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


Appendix IV

This appendix compares the costs and performance of the aircraft and
munitions used in the Desert Storm air campaign, as well as the
results from them.  Because the data collected in Desert Storm about
the performance of weapon systems contain numerous inconsistencies in
quality and quantity, they do not allow us to make a reliable
cost-effectiveness comparison of all the systems under review. 

For some aircraft, such as the F-117, there are relatively good data
about the number of sorties conducted, while for others, such as the
A-10, numerous questions remain about the most basic kind of
performance data.  For most systems, including the TLAM, there are
relatively few instances in which the effects of a particular attack
with a particular weapon on a given target can be separated out from
other attacks on the same target.  This is because BDA data often
were not collected until after several attacks had occurred. 

To approximate a measure of cost-effectiveness, we considered an
aircraft's total program unit cost; sortie cost; and Desert Storm
performance data such as survivability, sortie rate, and outcomes
achieved by target category.  Combining aircraft input and output
performance data with cost estimates permits us to present as
comprehensive a comparison as possible of the multiple weapon systems
used in the air campaign. 

-------------------------------------------------------- Appendix IV:1

------------------------------------------------------ Appendix IV:1.1

The following measures assisted us in our comparative evaluation of
the aircraft under review.  Dollar costs are in constant fiscal year
1994 dollars. 

---------------------------------------------------- Appendix IV:1.1.1

This measure includes research and development and procurement costs
identified in DOD's periodic Selected Acquisition Reports to the
Congress, to permit a comparison of aircraft per unit costs. 

---------------------------------------------------- Appendix IV:1.1.2

This is the cost to operate each type of aircraft under review on a
typical sortie.  These estimates of comparative costs were generated
by the Air Force at our request, using Air Force and Navy data and an
agreed-upon methodology. 

---------------------------------------------------- Appendix IV:1.1.3

This measure was derived by dividing total sorties for each aircraft
under review by the 43 days of the air campaign and by the number of
aircraft deployed.  Since these averages were clearly dependent upon
multiple factors--such as distance to target, which can vary greatly
for identical Navy aircraft on different carriers or identical Air
Force aircraft at different bases--there are various factors that can
explain differences between aircraft on this measure.\1 However, it
is a summary measure of overall aircraft availability and, as such,
permits an understanding of the range of the comparative availability
of each aircraft to perform its assigned mission at its own
particular level of effectiveness, which can vary by type while not
showing the explanation for differences.  Availability is commonly
regarded as advantageous, since it is assumed that it is better to be
able to attack the enemy more rather than less in a given time

\1 Other factors can include, aircraft reliability and
maintainability, mission planning requirements, aircrew fatigue and
availability, and ability or inability to operate out of forward
operating bases; all can vary by aircraft type. 

---------------------------------------------------- Appendix IV:1.1.4

This measure permits a comparison of the survivability of aircraft,
derived by dividing total sorties of each aircraft type by total lost
and damaged aircraft of each type. 

---------------------------------------------------- Appendix IV:1.1.5

This performance measure presents the type and number of munitions
delivered, by aircraft type, on all target categories.

---------------------------------------------------- Appendix IV:1.1.6

This performance measure compares each aircraft's delivery of
munitions, as measured by total Desert Storm tonnage and average
tonnage per day per aircraft.  The assumption is that, given a
specific munition type, it is advantageous to deliver more rather
than less tonnage per day against an enemy.  This measure is, of
course, complicated by variance in the type of munitions that
different aircraft types deliver.  Thus, it is also necessary to
review the effect that the various aircraft types had on targets with
their different munition combinations. 

---------------------------------------------------- Appendix IV:1.1.7

This measure compares aircraft on their capability to operate in two
stressful environmental conditions:  conducting combat flight
operations at night and in adverse weather.  First, we indicate
whether an aircraft was used for both day and night strikes in Desert
Storm (versus day or night only).  Second, we indicate whether an
aircraft had the capability to deliver munitions effectively under
adverse weather conditions.  We did not have sufficient data to know
whether pilots chose to release bombs in poor weather regardless of
accuracy degradation. 

---------------------------------------------------- Appendix IV:1.1.8

The strategic target categories we measure accounted for
three-quarters or more of an aircraft's Desert Storm strikes.  By
eliminating those categories for which only a comparatively small
number of aircraft strikes were performed, we obtained an overall
assessment of what target categories an aircraft was used most often
against.  This, we believe, is somewhat more useful and informative
than simply tallying up the gross number of target categories an
aircraft was used against, even if only a handful of strikes were
flown in some categories.  This latter methodology was used by the
Air Force and DOD in descriptions of the F-117s' contribution to the
air campaign. 

---------------------------------------------------- Appendix IV:1.1.9

Using the data discussed in appendix III, we compared the various
aircraft on the overall ratio of targets they attacked that were, or
were not, assessed as successfully destroyed.  At best, these ratios
reflect assessments of the level of success associated with the
various aircraft, though not necessarily exclusively attributed to

------------------------------------------------------ Appendix IV:1.2

Numerous measures could be used in comparing Desert Storm air
campaign systems, such as aircraft mission capable rates or aircraft
range.  We chose the measures that, in our view, offered the most
useful ways in which to compare systems used in Desert Storm, again
taking into account data availability and limitations.  Thus, for
example, rather than comparing mission capable rates, we compared
sortie rates actually flown in the air campaign:  we believe it more
informative to measure that a combat sortie was actually flown than
that an aircraft was determined "mission capable" yet may not have
actually flown a combat mission.  Similarly, aircraft range was not
compared because the availability of tanker aircraft in Desert Storm
tended to mask differences between aircraft on this dimension. 
(However, if fewer tankers are available in future conflicts, range
differences among aircraft could have a significant effect on

Finally, it is important to emphasize that no single measure should
automatically be given greater weight than others in assessments of
aircraft or munitions.  The comparison we proffer here intentionally
presents multiple dimensions on which to assess air campaign systems,
not least because of the data reliability problems already discussed. 
Further, aircraft have different missions, and effectiveness on one
type of mission may have been achieved through design requirements
that greatly limit performance on other missions.  Therefore, no one
single cost or performance measure will consistently capture all that
should be known or understood in comparing one aircraft type to

------------------------------------------------------ Appendix IV:1.3

Table IV.1 presents cost and performance data for the aircraft under

The following appear to be the major points that can be drawn with
regard to the issue of Desert Storm aircraft cost and performance. 
Comparatively, none of the air-to-ground aircraft examined
demonstrated overall consistently superior performance across the
measurable performance indicators.  Similarly, no aircraft performed
consistently poorly on all or most of these dimensions. 

Neither single-role bombers, nor multirole fighter-bombers
demonstrated obvious superiority compared to others in the
air-to-ground role.  Defensive air-to-air missions were predominantly
performed by single-role air-to-air aircraft, with single-role F-15Cs
credited with over 85 percent of Desert Storm air-to-air kills. 
While multirole aircraft did perform some support and some air-to-air
missions, their participation by no means eliminated the need for
single-role air-to-air and support aircraft.  The evidence from
Desert Storm points to the usefulness of single-role aircraft in
their respective missions and the usefulness of multirole aircraft
most predominantly in the air-to-ground mission. 

The data in table IV.1 reveal no clear link between the cost of
either aircraft or weapon system and their performance in Desert
Storm.  Neither relatively high-cost nor low-cost air-to-ground
aircraft demonstrated consistently superior performance across a
range of measures such as sortie rate, survivability, amount of
munitions delivered, and participation in successful target outcomes. 

                                    Table IV.1
                      Cost and Performance of Major U.S. and
                     U.K. Desert Storm Air-to-Ground Aircraft
                                     and TLAM

        Total                           Rate of lost
        program            Average per   and damaged
Platfo  unit     Sortie\       day per  aircraft per    Number   Number
rm      cost\a      b         aircraft        sortie    guided  unguided   Ratio
------  -------  -------  ------------  ------------  --------  --------  ------
F-117   $111.2\   $15.7            0.7             0     2,000     4       500:1
F-      $68.3     $24.9            0.9        0.0011     2,935    586        5:1
F-15E   $39.1     $11.5            1.0        0.0009     1,669   14,089      1:8
         with 2
         N pods
A-6E    $39.3     $27.8            1.1        0.0031       623   17,588     1:28
F-16    $18.9     $5.9             1.2        0.0006       159   38,438    1:242
         with 1
         N pod
F/A-    $35.9     $17.2            1.2        0.0022       368   11,179     1:30
GR-1    $32 -      \g              0.9        0.0076       497   1,346       1:3
A-10    $11.8      \g            1.4\h        0.0023     4,801     \g         \g
B-52    $163.8\    \g              0.6        0.0029      36\k   71,885   1:1,19
         j                                                                     6
TLAM    $2.85    $2,855.            \g   [DELETED]\l       297     0          \i

                             Tonnage                    Predominan
                                 per                    t target
               Total        aircraft      Strike        categories     FS:NFS
Platform      tonnage        per day      conditions    \c            ratio\d
----------  ------------  ----------  --  ------------  ----------  ------------
F-117          1,990            1.10      Night only;   C\3, LOC,      1.4:1
                                          no weather    MIB,
                                          capability    NBC, OCA

F-111F         2,004            0.71      Night only;   KBX, LOC,      3.2:1
                                          limited by    OCA

F-15E          5,593            2.71      Mostly        KBX, OCA,      1.0:1
                                          night; very   SCU
                                          few day
                                          limited by

A-6E           5,715            1.16      Mostly        KBX, NAV       1.1:1
                                          night; some
                                          day; limited
                                          by weather

F-16           20,866           1.93      Mostly day;   KBX, OCA       1.5:1
                                          some night;
                                          limited by

F/A-18         5,513            0.74      Day and       C\3, KBX,      0.8:1
                                          night;        OCA
                                          limited by

GR-1           1,090            0.38      Day and       KBX, OCA,      1.2:1
                                          night;        OIL
                                          limited by

A-10             \g               \g      Mostly day;   \KBX             \g
                                          some night;
                                          no weather

B-52           25,422           8.69      Mostly        KBX, MIB       0.7:1
                                          night; some
                                          day; no

TLAM            144           3.30\m      Day and       C\3, ELE,      1.1:1
                                          night;        GVC, NBC,
                                          limited by    SCU
\a In millions of fiscal 1994 dollars. 

\b In thousands of fiscal 1994 dollars.  Generic aircraft sortie
costs, not specific Desert Storm sortie costs.  Total program unit
cost and sortie cost for the TLAM are the same because a combat
sortie for the TLAM requires the physical destruction of the missile. 

\c Target categories in which approximately three-quarters of all
strikes by aircraft type were directed. 

\d Based on the analysis in appendix III and summarized in table

\e Lockheed data expressed in "then-year" dollars.  DOD data exist
but are classified at the "special access required" level.  Because
the specific "then-year" dollars were not identified by year and
amount, we were unable to convert them to fiscal 1994 dollars.  Even
though this figure understates the cost of the F-117, it is the best
figure we could obtain. 

\f Estimated costs obtained from public sources. 

\g Data were not available. 

\h A-10 sorties may have been undercounted; thus, 1.4 may be too low. 

\i Data were not applicable. 

\j Includes $6.8 billion in acquisition costs for 102 aircraft and $9
billion in modifications since the B-52H was deployed in the early
1960s.  A portion of the $9 billion spent on modifications were for
upgrades of its strategic-nuclear capability or upgrades subsequently
superseded by other modifications.  The data we received from the Air
Force did not specifically identify those modification costs relevant
to the B-52s as used in Desert Storm.  Also, the total program unit
costs attributed to other aircraft could be understated somewhat in
comparison to the B-52.  Cost data for all aircraft other than the
B-52 were obtained from Selected Acquisition Reports.  However, these
reports, which include modification costs, are no longer issued after
airframe production ceases.  Thus, modification costs for
out-of-production aircraft are not captured on these reports and are
not reflected in table IV.1.  Therefore, the costs cited here tend to
overstate the B-52's cost relative to the other aircraft. 

\k The B-52 launched 35 CALCMs (conventional variants of the
air-launched cruise missile). 

\l TLAM losses are based on a study by CNA/DIA that found that of the
230 TLAM C and D-I models launched, an estimated [DELETED] did not
arrive at their target areas. 

\m Tonnage per day for TLAMs is its total tonnage (144 tons) divided
by the number of days in the entire air campaign (43). 

Virtually every type of aircraft and the TLAM demonstrated both
significant strengths and limitations.  For example, no F-117s were
lost or damaged; it was the platform of choice among planners for
nighttime strikes against stationary, point targets, yet it was
employable in only highly limited conditions.  The much older,
nonstealthy F-111F achieved a somewhat higher target hit rate than
the F-117 against targets attacked by both with the same type of
munition (although the F-111F expended more munitions per target). 
The low-cost A-10s and F-16s made large contributions in terms of
missions flown and bomb tonnage delivered and performed as well on
other measures, such as survivability rates.  However, neither was
equipped to deliver LGBs, and the F-16's potential effectiveness with
unguided munitions was diminished by operating from medium and high
altitudes.  B-52s delivered much more tonnage individually and as a
force than any other aircraft, but accuracy from high altitude was

Similarly, the F-16s delivered about 21,000 tons of bombs, but this
worked out to only 1.93 tons daily per aircraft, compared to 2.71
tons for the F-15Es; F-15Es, however, accounted for only about
one-quarter as much total tonnage as the F-16s.  Thus, on one
performance measure, the F-15Es look better than the F-16s, but much
less impressive on another.  In addition, the F-15Es had sortie costs
about double those of the F-16s but also delivered a much greater
ratio of guided-to-unguided munitions (1 to 8 versus 1 to 242).  This
was a result, in part, of the command decision to assign the
available LANTIRN targeting pods, and thus the ability to deliver
LGBs, to F-15Es rather than to F-16s; it was also a result, in part,
of the decision to assign all but a few Maverick missiles to A-10s
rather than to the F-16 units that were trained to employ them. 

A comparably mixed picture can be seen for all the other aircraft
under review.  Overall, therefore, the data in table IV.1 present an
inconclusive picture when it comes to rank-ordering the costs and
performance of the aircraft as they were used in Desert Storm. 

\2 See Operation Desert Storm:  Limits on the Role and Performance of
B-52 Bombers in Conventional Conflicts (GAO/NSIAD-93-138, May 12,

------------------------------------------------------ Appendix IV:1.4

To facilitate comparative assessment of the aircraft, we examined the
extent to which the data above, in combination with data discussed in
appendixes II and III, can address four questions that involve
aircraft acquisition issues of concern to the Congress: 

1.  Did the F-117 stealth bomber differ in air-to-ground combat
performance and effectiveness from nonstealth aircraft, and what was
the contribution of stealth technology to the outcome of the air

2.  What were the contributions of single-purpose aircraft versus the
multirole or dual-role aircraft recommended by DOD's "Bottom-Up

3.  Was there a relationship between aircraft cost and performance? 

4.  How did the TLAM cruise missile perform compared to various

---------------------------------------------------- Appendix IV:1.4.1

Stealth was one of many options used to achieve portions of what was
accomplished in the air campaign.  It could not serve to achieve all
objectives given its operating limitations.  For example, it was not
designed to, and in Desert Storm it did not, engage targets (1) that
were mobile and required searching, (2) that were large "area
targets" requiring coverage by dozens of bombs, or (3) that planners
wanted to attack during the day.  Most notably, the F-117's bomb hit
rate was between 55 and 80 percent, and equally important, its weapon
release rate was only 75 percent. 

In addition, in some respects, other aircraft may have equaled the
F-117 on the very dimensions for which special claims had been made
for it.  The limited data available showed that the F-111F missions
were about as successful in hitting common targets.  Pilots of
aircraft other than the F-117 reported that they, too, achieved
surprise on many, and in some cases most, attacks, according to an
Air Force criterion for the success of stealth--namely that defensive
fire from SAMs and AAA did not commence until after the first bombs
detonated.  While the F-117 attacked targets in every strategic
category--more than any other aircraft--in some categories, very few
strikes were conducted, and every type of aircraft under review
attacked targets in no less than three-fourths of the categories. 
And unlike several other aircraft, the F-117 never faced the daytime
air defenses that turned out to be the war's most lethal. 

As the second most expensive aircraft in our study--costing almost
twice as much as the next most costly aircraft--the F-117 did not
perform as well as several other aircraft on the sorties- and
tons-per-day measures.  For example, the F-15Es averaged 1 mission
per day (about 50 percent higher than the 0.7 average for F-117s) and
averaged 2.71 tons of released munitions per day (246 percent more
than the F-117 average).  The F-16s averaged 1.2 sorties daily (70
percent more) and delivered 75 percent more tonnage daily than the

To maintain stealth, F-117s can carry bombs only internally; this
limits them to two LGBs.  As a result, each F-117 was clearly very
limited in the number of aimpoints it could hit before having to
return home.  Also, the F-117s were based more than 1,000 miles from
Baghdad, which meant a round-trip mission as long as 6 hours with
multiple refuelings.  One Air Force explanation for this basing
decision was the need to keep the F-117s out of range of Scud
missiles.  Another explanation was that the air base at Khamis
Mushayt was one of only a select few in-theater bases with sufficient
hangars to house the F-117 fleet and protect its sensitive radar
absorptive coating from the elements.  Another possible reason for
the F-117s being based so far away was the fact that a complex and
time-consuming mission planning process was necessary to exploit its
stealth characteristics.  The time this mission planning system took
and the fact that the F-117 was able to conduct combat operations
only at night could have meant that the time required to fly between
Khamis Mushayt and the Saudi border was not the key limiting factor
on the F-117's Desert Storm sortie rate.  Moreover, unlike other
aircraft, such as the A-10 and the F-16, the F-117 did not fly from a
more distant main operating base to a forward one from which multiple
sorties were generated. 

Other Desert Storm aircraft were also limited by their distance from
targets.  For logistics reasons, most B-52s flew from far more
distant bases than the F-117s, resulting in a slightly lower 0.6
average on daily sorties.  In contrast, the B-52s had nearly eight
times the daily average munition delivery (8.69 tons versus 1.10
tons) because of their greater carrying capability.  Navy planes on
carriers in the Red Sea were similarly limited in terms of sortie
rates because of the distance from targets and carrier rotations.\3

The A-6Es averaged 1.1 daily sorties and 1.16 tons per day in
munitions.  The F/A-18s averaged nearly the same number of daily
sorties (1.2), but delivered only an average of 0.74 tons of
munitions per day, approximately two-thirds that of the F-117s. 
F-111Fs were based 525 nautical miles from the Iraqi border, some
F-16s were 528 nautical miles from the border, and some F-15Es were
about 250 nautical miles away.  Thus, distance to targets was clearly
a factor in various aircrafts' sortie rates, but it was not the only
factor; additional reasons included complex mission planning
requirements, logistics needs, inability to operate out of forward
operating bases, and requirements to operate only from aircraft
carriers that could not be deployed close by. 

Nevertheless, distance to target alone cannot explain performance,
since the F-111Fs averaged 0.9 sorties per day (28 percent higher
than the F-117s) but released only 0.7 tons of munitions per day per
aircraft--36 percent below the F-117's average.  Similarly, the
British, Saudi, and Italian air-to-ground variants of the Tornado
flew slightly more sorties per day (0.9), yet they delivered less
than half the daily tonnage (0.4). 

At the same time, the F-117s were not able to perform tasks routinely
carried out by other aircraft because of the operating trade-offs
that were necessary to enable them to be stealthy and to deliver
LGBs.  Such routine tasks include strikes in poor weather or under
any conditions in daylight or dusk, attacks against mobile targets
that required searching, and missions that required deviation after
takeoff from planned flight paths. 

However, to the extent that air defense systems depend on radar, it
is surely an advantage to be less detectable by radar than other
aircraft, and the available evidence suggests that in Desert Storm,
the F-117 was not easily detectable by radar.  However, nonstealthy
aircraft were also able to escape engagement by radar-based defense
systems by other means such as by being masked by jamming support
aircraft or by virtue of the physical destruction of radars by SEAD
aircraft such as the F-4G.  Moreover, given the widespread jamming
that occurred in Desert Storm, the availability of fighter
protection, as well as the relatively rapid degradation of the Iraqi
IADS, it is clear that the F-117s sometimes also benefited from these
support factors and did not always operate independent of them. 

\3 Sortie rates and munition payloads cited here are for all Navy
aircraft, from both Red Sea and Persian Gulf carriers and for Marine
Corps F/A-18s and A-6Es based on land. 

---------------------------------------------------- Appendix IV:1.4.2

In its October 1993 "Bottom-Up Review," DOD expressed a strong
preference for multirole as opposed to "special-purpose" aircraft
because "multi-role aircraft, capable of air superiority, strike, and
possible support missions have a high payoff."\4 The use of both
types of aircraft in Desert Storm permits a comparison on some
dimensions of their performance and contribution. 

The Navy F/A-18 was the only multirole aircraft that was actually
employed in both air-to-ground strikes and air-to-air engagements.  A
large number of F/A-18 missions, especially in the early stages of
the air war, were escort, and one F/A-18 was credited with two
air-to-air kills.  Although the F-16 and the F-15E were equipped with
guns and missiles for self-defenses, neither of these Air Force
multirole aircraft performed any escort or air-to-air missions. 
Air-to-air engagements for Air Force aircraft were the domain of the
single-role, air-to-air, F-15C, which was neither equipped nor tasked
to air-to-ground missions.\5 While the exercise of air-to-air
capability by Air Force multirole aircraft was apparently strongly
discouraged, air supremacy meant that there was limited need for
air-to-air capability, and what did exist was adequately covered by
F-15Cs and F-14s.\6

With regard to support roles, F/A-18s and F-16s employed HARM
missiles and other munitions to suppress enemy air defenses. 
However, this supplemented rather than eliminated the role played by
specialized F-4Gs, EF-111s, and EA-6Bs--all of which were used
extensively in SEAD or jamming. 

The data available permit a limited comparison of multirole aircraft
and more specialized, single-role bombers (F-117, F-111F, A-10, A-6E,
GR-1, and B-52) in the air-to-ground mission.  In terms of unit cost,
the single-role aircraft are both the most and the least expensive
(the B-52 and the F-117 versus the A-10).\7 In terms of the average
daily sorties, only the single-role A-10 exceeded the multirole
F-16's and F/A-18's rate of 1.2 per day.  Excluding the B-52,
multirole aircraft had the highest as well as the lowest average
daily munition tonnage; the F-15E was the highest, at 2.71 tons, and
the F/A-18 was the lowest, at 0.74 tons. 

On other performance measures, the two aircraft types appear to be
generally indistinguishable.  All were very survivable, most had
comparable overall night and weather capability, as well as similar
night and weather limitations, and most delivered a mix of guided and
unguided munitions. 

In terms of the ratios reflecting rate of participation against
successfully and not fully successfully destroyed targets, the
single-role F-111F had the highest ratio and the single-role B-52 had
the lowest ratio.  However, the multirole F/A-18 had a ratio that was
nearly identical to that of the B-52, and the multirole aircraft with
the highest ratio--the F-16 at 1.5:1--had a ratio that was 47 percent
of the F-111F's 3.2:1. 

In sum, in air-to-air combat, multirole aircraft had only minimal
opportunity, accounting for only 2 of 38 air-to-air kills.  Some
multirole aircraft were used in air-to-air support SEAD missions, but
their use did not halt the need for aircraft specialized for those
type of missions.  Both single and multirole aircraft appeared at
both ends of the cost scale.  As a generic type, multirole aircraft
did not demonstrate any major payoff in the air-to-air role since the
more specialized F-15Cs accounted for almost all air-to-air kills. 
In the air-to-ground role, multirole performed at the same overall
level as specialized aircraft.  Generally, the multirole aircraft did
not perform as multirole aircraft in Desert Storm. 

However, using Desert Storm data, it is not possible to reach firm
conclusions about the multirole aircraft's potential payoff, relative
to single-role aircraft.  With greatly varying total program unit
costs, as well as a wide range of daily average bomb tonnage dropped
and especially the apparent lack of need for multirole aircraft on
missions other than air-to-ground attack, the case for or against
multirole and single-role aircraft is not readily apparent solely
from Desert Storm experiences. 

\4 DOD, Report on the Bottom-Up Review, Les Aspin, Secretary of
Defense, (Oct.  1993), p.  36.  Secretary of Defense William J. 
Perry endorsed the Bottom-Up Review and has not altered the review's
advocacy of multirole aircraft over special purpose aircraft. 

\5 U.S.  F-15Cs were credited with 31 coalition air-to-air kills, 87
percent of the Desert Storm total.  F-14s were also assigned to the
air-to-air mission; however, none had any air-to-air kills of
fixed-wing aircraft (though one enemy helicopter was shot down by an

\6 Pilots told us that Gen.  Horner said the first F-16 pilot to
unload his bombs in order to attack an Iraqi aircraft would be "sent

\7 In terms of sortie cost, the single-role A-6E and the F-111F were
high and the multirole F-16, F/A-18, and F-15E were lowest and lower;
however, it is not clear whether it was the A-6's and F-111F's much
older age than the multirole aircraft that explains their higher cost
or their single role. 

---------------------------------------------------- Appendix IV:1.4.3

It is often asserted that, on average, the more that something
costs--such as a passenger car--the better it is, compared to similar
things that cost less.  A more expensive automobile is assumed to
possess certain performance qualities that make it superior to a
low-cost car:  it might accelerate more quickly, handle more
precisely, or ride more comfortably.  Moreover, these advantages are
assumed not to have limitations that would prevent the car from being
used as frequently as one chose or under a wide variety of
conditions.  Similarly, a common impression of military hardware is
that an airplane that costs much more than others would have greater
capabilities that distinguish it from other aircraft, making it
overall a "better" aircraft.  While this perception may appear to be
simplistic, it has been sufficiently widespread, even among military
experts, to warrant examination in light of the Desert Storm data. 
Moreover, DOD commonly justifies very costly aircraft and other
weapons to the Congress, and to the public, on the grounds that they
are more capable than other aircraft and they offer unique
capabilities that warrant the greater cost. 

In this section, we consider aircraft total program unit costs and
whether there was any discernible correlation between those costs and
the Desert Storm performance measures cited above.  As noted above,
at $111 million and $164 million, respectively, the F-117 and B-52H
cost far more than the next most expensive aircraft under review (the
$68 million F-111F).  The A-10, at $12 million, was the least
expensive U.S.  air-to-ground aircraft in Desert Storm; the
F-16--with one LANTIRN pod--was the next least expensive at $23

Survivability and Operating Conditions.  In terms of aircraft
survivability, high- and low-cost aircraft were almost identical at
night and at medium-to-high altitudes (as shown in app.  III,
statistically speaking, there was no meaningful difference in the
survivability rates of any of the Desert Storm air-to-ground
aircraft).  Most high- and low-cost aircraft were able to operate
both day and night, although high-cost F-117s, F-111Fs, and F-15Es
were used almost exclusively in the more survivable nighttime
environment.  In effect, in general, high cost did not correlate with
improved survivability, although it may correlate with it in the case
of the F-117, which, as intended, operated only at night and at
medium altitudes--an environment where substantially fewer aircraft
casualties occurred in Desert Storm. 

In terms of other environmental conditions, there was no pattern of
high-cost aircraft offering consistently better performance in
adverse weather.  Indeed, the more costly aircraft with LGB
capability were more likely to be vulnerable to weather degradation
than were aircraft that used unguided ordnance.  For, while both
types of aircraft delivered guided and unguided ordnance, most of the
more costly aircraft delivered more guided, relative to unguided,
bombs.\8 One reason for this was that low-cost aircraft were not
equipped to deliver LGBs, which can partially account for aircraft
cost differentials.\9 Whether the capability to deliver LGBs versus
unguided munitions made the platform more or less effective would
depend on an assessment of the relative merit of those munition
types, discussed later in this appendix. 

Number Deployed.  All other things being equal, one would expect that
the more costly an aircraft, the fewer would be available to be
deployed in combat, since fewer would likely have been produced.\10
This proved to be the case in Desert Storm, with 251 F-16s and 148
A-10s deployed compared to 42 F-117s, 48 F-15Es, and 66 F-111Fs. 
Although obvious, it may be worth recalling that, in terms of total
program unit costs, a single F-117 costs about as much as about 9
A-10s; a single F-111F equals 3 F-16s with LANTIRN pods. 

Thus, in assessing an overall force, the appropriate comparisons
should not be between one high-cost aircraft and one low-cost
aircraft because to acquire equal forces of the two would obviously
require vastly different amounts of money.  A more appropriate way to
measure aircraft forces might be the number of aircraft that an equal
amount of acquisition funding can purchase.  For example, the fleet
of 42 F-117s deployed to Desert Storm cost $4.7 billion to develop
and build, while the three times larger fleet of 148 A-10s cost $1.7
billion; that is, 106 additional aircraft for $3 billion less. 
Similarly, for the same amount of money, very different sized fleets,
and capability, can be procured.  For example, $1 billion in funding
would procure 9 F-117s or 85 A-10s.  The Desert Storm performance
data reveal that the 9 F-117s would have carried out fewer than 7
sorties per day; in contrast, the 85 A-10s would have flown 119. 
While the design missions of the two aircraft differ substantially,
their use in Desert Storm demonstrated that they are not necessarily
mutually exclusive.  Nearly 51 percent of the strategic targets
attacked by the stealthy F-117s were also attacked by less costly,
conventional aircraft--such as the F-16, F-15E, and F/A-18.\11

Based on its performance in Desert Storm, advocates of the F-117 can
argue that it alone combined the advantages of stealth and LGBs,
penetrated the most concentrated enemy defenses at will, permitted
confidence in achieving desired bombing results, and had perfect
survivability.  Advocates of the A-10 can argue that it, unlike the
F-117, operated both day or night; attacked both fixed and mobile
targets employing both guided and unguided bombs; and like the F-117,
it suffered no casualties when operating at night and at medium
altitude.  In short, the argument can be made that to buy more
capability, in the quantitative sense, the most efficient decision
could be to buy less costly aircraft.  Moreover, to buy more
capability in the qualitative sense, it may be a question of what
specific capability, or mix of capabilities, one wants to buy:  in
the F-117 versus A-10 comparison, each aircraft has both strengths
and limitations; each aircraft can do things the other cannot. 
Therefore, despite a sharp contrast in program unit costs, based on
their use, performance, and effectiveness demonstrated in Desert
Storm, we find it inappropriate to call one more generally "capable"
than the other. 

The data did not demonstrate a consistent relationship between the
program unit cost of aircraft and their relative effectiveness
against strategic targets, as measured by the ratio of fully
successful to not fully successful target outcomes for the set of
strategic targets attacked by each type of aircraft.  For example,
while the high-cost F-111F had the highest ratio of all aircraft
reviewed, the relatively low-cost F-16 had a higher ratio than either
the F-117 or the F-15E, both of which were on the high end of the
cost scale.  The F/A-18, in the middle of the cost scale, had a low
ratio of participation against successfully destroyed targets
relative to unsuccessfully destroyed targets, but the medium-cost
A-6E had a ratio that was higher than or equivalent to the F-15E and
F-117, both much higher cost aircraft.  However, the F-117 and the
F-111F, two high-cost, LGB-capable aircraft, ranked first and third
in participation against successful targets.\12

Summary.  We found no clear link between the cost of either aircraft
or weapon system and their performance in Desert Storm.  Aircraft
total program unit cost does not appear to have been strongly
positively or negatively correlated with survivability rates, sortie
rates or costs, average daily tonnage per aircraft, or success ratio
of unguided-to-guided munition deliveries.  No high-cost aircraft
demonstrated superior performance in all, or even most, measures, and
no low-cost aircraft was generally inferior.  On some measures
low-cost aircraft performed better than the high-cost ones (such as
sortie rate, sortie cost); on some measures, the performance of low-
and high-cost aircraft was indistinguishable (such as survivability
and participation against targets with successful outcomes). 

\8 Two prominent exceptions to this are the high-cost B-52, which
delivered very few guided munitions, and the low-cost A-10, which
delivered about 4,800 guided Maverick missiles. 

\9 For example, providing the low-cost A-10 with LGB-capability
would, at a minimum, raise the A-10 unit price by about $7.4 million
by adding two LANTIRN pods, to $19.2 million, an increase of 63
percent.  This increase would, however, result in the A-10's
continuing to be the lowest cost aircraft under review.  It should
also be noted that since the war, the relatively low-cost F-16 has
been equipped with both types of LANTIRN pods, thus enabling it to
deploy LGBs. 

\10 The statistical correlation between aircraft unit cost and
numbers deployed was r = -0.54.  Unit cost data for the Tornado was
calculated as the average of the highest and lowest fiscal 1994 unit
cost figures that were available.  The number of Tornado GR-1s
deployed to Desert Storm was taken from the British AOB for February
1991 cited in the British Ministry of Defense Gulf War Lessons
Learned report.  The number of all other aircraft deployed to Desert
Storm was taken from DOD's title V report.

Because the number of aircraft deployed to battle is likely to be
related to the number available for deployment, we also examined,
where the data permitted, the correlation between the number of
aircraft produced and unit cost.  The correlation was r = -0.54,
indicating that more costly aircraft are produced in smaller numbers,
thus leaving fewer available for deployment, relative to less costly
aircraft.  GR-1 data are not included because production numbers for
these aircraft were unavailable. 

\11 The incompleteness of A-10 strike data prevents our identifying
the extent, if any, to which, A-10 and F-117 target taskings
overlapped.  However, each type of aircraft performed 40 or more
strikes in the following strategic target categories:  C\3 , KBX,
OCA, SAM, and SCU. 

\12 Participation by each type of air-to-ground aircraft against
targets assessed as fully successful targets was as follows:  F-117 =
122, F-16 = 67, F-111F = 41, A-6E = 37, F/A-18 = 36, F-15E = 28, B-52
= 25, and GR-1 = 21.  No data were available for the A-10.  The TLAM
participated against 18 targets assessed as fully successful. 
Participation against FS targets by type of aircraft is a function of
two factors--the breadth of targets tasked to each type of aircraft
(see app.  III) and their FS:NFS ratio as presented previously. 

---------------------------------------------------- Appendix IV:1.4.4

The Navy's TLAM cruise missile is substantially different from the
aircraft reviewed.  Its unit cost of approximately $2.9 million is
clearly well below that of any aircraft, but because it is not
reusable, it had the highest cost per sortie.  Moreover, there were
major categories of strategic targets (mobile, very hard, or buried
targets) that it was inherently incapable of attacking or destroying. 
Also, like many guided munitions, the TLAM's optical guidance and
navigation system (employed in the last portion of flight) can be
impeded by [DELETED].\13 This means that the costs of hitting any
given target are substantial, given that TLAMs are single-use

These TLAM characteristics must be balanced against the fact that its
employment does not risk an aircraft or its pilot.  There is, of
course, essentially immeasurable benefit to avoiding the loss or
capture of pilots.  However, TLAMs are limited in their applicability
compared to some aircraft because many target types (for example,
very hard targets) are not vulnerable to TLAMs or are not feasible as
TLAM targets (for example, mobile ones).  Further, given the TLAM's
high-unit cost and demonstrated P(k), consideration must be given to
whether a given target is sufficiently valuable to be worth using a
TLAM.  High costs mean that relatively few targets in an air campaign
would be worth targeting with TLAMs, especially if aircraft
survivability is high. 

\13 Even if the P(k) for a single TLAM against a given target is
[DELETED], no less than [DELETED] missiles would be required to guard
against reliability failure if the target is deemed to have urgent or
high value. 

-------------------------------------------------------- Appendix IV:2

A review of the cost and use of the air-to-ground munitions in Desert
Storm supplements the foregoing assessment of aircraft to examine
what aircraft-munition combinations may have been the most effective
in the air campaign.  The GWAPS study presented data on air
combat-related ordnance expended in Desert Storm by U.S.  forces. 
Neither a separate breakout nor ordnance dropped by other coalition
air forces was available.\14 Five major types of ordnance were
released by U.S.  air-to-ground aircraft in Desert Shield and Desert
Storm.  Table IV.2 shows these and their cost. 

                               Table IV.2
                  Desert Shield and Desert Storm Air-
                 Related Ordnance Expenditures by U.S.

                                                    Number      Cost\a
----------------------------------------------  ----------  ----------
Bombs and noncruise missiles
Unguided bombs                                     210,004      $432.0
Guided bombs                                         9,342       298.2
Antiradiation missiles                               2,039       510.9
Air-to-surface guided missiles                       5,448       549.1
Total                                              226,833    $1,790.2
Cruise missiles
TLAMs                                                 297\      $861.3
CALCMs                                                  35        52.5
Total                                                  332      $913.8
Total bombs and missiles                           227,165    $2,704.0
\a In millions of fiscal 1990 dollars. 

Source:  GWAPS, vol.  v, pt.  I (Secret), pp.  581-82, and DOD
Selected Acquisition Report on TLAM. 

It is evident from table IV.2 that while the vast majority of the
expended ordnance was unguided--92.4 percent--the inverse was true
for cost.  About 84 percent of cost was accounted for by the 7.6
percent of ordnance that was guided.  If the 332 cruise missiles are
excluded--with their extremely high unit costs--unguided ordnance
still represented about 92.6 percent of the total number expended,
but the percentage of cost for ordnance that was guided decreases to
75.9 percent. 

The points summarized in table IV.3 concerning the relative strengths
and weaknesses of guided and unguided munitions are supported in the
discussion below. 

                                    Table IV.3
                      Relative Strengths and Limitations of
                     Guided and Unguided Munitions in Desert

Measure           Relative strengths              Relative limitations
----------------  ------------------------------  ------------------------------

Cost              No demonstrated strengths.      High unit cost; cost ratio of
                  LGBs, Mavericks, and other      LGBs to unguided unitary bombs
                  guided munitions were much      ranged up to 48:1; for
                  more expensive than unguided    Mavericks, 164:1.

Survivability     Varying amounts of standoff     Standoff capability did not
                  capability avoided defenses     negate defenses not at the
                  collocated with the target.     target. [DELETED]
                  LGB and other guided munition
                  use permitted medium-and high-
                  altitude releases while
                  retaining accuracy, thus
                  reducing aircraft
                  vulnerability to AAA and IR

Operating         Night-capable, clear weather    Adverse weather, clouds,
characteristics   (except for most EO guidance    smoke, dust, haze, and
                  systems); some correctable      humidity either eliminated or
                  accuracy degradation from high  seriously restricted
                  winds.                          employment. Sometimes required
                                                  precise intelligence and more
                                                  demanding mission planning.

Effectiveness     Sometimes highly accurate even  "One target, one bomb" is an
                  from high altitudes, even       inappropriate and illusive
                  against point targets; lower    characterization of LGB
                  likelihood of collateral        effectiveness; no consistent
                  damage.                         relationship between use of
                                                  guided munitions and targets
                                                  that were successfully


Cost              Low unit cost; made up 92       No cost disadvantages
                  percent of the munitions used   identified.
                  but only 16 percent of
                  munitions cost.

Survivability     Permitted higher pilot          Little or no standoff
                  situation awareness and more    capability from defenses at
                  ready ability to maneuver to    target except for use at high
                  evade threats.                  altitude, which severely
                                                  degraded accuracy.

Operating         Exploited radar bombing         Nonradar unguided bombing
characteristics   systems impervious to weather   systems had virtually as many
                  but only for missions           limitations from weather,
                  requiring limited accuracy.     smoke, dust, and so on as
                                                  guided munition sensors;
                                                  accuracy seriously degraded by
                                                  winds, especially when used at
                                                  medium-to-high altitude.

Effectiveness     Of all munitions used, 92       Not accurate from medium-to-
                  percent were unguided;          high altitude against point
                  unguided munition use was an    targets. Higher likelihood of
                  essential part of the air       collateral damage; no
                  campaign, especially against    consistent relationship
                  area targets and ground         between use of unguided
                  forces.                         munitions and targets that
                                                  were successfully destroyed.

\14 The ordnance included cruise missiles, of which 35 were CALCMs
launched from B-52s, and 297 TLAMs from Navy ships and submarines. 
We include both types of missiles because they were integral to the
air campaign in terms of their targets and their role in the planning
of the air campaign. 

---------------------------------------------------- Appendix IV:2.0.1

We analyzed and compared the munitions used in Desert Storm,
calculating the weighted average unit cost for each munition, which
is based on the different numbers of each type used and their unit
cost.  Table IV.4 compares these costs for unguided unitary bombs,
unguided cluster bombs, LGBs, the IR/EO guided GBU-15, and the
Maverick and Walleye air-to-surface munition. 

The data in table IV.4 show that there are very large differences in
the unit costs between the categories of guided and unguided
munitions, as well as substantial cost variations within each
category.  The unguided unitary bombs used in the air campaign cost,
on average, $649 each, while LGBs cost, on average, more than $31,000
each--a cost ratio of about 1:48.  The cost ratio of the average
unguided unitary bomb to the other major type of guided munition, the
Maverick, was 1:164.  Even the cost for more expensive unguided
cluster munitions was just one-fifth the average LGB and
one-eighteenth the cost of a Maverick.\15

In terms of munition expenditures, 17 times more unitary unguided
bombs were dropped than LGBs and 30 times more unguided unitary bombs
than Mavericks.  Six times more cluster munitions were used than
LGBs, 11 times more clusters than Mavericks. 

                               Table IV.4
                 Unit Cost and Expenditure of Selected
                Guided and Unguided Munitions in Desert

                                Unit    Number                    unit
Munition                        cost  expended    Total cost    cost\b
--------------------------  --------  --------  ------------  --------
Unguided unitary
 MK-82LD                        $498    69,701   $34,711,098
 MK-82HD                       1,100     7,952     8,747,200
 MK-83                         1,000    19,018    19,018,000
 MK-84GP                       1,871     9,578    17,920,438
 MK-84HD                       2,874     2,611     7,504,014
 M-117                           253    43,435    10,989,055
 Subtotal                              152,295   $98,889,805      $649
Unguided cluster
 CBU-52/58/71                  2,159     7,831   $38,497,129
 CBU-87                       13,941    10,035   139,897,935
 CBU-89                       39,963     1,105    44,159,115
 CBU-72                        3,800       254       965,200
 CBU-78                       39,963       209     8,352,267
 MK-20                         3,449    27,987    96,527,163
 Subtotal                               57,421  $328,398,809    $5,719
Laser guided
 GBU-10                      $22,000     2,637   $58,014,000
 GBU-12                        9,000     4,493    40,437,000
 GBU-16                      150,000       219    32,850,000
 GBU-24                       65,000       284    18,460,000
 GBU-24/109                    5,000       897    76,245,000
 GBU-27                       75,539       739    55,823,321
 GBU-28                      100,000         2       200,000
 Subtotal                                9,271  $282,029,321   $30,421
IR GBU-15                   $227,600        71   $16,159,600  $227,600
IR and EO Maverick
 AGM-65B                     $64,100     1,673
 AGM-65C                     110,000         5  $107,239,300
 AGM-65D                     111,000     3,405       550,000
 AGM-65E                     101,000        36   377,955,000
 AGM-65G                     269,000       177     3,636,000
 Subtotal                                5,255    47,613,000  $102,187
Walleye II
 AGM-62B                     $70,000       133    $9,310,000   $70,000
Total                                  224,446  $1,271,680,8
\a In fiscal 1990 dollars. 

\b The weighted average unit cost for each general munition type
takes into account the different numbers of each munition type
actually used. 

Source:  GWAPS, vol.  V, pt.  I (Secret), pp.  581-82. 

Similar to our findings regarding the relationship between aircraft
cost and numbers deployed, the data in table IV.4 show that the more
costly a munition, the fewer were expended, for both guided and
unguided categories of munitions.\16 More than 150,000 unguided
unitary bombs were expended, costing just under $100 million, while
in contrast, the 9,271 LGBs used cost over $282 million.  Only 5,255
Maverick missiles were used, but these cost over $536 million, or 30
percent of all noncruise missile costs, while representing about 2.3
percent of ordnance expended.  Even if cruise missile costs are
included, Mavericks were 21.5 percent of total ordnance costs, or
nearly 10 times their share of total ordnance numbers. 

\15 Some unguided munitions were more expensive than some guided: 
CBU-89s cost four times more than GBU-12s, [DELETED]. 

\16 The Pearson correlation coefficient between the number of
munitions expended and cost was negative and moderate in size, r =
-0.42.  The correlation between the number of unguided munitions
expended and unguided munition cost was r = -0.44, while the
correlation between the number of guided bombs expended and munition
cost was slightly stronger r = -0.52, although still in the moderate

------------------------------------------------------ Appendix IV:2.1

The data available permit us to calculate the munition costs to
attack the targets assessed in appendix III as fully successfully
destroyed and not fully successfully destroyed.  These data are shown
in table IV.5, grouped into target categories, for targets that we
were able to evaluate from DIA phase III damage assessments.  Data
for the A-10 are not included, for the reliability reasons noted

                                    Table IV.5
                     Number and Cost of Munitions Expended by
                        Target Category and Success Rating

Ta                                              S:                   Total
rg              Number   Munitions   Munitions  FS    Total BE        cost  NFS:
et  Rating    of BEs\a    expended      per BE  \      targets    per BE\b   FS
--  --------  --------  ----------  ----------  --  ----------  ----------  ----
C\  FS              62         974        15.7  2.         105        $190  1.58
 3                                              41
    NFS             43       1,626        37.8                         300
EL  FS              10       1,298       129.8  1.          14         391  0.30
 E                                              92
    NFS              4         996       249.0                         119
GV  FS              10         139        13.9  0.          21         186  1.91
 C                                              90
    NFS             11         133        12.1                         356
LO  FS              28         605        21.6  2.          40         300  1.01
 C                                              47
    NFS             12         641        53.4                         302
MI  FS              17       4,814       283.2  1.          50       1,590  0.69
 B                                              11
    NFS             33      10,378       314.5                       1,091
OC  FS              22       7,682       349.0  0.          34       4,661  0.95
 A                                              73
    NFS             12       3,059       254.9                       4,445
OI  FS               4       1,017       254.3  0.          16         447  1.07
 L                                              44
    NFS             12       1,353       112.8                         478
NA  FS               3         132        44.0  2.          13         323  4.13
 V                                              10
    NFS             10         939        93.9                       1,334
NB  FS              15       1,458        97.2  2.          20       1,600  2.72
 C                                              79
    NFS              5       1,354       270.8                       4,346
SA  FS              10         189        18.9  0.          14          51  4.84
 M                                              63
    NFS              4          48        12.0                         248
SC  FS               5         972       194.4  0.          20         929  1.52
 U                                              90
    NFS             15       2,633       175.5                       1,416
\a BEs attacked exclusively by cruise missiles are not included. 

\b Costs are in thousands of fiscal 1991 dollars.  As official data
on the cost of British munitions were not available to us, we assumed
that the cost of the U.K.  1000 LGB was equivalent to the GBU-10, the
most common U.S.  LGB. 

Few, if any, consistent patterns can be discerned from the data shown
in table IV.5.  Among targets rated FS, the average number of
munitions used per BE ranged from about 12 to 350; among NFS targets,
the average per BE ranged from 12 to 315.  The ratio of munitions
used on targets rated NFS versus FS within each category also showed
great variation--from 0.44 for OIL (on average, less than half as
many munitions were used on OIL targets rated NFS as on those rated
FS) to 2.8 for NBC targets (NBC NFS targets received nearly three
times as many munitions per BE as those rated FS).  Moreover, in 5 of
the 11 target categories, more munitions were expended on FS targets
than on NFS targets; however, in 6 categories, the NFS targets
received more munitions than FS ones.  In other words, success across
categories did not clearly correlate with the amount of munitions

Weapon costs and target success showed some degree of pattern, but it
was counterintuitive:  in most categories, nonsuccess was more costly
than success in terms of the munitions employed.  In three categories
(ELE, MIB, and OCA), the successfully attacked target costs were
higher than those not fully successful.  In the other eight
categories, target costs were higher for the NFS targets. 

To control for outliers, or unrepresentative data from small samples,
we looked at the two categories that received the most munitions, MIB
(15,192 weapons on both FS and NFS) and OCA (10,741 total weapons). 
Even between these two categories there were notable variations.  The
ratio of weapons used on NFS versus FS targets was 1.11 for MIB and
0.73 for OCA--that is, in one target category, FS targets received
more munitions on average than NFS targets, and in the other
category, they received less.  The same was true of cost--in one
category FS targets had higher munitions costs, on average, than NFS
targets and in the other target category, the relationship was
reversed.  In addition, the cost of weapons used for each FS target
was about three times greater for OCA than for MIB ($4.7 million for
OCA targets versus $1.6 million for MIB targets).  Also, because
there were less than twice as many munitions used against FS OCA
targets as FS MIB targets, it is apparent that more expensive
munitions per unit were used against OCA targets than MIB targets. 
However, the ratio of success against MIB targets was more favorable
than against OCA targets. 

Any generalizations must be tempered by the fact that the data are
incomplete in at least three regards:  (1) A-10 weapons expenditures
are absent, and these aircraft conducted approximately 8,000 combat
sorties during Desert Storm, although the great majority were in the
KBX category not listed in table IV.4; (2) the 357 BE-numbered
targets for which FS and NFS evaluations could be made are a subset
of all targets with BEs and a considerably smaller subset of all
targets against which munitions were delivered during the air
campaign; and (3) data on TLAMs and CALCMs are not included. 

Given these limitations, the data shown must be treated as indicators
of Desert Storm performance, not definitive measures.  Two
conclusions seem apparent:  (1) there was great variability in the
number and cost of munitions used to attack targets, whether
successfully or unsuccessfully, and (2) neither greater numbers of
munitions used nor greater munition costs consistently coincided with
success across target categories.  In 6 of 11 categories, greater
numbers of munitions used coincided with NFS, and in 8 out of 11
comparisons, greater cost of munitions more closely coincided with
NFS assessments. 

The use of guided and unguided munitions against the rated targets
can also be compared.  Costs of the weapons delivered, per BE, in
each target category are illustrated in table IV.6.  (Note, data on
TLAMs and A-10s are not included; therefore, both the weight of
effort and costs are somewhat understated.)

Two points can be made from the data shown in table IV.6.  First, in
8 of the 11 target categories, the cost per BE of precision-guided
munitions used on FS targets exceeded the cost of unguided munitions. 
The same is true of the NFS targets in 7 of 11 categories.  However,
in all cases but one (GVC, NFS), more unguided munitions were used
than guided munitions against any target, whether it was successfully
destroyed or not.  Thus, even though more unguided munitions were
almost always used than guided, the cost to use guided munitions was
usually greater. 

Perhaps more importantly, the data in table IV.6 permit an analysis
of whether an increase in the number of either guided or unguided
munitions coincided with successfully destroyed targets.  In only 4
of 11 categories, more PGMs were used on average against the FS than
NFS targets; in 7 of 11, more PGMs were used against NFS targets.  In
5 of 11 categories, more unguided munitions were used against the
successfully destroyed targets.  In other words, the data do not show
that a key difference between successfully and not fully successfully
destroyed targets was that the former were simply bombed more than
the latter.  This was the case for both types of munitions, PGMs and

                               Table IV.6
                     Munition Costs Associated With
                Successfully and Not Fully Successfully
                           Destroyed Targets

                          Number   PGMs     cost   Unguided   Unguided
                              of    per      per  munitions   cost per
Target                     BEs\a     BE     BE\b     per BE      BE\b\
-----------------------  -------  -----  -------  ---------  ---------
C\3                           62    3.5   $160.4       12.2      $29.6
ELE                           10    2.0    307.3      127.8       83.4
GVC                           10    5.9    167.3        8.0       19.0
LOC                           28    6.2    261.7       15.0       38.3
MIB                           17   16.3    982.7      266.9      607.5
NAV                            3    4.3    287.0       39.6       35.9
NBC                           15   19.3  1,194.5       77.9      405.4
OCA                           22   51.9  3,498.4      297.0    1,162.6
OIL                            4      0        0      254.2      446.6
SAM                           10    0.8     22.4       18.1       28.9
SCU                            5   11.6    243.9      182.8      685.3
Total                        186   12.1   $730.7       91.5     $277.0

                          Number   PGMs     cost   Unguided   Unguided
                              of    per      per  munitions   cost per
Target                     BEs\a     BE     BE\b     per BE       BE\b
-----------------------  -------  -----  -------  ---------  ---------
C\3                           43    3.8   $254.4       34.0      $45.8
ELE                            4    0.5     11.0      248.5      107.9
GVC                           11    7.7    345.2        4.4       10.4
LOC                           12   10.7    281.4       42.7       20.3
MIB                           33    7.5    775.1      306.9      316.3
NAV                           10    9.1  1,210.6       84.8      123.7
NBC                            5   51.0  4,051.7      219.8      294.6
OCA                           12   39.7  3,355.6      215.3    1,089.6
OIL                           12    0.3     25.1      112.5      452.9
SAM                            4    4.8    104.5        7.3      143.9
SCU                           15    6.0    372.0      169.5    1,043.7
Total                        161    9.7   $761.9      134.1     $314.6
\a BEs attacked exclusively by cruise missiles are not included. 

\b Costs are in thousands of fiscal 1991 dollars.  As official data
on the cost of British munitions were not available to us, we assumed
that the cost of the U.K.  1000 LGB was equivalent to the GBU-10, the
most common U.S.  LGB. 

------------------------------------------------------ Appendix IV:2.2

With regard to the effectiveness of individual munitions, the Desert
Storm data do not permit a comprehensive comparison, since the
effects of one type of weapon were almost never identified before
other weapons hit the target.  However, pilots did report both pluses
and minuses with both guided and unguided munitions. 

With guided munitions, pilots reported three negative consequences as
delivery altitude increased.  First, because the slant range to
targets was increased by higher altitude, [DELETED]. 

Second, the higher altitude deliveries made LGBs more subject to
winds, and pilots had to correct the [DELETED]. 

A third problem reported by F-117 and F-15E pilots was the need to
revise some of the computer software for LGBs to accommodate the
higher altitude tactics.  [DELETED]

While each of these problems affected accuracy, they were correctable
or caused problems only on the margin.  The accuracy problems
encountered by unguided munitions were more difficult, if not
impossible, to overcome.  Pilots of virtually every type of aircraft
remarked that they had little confidence in hitting point targets
with consistent accuracy from high altitudes with unguided bombs. 

Several reasons were cited.  First, pilots stated that much of their
training before Desert Shield had been for low-altitude tactics.  As
a result, some pilots had to learn high-altitude bombing techniques
either just before or during the war.  Second, the Persian Gulf
region experienced winds that were both strong (as much as 150 mph in
jet streams) and unpredictable.  The high-altitude tactics
exacerbated the effects of these winds wherever they occurred,

As a result, pilots reported considerable difficulty attacking small,
point targets, such as tanks, from high altitude with unguided bombs. 
Some expressed a high level of frustration in being assigned to do so
and said that it was simply inappropriate, even "ridiculous," to
expect that unguided bombs were capable of hitting a target like a
tank from high altitude with any consistency.  It was also clear that
such inaccuracy made unguided munitions inappropriate for use in
inhabited areas, where civilian assets could easily be hit in error. 

The large number of circumstances using unguided munitions was
described by pilots as both appropriate and effective.  These
included military units in the field or other large, area targets,
such as buildings or complexes of buildings, when not near civilian

------------------------------------------------------ Appendix IV:2.3

Beyond the experiences and observations of pilots, the data permit
some analyses that shed some additional insights about the relative
effectiveness and cost of different munition types. 

CNA was able to analyze U.S.  Navy attacks against certain bridges
that employed LGBs, unguided bombs, and Walleye electro-optical
guided bombs.\17 The CNA data and analysis are only one of a few
instances where it is possible to link target damage with the use of
specific types and numbers of munitions.  The analysis separated out
the effects of attacks with the different munitions, and it found 29
strikes on bridges where the BDA was unambiguous--that is, when no
other attack was scheduled between the time of the attack and the
collection of the BDA. 

The study found that in eight strikes against bridges using Walleye,
[DELETED].  The same rate of success was found when unguided
munitions were used--[DELETED].\18 CNA data also reveal that, on
average, more unguided munitions were delivered per bridge strike
than guided munitions.  On Walleye missions, an average of 1.3 bombs
were used per strike.  When LGBs were employed, an average of 3.2
were delivered per strike.  And when unguided bombs were selected, an
average of 15 were used per bridge. 

Table IV.7 presents the cost of each type of munition employed and
calculates the average cost of munitions per dropped span. 

                               Table IV.7
                  Number and Cost of Munitions Used in
                   Naval Air Attacks on 13 Bridges in
                              Desert Storm

                                                   Average       Total
                                                 munitions        cost
               Numbe                              per span    per span
Munition type      r  Cost\a  Spans dropped        dropped     dropped
-------------  -----  ------  --------------  ------------  ----------
Walleye            8  $560,0    [DELETED]        [DELETED]   [DELETED]
LGB               34  1,260,        3                 11.3     420,000
Unguided          85  120,05        1                 85.0     120,052
\a The range in costs of guided munitions used against these bridges
was from $22,000 for the GBU-10 to $150,000 for the MK-83 LGB.  The
range in costs for unguided munitions was from $498 for the MK-82 to
$1,871 for the MK-84. 

Based on table IV.7, we find that (1) far fewer--as few as
one-tenth--the number of guided munitions than unguided were
required, on average, to destroy a bridge; (2) there is an inverse
relationship regarding cost--that is, the cost to drop a span with
guided munitions was three-to-four times more than the cost of
unguided munitions; and (3) as with our previous analysis, the Desert
Storm evidence did not substantiate the "one-target, one-bomb"
claim--rather, on average, 11 laser-guided bombs were used for each
span dropped.  (See app.  III.)

These conclusions must be treated cautiously.  The sample is from only
13 bridges and consists only of Navy aircraft and munitions.  Within
these limitations, the data support our previous findings concerning
the relationship between the cost and effectiveness of guided and
unguided munitions and the numbers actually used to achieve target

\17 CNA, Desert Storm Reconstruction Report; Volume II:  Strike
Warfare (Secret), Alexandria, Va.:  1992. 

\18 Using these data, CNA concluded (on p.  6-41) that "Irrespective
of weapon employed, for those bridge strikes with directly associated
BDA, 17 percent of the strikes dropped a [bridge] span.  When
considering individual weapon types, the percentages of strikes
resulting in dropped spans are similar, although the percentage for
LGB/GBU strikes is somewhat higher.  When the indeterminate BDA cases
are considered, the individual results become indistinguishable."

------------------------------------------------------ Appendix IV:2.4

Using the Missions database and the phase III BDA messages, we
performed a second analysis of attacks against bridges.  The phase
III messages included 24 bridges attacked by both Air Force and Navy
aircraft.  Nineteen of these were successfully destroyed; five were
not.  The BDA did not, in these cases, allow any distinctions of what
munition type effected the damage.  Using the munition cost data in
table IV.4, we calculated the munition cost to successfully destroy a
bridge with both Air Force and Navy aircraft and guided and unguided
munitions.  The results are shown in table IV.8. 

                                    Table IV.8
                     Munitions Costs to Attack 24 Bridges in
                                   Desert Storm

Assessment                     Cost of                     Cost of       Cost of
of bridges        Guided        guided      Unguided      unguided       bridges
attacked       munitions   munitions\a     munitions   munitions\b    attacked\c
----------  ------------  ------------  ------------  ------------  ------------
FS                  10.8      $237,600          18.2       $34,052      $271,652
NFS                  7.2       158,400          14.2        26,568       184,968
\a This assumes that the guided munition used was the GBU-10, with a
unit cost of $22,000. 

\b This assumes that the unguided munition used was the MK-84 GP,
which pilots stated to be the unguided munition of choice against
bridges.  The unit cost of this munition was $1,871. 

\c The average costs to attack bridges presented in tables IV.7 and
IV.8 are not directly comparable for two reasons.  First, the
chronological BDA and strike data compiled by CNA allowed the
calculation of costs up to and including the first successful strike
against a bridge.  Unambiguous chronological BDA and strike data were
not available through the missions and phase III databases; thus,
costs include strikes before, during, and after the initial
successful attack.  Second, the criteria for success are different. 
In the CNA study, the criterion was "span dropped." In our
interpretation of DIA's phase III messages, the criterion was that
the mission objective was met, which often equated with the absence
of a restrike recommendation.  In addition, ambiguous BDA was
included in the NFS category. 

While these data do not distinguish the effects of different types of
munitions, they do support many of the points made earlier.  First,
as with the CNA data, it is clear that while fewer guided munitions
were used, their cost was substantially higher.  Second, in both
analyses, about 11 LGBs were used per destroyed bridge.  Thus, the
data from the CNA analysis--with unambiguous BDA--suggest that our
analysis of the number of LGBs dropped per successful target--in this
case a bridge--is not inappropriate.  It also reinforces the point
that it is misleading to characterize LGBs as "one-target, one-bomb"
weapons.  Third, and finally, there are so few cases where BDA
permits a reliable analysis of the exact number of a specific type of
munition used per successful mission; thus, the data available from
Desert Storm do not permit supportable general conclusions about the
comparative effectiveness of guided versus unguided munitions. 

------------------------------------------------------ Appendix IV:2.5

One characteristic pilots cited as a strong advantage of guided
munitions over unguided was the ability to release a munition at a
substantial standoff distance from a target, thereby limiting
exposure to any defenses at the target.\19 There were, however,
limitations to the advantages of standoff capability.  [DELETED] A-10
pilots noted that Iraqi defenses were not always directly collocated
with the target, with the result that launching a weapon from maximum
delivery range could still expose aircraft to defenses not at the
target.  Standoff capability distanced aircraft from defenses
collocated with the target, but that was not necessarily all the

Another factor cited by pilots about guided munitions was the
relatively high workload required to employ them.  [DELETED]\20


Pilots delivering unguided munitions experienced different problems: 
vulnerability to AAA was high when releasing at the low altitude that
maximized the accuracy of unguided munitions.  Thus, CENTAF's order
to cease low-level deliveries after the third day of the campaign
meant a trade-off of reduced accuracy with unguided bombs for
improved survivability. 

In sum, delivery tactics for guided and unguided munitions both
compromised aircraft survivability but in different ways.  The
advantage of guided munitions to standoff from a target's defenses
varies by PGM type, and some pilots reported that standoff from
target defenses did not always ensure standoff from all relevant
defenses.  Moreover, guided munitions can make aircraft more
vulnerable [DELETED], while maximum accuracy for unguided bombs
requires more dangerous low-altitude delivery. 

\19 Different guided munitions could be delivered at standoff
distances greater or lesser than others:  specifically, the IR
version of the GBU-15 had a standoff capability of up to [DELETED]. 
Maverick missiles stood off at slant ranges of [DELETED].  Unpowered
LGBs were described by some pilots as having a limited standoff

\20 The "heads down" and subsequent situational unawareness problem
was much less of a problem in two-seat aircraft (the F-15E, A-6E, and
the D model of the F/A-18).  In these, the pilot could concentrate on
external threats while the weapon systems officer performed the
"heads down" tasks necessary to deliver the guided munition. 
However, this advantage of two-seat air-to-ground aircraft did not
appreciably reduce the "wings level" time of the aircraft. 

------------------------------------------------------ Appendix IV:2.6

As discussed in appendix II, night, clouds, haze, humidity, smoke,
dust, and wind had significant, but different, effects on guided and
unguided munitions.  Delivery of guided munitions was either limited
or prevented altogether by weather or other conditions that impaired
visibility.  In contrast, when weather and other environmental
conditions affected infrared or optical search sensors for unguided
munitions, they could still be delivered with radar.  Doing so meant
that the ability to identify valid targets among relatively
indiscriminate radar returns was usually poor and accuracy from high
altitude was also poor, but the employment of unguided munitions was
still possible. 

Another operating characteristic was the support that the different
munition types normally required.  Pilots reported varying levels of
intelligence and mission planning they needed for guided and unguided
munitions.  For example, [DELETED].  (In F-111F LGB missions, such as
"tank plinking," detailed information and planning were not
necessary.) Although they strongly preferred receiving detailed
target and mission planning data, pilots using unguided munitions
reported that they often had less support.  For example, B-52 pilots
stated that they sometimes received new targets just before takeoff,
or even when they were en route to a previously planned target, but
the new targeting information was sometimes little more than
geographic coordinates. 

In addition, "precision" for guided munitions requires not only
precise accuracy from the munition but also precise intelligence
support.  Pinpoint accuracy is impossible if the right aimpoint is

In sum, to achieve accuracy, guided munitions were normally more
limited by weather and by their support and intelligence needs than
unguided munitions.  In contrast, unguided munitions were usable in
poor weather, but they were also less accurate. 

-------------------------------------------------------- Appendix IV:3

In this appendix, we found that each type of aircraft and munition
under review demonstrated both significant strengths and weaknesses. 
There was no consistent pattern indicating that either high-cost or
low-cost aircraft or munitions performed better or were more
effective in Desert Storm.\21

The limited data do not show that multirole aircraft were either more
or less effective in the air-to-ground capacity than more
specialized, single-role aircraft.  However, air-to-air missions were
predominantly performed by single-role air-to-air aircraft, and while
multirole aircraft did perform some air-defense escort and some
support missions, their use did not eliminate the need for
single-role, air-to-air, and other support aircraft.  The evidence
from Desert Storm would seem to suggest the usefulness of single-role
aircraft in their respective missions and the usefulness of multirole
aircraft most predominantly in the air-to-ground mission. 

The high-cost F-117 stealth bomber has significant operating
limitations that affect when, where, and how it can be used; its
target hit rate appears to have been matched by the F-111F against
similar targets.  Although the F-117 was often, but certainly not
always, tasked against different targets, on certain performance
dimensions--such as sortie rate, operations in weather, and tonnage
delivered--it did not match the performance of several moderate- and
even low-cost aircraft. 

Guided munitions are many times more costly than unguided munitions,
and their employment was constrained by poor weather, clouds, heavy
smoke, dust, fog, haze, and even humidity.  However, guided munitions
were less affected by winds and, unlike unguided munitions, they were
more consistently accurate from medium-to-high altitude.  Although
quite inexpensive and less restricted by low visibility, unguided
munitions cannot reliably be employed against point targets from the
medium and high altitudes predominantly used in Desert Storm. 

Both guided and unguided munitions have important implications for
aircraft survivability.  To be accurate, unguided munitions need
low-altitude delivery, which in Desert Storm was found to be
associated with too many casualties.  While guided munitions can be
accurate from high altitude, their standoff capability does not
necessarily protect them from defenses not at the target.  [DELETED]

While guided munitions are clearly more accurate from medium and high
altitudes, their high unit cost means that they may not be the least
expensive way to attack certain targets, sometimes by a considerable
margin, compared to unguided bombs.  There was no apparent pattern
indicating that guided munitions were, overall, more effective than
unguided munitions in successfully destroying targets or that the
difference between targets that were successfully destroyed and that
were not fully successfully destroyed was simply that the latter were
not attacked as often as the former by either guided or unguided

The TLAM cruise missile demonstrated a high-cost sortie rate, low
survivability, and severe employment limitations.  Its accuracy was
substantially less than claimed; however, unlike any aircraft, its
use does not risk an aircraft or, more importantly, its pilot. 

\21 Despite the absence of an overall, consistent pattern, there were
clearly cases where both types were ineffective:  weather either
seriously degraded or rendered unusable guided munitions;
high-altitude deliveries made unguided munitions highly inaccurate,
according to pilots who termed the use of unguided munitions against
point targets, "ridiculous." Conversely, there were conditions where
the data indicated that both munitions were effective.