; lines of communication  Iraqi ground forces
                  in KTO effectively              experienced some shortages
                  interdicted.                    but, overall, remained
                                                  adequately supplied up to
                                                  ground war start.

Republican Guard  Iraqi forces' overall combat    Frontline troops and equipment
and other ground  effectiveness "reduced          apparently hit hard, but
forces in the     dramatically," "significantly   morale apparently very low
KTO               degraded"; "not every           before the air campaign.
                  Republican Guard division was
                  hit equally hard."              Static tactics of Iraqi ground
                                                  forces aided targeting.
                  Those south of Basrah
                  "received less damage."         Some RG heavy armor divisions
                                                  escaped with large inventory.
                  RG forces overall less damaged
                  than frontline forces.

\48 The only strategic target category not clearly subsumed under one
of several broader sets is that of naval-related targets, including
port areas.  These targets were not a major focus of our study.  Both
DOD's title V report and GWAPS reported that the air campaign was
highly effective in eliminating Iraq's naval forces. 

--------------------------------------------------- Appendix III:9.2.1

Using DOD's definition of air supremacy, we can state that the
coalition rapidly achieved and maintained it--meaning that there was
no effective opposition to coalition aircraft from the Iraqi air
force within just a few days of the onset of the air campaign.\49

However, coalition aircraft were never safe from AAA or handheld IR
SAMs while flying at either low or medium altitude at any time during
the conflict, and actual damage to the Iraqi air force was less than
implied by the claim of air supremacy. 

The primary response of the Iraqi air force to coalition attacks and
capabilities was either to flee to Iran or to try to remain hidden in
hardened aircraft shelters or in civilian areas.  As a result, after
some initial resistance--including the likely shooting down of an
F/A-18--the Iraqi air force retreated, offering little threat to
either coalition aircraft or to coalition ground forces.  At the same
time, an estimated 290 (40 percent) of Iraq's 724 fixed-wing aircraft
were destroyed in the air or on the ground by the coalition; another
121 escaped to Iran, leaving 313 (43 percent) intact and inside Iraq
at the end of the war.  GWAPS' conclusion that the "Iraqi Air Force
was not completely destroyed by the war's end" may be an
understatement, since more fixed-wing aircraft survived than were
destroyed.\50 While the Iraqi air force never posed a serious threat
to a qualitatively and quantitatively superior coalition force, more
than enough of it survived to remain a regional threat. 

Similarly, as evidenced by pilots' accounts and low-level losses that
continued throughout the war, coalition aircraft were not able to
defeat the AAA or portable IR SAM threats because of the very large
number of these systems and the difficulty in finding such small,
mobile, nonemitting systems.  This meant that while coalition
aircraft had a high-altitude sanctuary, medium- and especially
low-altitude deliveries remained hazardous throughout the war. 

Moreover, although radar-guided SAMs accounted for almost no damage
or losses after the first week of the air war--because they were
being launched unguided--the number of launches remained quite
substantial throughout the campaign.  About 151 SAMs were launched in
the last 8 days of the air war, although only 2 resulted in loss or
damage to coalition aircraft.\51 Eleven coalition aircraft were shot
down in the last 3 days of the war, almost all at low altitudes
(either in advance of the ground war or during it), from AAA or IR
SAMs.  Of a total 86 coalition aircraft lost or damaged during the
war, 21 losses (25 percent) occurred in the last
7 days--long after air supremacy had been declared. 

\49 On January 27, 1991, Gen.  Schwarzkopf declared that coalition
air forces had achieved air supremacy.  (DOD title V report to the
Congress [Apr.  1992], pp.  124, 127, and 129.  See glossary for

\50 GWAPS, vol.  II, pt.  II (Secret), p.  156.  GWAPS also notes
that there are some questions about the exact number of aircraft;
this reflects data gaps and counting issues.  Therefore, all numbers
cited are estimates. 

\51 GWAPS, vol.  II, pt.  II (Secret), p.  140, fig.  10.  Numbers
are our estimates based on the bar charts shown in the figure. 

         CONTROL, AND
--------------------------------------------------- Appendix III:9.2.2

The effectiveness of the air war against the Iraqi "national command
authority" is less clear than for air supremacy, not least because
there is no readily quantifiable measure about what it would have
meant to "disrupt" command, control, and communications.  There are
no agreed-upon yardsticks about how many communication nodes or lines
need to have been destroyed, how much dispersion or degradation of
authority fulfills the term "disrupt," or what it means to "isolate"
Saddam from the Iraqi people or to force him to "cry uncle."

Moreover, while the kind of targets that were related to C\3 were
fairly apparent, they were also diverse--including the "AT&T
building," the presidential palace, numerous deeply buried command
bunkers, military headquarters, telecommunication switching
facilities, and so forth.  Further, even if all these had been
destroyed--and analysis of the DIA phase III messages shows that at
least 57 percent of the C\3 category and 52 percent of the GVC
were--the fact that C\3 could be and was maintained through radios
meant that C\3 was very difficult to disrupt.  In effect, the extent
of communications disruption was "unknown."\52 It is clear, however,
that the air campaign against the Iraqi leadership did not cause the
regime to collapse and thereby preclude the need for a ground

\52 GWAPS, vol.  II, pt.  II (Secret), p.  348, notes that "the
available evidence will not permit even a rough quantitative estimate
as to how much Baghdad's national telecommunications and C\3 were
disrupted by strategic air attack."

--------------------------------------------------- Appendix III:9.2.3

The attacks on electricity-related targets largely achieved their
objective of sharply reducing generated electricity but apparently
did not succeed in weakening popular support for the regime, as hoped
by air war planners.  Oil supplies were somewhat reduced by air
attacks but not enough to affect the Iraqi forces.  Table III.17
reports that 38 and 57 percent of the oil and electric facility
targets, respectively, were assessed as fully successfully destroyed. 
These data are consistent with GWAPS and title V accounts of the
damage to the oil and electricity infrastructure, which concluded
that the campaign was more successful in achieving its goals in the
electricity category than in the oil category. 

With regard to electricity, both accounts agree that attacks on
electric power plants and transformer facilities in the first 2 days
resulted in a fairly rapid reduction in generating capacity.  By
January 20, capacity had dropped from about 9,500 megawatts to about
2,500; after numerous restrikes against smaller plants, it was
eventually reduced to about 1,000 megawatts, or about 15 percent of
prewar capability.  While the lights did go off in Baghdad as well as
in much of the rest of central and southern Iraq, GWAPS found no
evidence that this negatively affected the popularity of the Hussein

GWAPS notes that damage to electric generator halls was somewhat
greater than had been planned.  While the planners had wanted only
the electrical transformers and switching systems hit, to avoid
long-term damage, the pilots, perhaps unaware of these plans, hit the
generators.  Forcing the Iraqis to rely on secondary backup power
sources was an undoubted hindrance to overall capabilities. 

With regard to oil, the air campaign focused on reducing refining
capability and destroying stored refined oil.  Iraqi oil production
was concentrated at three major refineries.  According to GWAPS, the
CIA estimated that more than 90 percent of the total Iraqi refining
capability was rendered inoperative by air strikes.  However, only
about 20 percent of the refined product storage capacity was
destroyed, perhaps because fewer than 400 sorties struck these
facilities.  Further, because Iraqi units had sufficient stocks to
last for weeks, if not months, when the ground war started, the
attacks on oil had no significant military impact on Iraqi ground

\53 GWAPS, vol.  II, pt.  II (Secret), p.  308. 

--------------------------------------------------- Appendix III:9.2.4

The overall record against mobile Scuds strongly suggests that even
under highly favorable circumstances--namely, in a condition of air
supremacy with no jamming of airborne sensors and with Scud launches
lighting up the night sky--the United States did not have the
combination of real-time detection and prosecution required to hit
portable launchers before they moved from their launch points.  There
is no confirming evidence that any mobile Scud launchers were
destroyed, and data to support the deterrent effect of the
Scud-hunting campaign are weak because the rate of firings does not
appear to have been related to the number of anti-Scud sorties. 

The launches of Scud missiles at Israel and Saudi Arabia forced a
major unplanned diversion of air resources into trying to locate and
target trucks and other vehicles being used as mobile launchers. 
Preventing these launches became an urgent mission, yet both GWAPS
and DOD title V reported that there is not a single confirmed kill of
a mobile launcher; a draft Rand analysis reached essentially the same

In 42 instances, F-15s on Scud-hunting missions were directed to an
area from which a Scud had been launched but prosecuted only 8 to the
point of delivering ordnance.  However, both GWAPS and DOD credit the
anti-Scud campaign with suppressing the number of launches after the
initial 10 days of the war.  There was a clear drop-off in Scud
launches after day 10 of the war, but an increase again starting with
day 36.  The firing rate of Scuds averaged about 5 per day for the
first 10 days--but with large daily variations--and declined to
approximately 1 per day until the last week of the war, during which
it averaged 3 per day.\55

The number of launches on a given day shows no consistent
relationship to the number of planned counter-Scud sorties.  This can
be seen from the fact that while the number of anti-Scud sorties
ranged from about 45 to 90 on days 2 through 12, the number of Scud
launches varied from 0 to 14 per day during that period. 

\54 Rand, "Technology Lessons From Desert Storm Experience:  A
Preliminary Review and Assessment," draft report (Oct.  1991), p.  3
and chart 25. 

\55 Institute for Defense Analyses, Desert Storm Campaign, P-2661
(Apr.  1992), p.  I-16. 

--------------------------------------------------- Appendix III:9.2.5

The coalition's objective was to eliminate Iraq's capabilities to
build, deploy, or launch nuclear, biological, or chemical weapons. 
The goal of eliminating Iraq's NBC capabilities was not even
approximated by the air campaign; very substantial NBC capabilities
were left untouched.  An intelligence failure to identify NBC targets
meant that the air campaign hit only a tiny fraction of the nuclear
targets and left intact vast chemical and biological weapons

While 3 nuclear-related facilities were severely or moderately
damaged by air power, these turned out to be only less than 15
percent of those identified by U.N.  inspection teams after the war. 
The United Nations identified 16 "main facilities." Moreover, some
facilities may have remained shielded from the United Nations. 
Therefore, effectiveness against this target category was probably
even less than can be estimated from damage to known sites.  The
unclassified title V report stated
(on p.  207) that the nuclear program "did not suffer as serious a
setback as was desired."

With regard to chemical warfare production facilities, DIA concluded
that by February 20, 1991, a 75-percent degradation of production and
filling facilities had been achieved.  However, it was also the case
that large stocks of chemical weapons were not destroyed:  "it took
numerous inspections and much effort after the war by U.N. 
inspectors to begin even to approach eliminating the bulk of Iraq's
chemical weapons."\57 For example, in April 1991, Iraq admitted to
the U.N.  that it still had 10,000 nerve gas warheads, 1,500
chemical-weapon bombs and shells, and 1,000 tons of nerve and mustard
gas.  Later, it conceded that it still had 150,000 chemical
munitions.  Therefore, it is readily apparent that, as with the
nuclear weapons targets, much was missed, either through lack of
target information or through ineffective attacks. 

For several years following the cease-fire, U.N.  inspection teams
were unable to find conclusive evidence that Iraq had produced
offensive biological weapons.  However, in mid-1995, in response to
U.N.  inspection commission evidence, the Iraqis admitted to
producing large quantities of two deadly agents--the bacteria that
cause botulism and anthrax--on the eve of the Gulf War.  Several
suspected production facilities were hit during the war, as were
suspected research facilities at Taji and Salman Pak.  In addition, a
number of refrigerated bunkers believed to contain biological weapons
were hit.  DOD's classified title V report stated
(on p.  224) that the biological warfare program "was damaged and its
known key research and development facilities were destroyed. 
Further, most refrigerated storage bunkers were destroyed." Whether
these constituted the entirety of Iraq's biological warfare program
is not yet known. 

\56 It is fair to note that although the air campaign was not
directly effective in destroying the vast majority of Iraq's NBC
warfare capabilities by the end of the war, the campaign was
instrumental in securing the coalition victory and motivating Saddam
Hussein to accept U.N.  resolutions and on-site inspection teams. 
Thus, the air campaign indirectly led to the achievement of this
campaign objective following the cease-fire. 

\57 GWAPS, vol.  II, pt.  II (Secret), p.  331. 

--------------------------------------------------- Appendix III:9.2.6

Destroying railroads and bridges as well as supply convoys was seen
as the key to meeting several related objectives--cutting supply
lines to the KTO to degrade and demoralize Iraqi forces and blocking
the retreat of those forces, leading to their destruction in the
ground campaign.  While large numbers of bridges, railroad lines, and
other LOC targets were destroyed by air attacks, the sheer amount of
in-place stocks, as well as the number of available transport
vehicles, apparently served to keep most of the Iraqi ground forces
adequately supplied, up to the start of the ground war.  Thus, the
goal of cutting lines of communication was only partially met. 

Table III.17 indicated that approximately two-thirds of the LOC
targets assessed were determined to be successfully destroyed.  GWAPS
and the
title V report stated that so many bridges over the Euphrates and
Tigris rivers were destroyed that supply flows were severely reduced
to frontline troops.  GWAPS stated (on p.  349) that "all important
bridges [were] destroyed"; the title V report noted that
three-fourths of the bridges from central Iraq to the KTO were
destroyed or heavily damaged.  It is estimated that attacks on LOC
targets reduced the carrying capacity of traffic on the
Baghdad-to-KTO highways from about 200,000 metric tons per day to
about one-tenth that amount by the end of the war.  In addition,
damage to railroad bridges completely cut the only rail line from
Iraq to Kuwait. 

However, GWAPS noted (on p.  371) that the Iraqis' stocks of material
in theater were so large that "by the time the ground war began, the
Iraqi army had been weakened but not 'strangled' by air interdiction
of its lines of communications." For example, at the start of the air
campaign, Iraq had 40,000 to 55,000 military cargo trucks, 190,000
commercial vehicles, and 120,000 Kuwaiti vehicles.  In addition, Iraq
had 300,000 metric tons of ammunition in dozens of locations in the
KTO; only an estimated 10 percent of this was destroyed before the
ground war.\58 The GWAPS report stated that logistic movement
difficulties within Kuwait may have resulted as much from Iraqi
ineptitude as from air attacks; the effect of the latter is
impossible to separate out.  Moreover, despite the air attacks, GWAPS
found that the Iraqi forces were adequately sustained overall
throughout the air campaign, although some units reported food

\58 GWAPS, vol.  II, pt.  II (Unclassified), p.  194. 

--------------------------------------------------- Appendix III:9.2.7

Assessments differ about the extent to which the effectiveness of the
Iraqi forces in the KTO was reduced before the ground war.  Estimates
of overall effectiveness must take into account not only the
inventory of weapons but also morale and readiness.  Moreover, not
all equipment was equally valuable, and some, such as artillery, was
potentially more lethal against an attacking force (including feared
chemical munitions) but less important than tanks for degrading Iraqi
offensive capabilities. 

The Iraqi ground forces were diverse in a number of ways:  the
better-equipped, elite Republican Guards were kept relatively far
back from the front while the lesser supplied frontline troops were
heavily composed of ethnic groups out of favor and out of power
within Iraq.  Evidence from interviews with Iraqi prisoners of war
suggests it was not just the air campaign that destroyed the
effectiveness of their ground forces:  they characterized themselves
not as "battle hardened" after
8 years of war with Iran but, rather, as "war weary." U.S.  Army
intelligence summaries of the statements of prisoners stated the

     "War weariness, harsh conditions, and lack of conviction of the
     justice of the invasion of Kuwait caused widespread desertion in
     the Iraqi Army prior to the air campaign, but in some units the
     genuine foot race north [that is, desertion] really commenced
     when the bombs began to fall."\59

In effect, the air campaign was a factor in that collapse of morale,
but it was clearly not the only cause:  the fact that the Iraqi
forces were in a preexisting state of low morale cannot be ignored. 

Another measure of the effect of air power against Iraqi ground
forces is its destruction of Iraqi equipment.  GWAPS stated that the
operations plan set a requirement that Iraqi ground forces in the KTO
were to be reduced to no more than 50-percent effectiveness by the
start of the ground war.  According to some sources, this meant a
50-percent reduction not in the number of weapons in each and every
category but, rather, in overall capabilities.  However, GWAPS stated
(on p.  203) that phase III of the air campaign had been designed to
"reduce Iraqi armor and artillery by that planned amount." The broad
objectives were not only to reduce the capability of these units to
inflict casualties but also--as the title V report states at least
three times--to "destroy" the Republican Guard. 

In effect, several competing objectives existed under the broader
umbrella of meeting the goal of reducing the Iraqi ground forces by
50 percent.  For while the commander in chief of the Central Command
ordered that attrition against Iraqi frontline forces be maximized,
this meant that fewer sorties were flown against the less-threatening
"third echelon" Republican Guard divisions, and fewer against the
Republican Guard heavy armor divisions, than against the infantry
divisions closer to the front.\60 As a result, destruction of the
three "heavy" Republican Guard divisions ("holding the bulk of all
the armor") was considerably less than that against either the
frontline forces or the Republican Guard infantry divisions.\61 All
frontline forces had been reduced to less than 50-percent
effectiveness just before the ground war, while most of the rear
units were above 75-percent effectiveness.  The consequence of the
much greater weight of effort on the front lines was that very large
numbers of Republican Guards and their armor were either not attacked
or only sporadically attacked during the air campaign.  The end
result was that many of these forces escaped back into central Iraq,
leaving some of the most formidable Iraqi forces intact. 

The CIA estimated that no more than about 30 percent of the tanks of
the three key Republican Guard "heavy" divisions were destroyed by
air power before the ground campaign.  Total tank losses by the end
of the ground war for those three heavy divisions were 50 percent,
according to the CIA, compared to an estimated 76 percent for all
Iraqi tanks in the KTO.  Our analysis of the Missions database found
that targets most closely associated with ground troops received by
far the most strikes and the most bombs and bomb tonnage compared to
other target categories.  These targets received at least nine times
more strikes, five times more bombs, and five times more bomb tonnage
than the next highest strategic target category, MIB. 

Whatever the exact cause of armor or personnel losses, the fact
remains that large numbers of Republican Guard armor were able to
avoid destruction or capture by U.S.  ground war forces.  They were
then available to Saddam for maintaining his power and to threaten
Kuwait in October 1994. 

\59 Department of the Army, "The Gulf War:  An Iraqi General
Officer's Perspective," memorandum for the record, 513th Military
Intelligence Brigade, Joint Debriefing Center (Mar.  11, 1991), p. 

\60 The title V report states that there were fewer sorties against
the rearward Republican Guard units because they were better dug in
and had better air defenses, requiring more air support and more
sorties.  The Republican Guard infantry divisions formed a "second
echelon" reserve, well behind the front lines but in front of the
heavy, armored divisions. 

\61 GWAPS, vol.  II, pt.  II (Secret), p.  161. 

------------------------------------------------------ Appendix III:10

Many claims of Desert Storm effectiveness show a pattern of
overstatement.  In this appendix, we addressed the effectiveness of
different types of aircraft and munitions used in Desert Storm and
the overall effectiveness of the air campaign in achieving its
objectives.  The Desert Storm input and BDA data did not permit a
comprehensive aircraft-by-aircraft or munition-by-munition comparison
of effectiveness; however, we were able to combine input and outcome
data to (1) reveal associations of greater and lesser success against
targets between types of aircraft and munitions and (2) examine the
effects of selected types of munitions and aircraft where they were
used in similar ways.  Thus, we were able to work within the data
constraints to examine several aspects of aircraft, munition, and
campaign effectiveness. 

While the available Desert Storm input and outcome data did not allow
direct effectiveness comparisons between all aircraft types, they did
indicate that overall effectiveness varied somewhat by type of
aircraft and more so by type of target category attacked.  The data
also revealed patterns of greater and lesser success against targets,
both between types of aircraft and munitions over the course of the
campaign and with respect to individual target categories. 

There was no consistent pattern indicating that the key to success in
target outcomes was the use of either guided or unguided munitions. 
On average, targets where objectives were successfully achieved
received more guided and fewer unguided munitions than targets where
objectives were not determined to have been fully achieved.  But in
several target categories, the reverse was true.  Nor were there
major differences in the apparent effect of platform type on strike
performance.  When attacking the same targets with LGBs, the F-111Fs
reported achieving only a slightly greater target hit rate than the
F-117s.  Similarly, there was little difference in the rates of
success achieved by F/A-18s and F-16s when delivering the MK-84
unguided munition. 

The results of our analyses did not support the claim for LGB
effectiveness summarized by "one target, one bomb." Moreover,
planners apparently ordered restrikes either because BDA revealed
that one bomb did not achieve target objectives or they did not
believe that "one target, one bomb" was being achieved. 

Desert Storm data also do not clearly support a number of major DOD
claims for the F-117.  For example, according to some, the accuracy
of the F-117 in combat may have been unprecedented; our estimates of
the bomb hit rate for the F-117 show that it was between 55 and 80
percent.  Of equal importance, the rate of weapon release for the
F-117 during Desert Storm was only 75 percent--largely because of a
weather abort rate far higher than for other strike aircraft.  Thus,
the effectiveness of scheduled F-117 strikes was between 41 and 60
percent.  And the accuracy and effectiveness of the TLAM was less
than generally perceived. 

Our analysis of manufacturers' claims revealed the same pattern of
overstatement.  All the manufacturers whose weapon systems we
reviewed made public statements about the performance of their
products in Desert Storm that the data do not fully support.  And
while the manufacturers' claims were often inaccurate, their
assertions were not significantly different from, nor appreciably
less accurate than, many of the statements of DOD officials and DOD
reports about the same systems' performance in Desert Storm. 

Finally, we found that the available quantitative and qualitative
data indicate that damage to several major sets of targets was less
complete than DOD's title V report to the Congress made clear and,
therefore, that the objectives related to these target sets were only
partially met.  The gap between what has been claimed for air power
in Desert Storm and what actually occurred was sometimes substantial. 
In effect, even under the generally favorable tactical and
environmental conditions prevalent during Desert Storm, the
effectiveness of air power was more limited than initially expected
(see app.  V) or subsequently claimed. 

In light of the favorable conditions under which the air campaign was
pursued and the technological and numerical advantages enjoyed by the
coalition, it would not have been surprising if the effectiveness of
the individual aircraft and munitions had been quite high.  However,
the commander of the U.S.  air forces clearly stated at the onset of
the war that his top priority in the air campaign was survivability. 
Conducting the war from medium and high altitudes precluded some
systems from being used in ways that would probably have maximized
their effectiveness.  At the same time, the basically flat terrain,
the attainment of air supremacy, and the dearth of Iraqi
countermeasures provided favorable delivery conditions.  Aircraft,
munitions, and campaign effectiveness, to the extent that they can be
measured, should be extrapolated only with care to another enemy in
another contingency.