Solving the Mystery of the “Missile Sightings”

The Crash of TWA Flight 800

Randolph M. Tauss

“We just saw an explosion out as far as 40 miles away in Con-

here on Stinger Bee five oh necticut, beachgoers, pleasure
seven.”1 boaters, construction workers,
diners, and others made similar
It was Wednesday, 17 July 1996, sightings. And from more than
8:31:51 p.m. Capt. David 22,000 miles overhead, a mili-
McClaine of Eastwind Airlines, tary satellite had detected heat
piloting a Boeing 737 commuter from something in the vicinity.
flight near the coast of Long
Island, had just become the first But it would be hours, some-

“
Capt. David McClaine
of Eastwind Airlines,
piloting a Boeing 737
recorded eyewitness to one of the
most deadly and mysterious com-
mercial air crashes in US
aviation history. It would be
almost a minute before the
times days or weeks, before these
observations would be assimi-
lated into the official record. For
now, the scope of what had hap-
pened was unfolding before a
commuter flight near importance of what McClaine very few.
the coast of Long had seen would become clear to
the Boston air traffic controller 8:32:56: “TWA eight hundred,
Island, had just with whom he was speaking. [call] Center.”
become the first
recorded eyewitness to 8:31:57: “Stinger Bee five oh It was now 65 seconds after
one of the most deadly seven, I'm sorry. I missed it. Ah, McClaine’s first sighting. The
you’re on eighteen. Did you say controller at Boston Air Route
and mysterious something else?” Traffic Control Center, coordinat-
commercial air crashes ing the flights of the dozen or so
in US aviation history. 8:32:01: “We just saw an explo- aircraft in the vicinity, had
sion up ahead of us here noticed one was missing from his

” something [like] about sixteen

thousand feet [altitude] or some-
thing like that. It just went
down—to the water.”
radar screen. In a professional
but increasingly urgent tone, the
controller was trying to elicit a
response.

Within seconds, other pilots in 8:33:04: “TWA eight hundred,

the vicinity corroborated Center.”
McClaine’s observation.
TWA Flight 800, a Boeing 747
8:32:25: “Boston, Virgin zero zero jumbo jet, had departed New
nine. I can confirm that, out of York’s John F. Kennedy airport
Randolph M. Tauss, a senior my, my nine o’clock position, we 14 minutes earlier. It was en
just had an ex… it looked like an route to Paris with 230 passen-
weapons analyst in the
explosion about five miles away, gers and crew. In his last contact
Directorate of Intelligence when six miles away.” with Boston Center at 8:30:17,
this work was done, received the Capt. Ralph Kevorkian, a pilot
Intelligence Medal of Merit for Up and down the coast of Long with 21 years experience flying
his participation in this effort. Island, and from vantage points for TWA, acknowledged clear-

1
TWA Flight 800

ance from the Boston Center CIA Participation of radar and satellite data, the
controller to climb to 15,000 feet. OTI analysts focused on detailed
But now the plane was nowhere The crash of TWA Flight 800, analysis of eyewitness state-
to be seen. potentially one of the most lethal ments provided by the FBI. After
international terrorist acts ever eight months of work, they con-
8:33:09: “TWA eight hundred, if perpetrated against the United cluded with confidence and full
you hear Center ident[ify].” States, touched off the most substantiation that the eyewit-
extensive, complex, and costly air nesses had not seen a missile. On
The Boston controller was becom- disaster investigation in US his- 28 March 1997, CIA’s Deputy
ing increasingly uneasy, as were tory. Had it been the result of Director for Intelligence sent a
state-sponsored terrorism, it memorandum to FBI Assistant
pilots in the area.
would have been considered an Director James Kallstrom sum-
act of war. marizing the results:
8:33:36, from Alitalia 609: “…just
for your information, sir, we are
FBI and National Transporta- Our analysis demonstrates that
just overhead the explosion, right tion Safety Board (NTSB) the eyewitness sightings of
overhead at this time.” investigators almost immedi- greatest concern to us—the ones
ately focused on three possible originally interpreted to be of a
8:33:48, from McClaine: “Stinger causes: a bomb, a missile, or a possible missile attack—took
Bee, ah, Boston, we are directly mechanical failure. The missile place after the first of several
over the site where that airplane theory seemed particularly plau- explosions aboard the air-
or whatever it was just exploded sible because of reports from craft…combined with the total
and went into the water.” dozens of eyewitnesses in the absence of physical evidence of
Long Island area who, on the a missile attack, [this] leads
8:34:01, from Boston Center: evening of 17 July, recalled see- CIA analysts to conclude that
“Roger that. Thank you very ing something resembling a flare no such attack occurred.
much, sir, we’re investigating or firework ascend and culmi-
that right now. TWA eight hun- nate in an explosion. These findings also were docu-
dred, Center. TWA eight zero zero, mented in an unclassified video
Because of the possibility that produced with the aid of anima-
if you hear Center, ident.”
international terrorists may have tion experts and the Directorate
been involved, the FBI requested of Intelligence’s Video/Multime-
8:35:36, again from Boston: “TWA
CIA’s assistance. In accordance dia Production Center.2
eight hundred, Center.”
with the National Security Act of
1947 and Executive Order 12333, The FBI and others involved in
The gravity of the situation was
the CIA responded to the FBI’s the investigation ultimately con-
now evident to all in the small request within 24 hours of the curred with the CIA’s
group privy to the interchange crash. This support consisted pri- conclusions. On 18 November
that had begun four minutes marily of help from the 1997, Kallstrom aired the CIA
earlier. Counterterrorist Center in the video in its entirety during a
Directorate of Operations and national news conference
8:35:43, from McClaine: “I think from a small group of analysts, announcing the suspension of the
that was him.” including myself, in the Office of worldwide criminal probe:3
Weapons, Technology and Prolif-
8:35:45, from Boston: “I think so.” eration (now the Office of We feel very, very comfortable
Transnational Issues, or OTI) in that what [the CIA] videotape
8:35:48, from McClaine: “God the Directorate of Intelligence. portrays is very close...to what
bless him.” Following a preliminary review happened. The tape was made

2
 TWA Flight 800

to explain, first to [the FBI],

and then to the public, and the
families, and the naysay-
ers…and anyone else out
there—the aeronautical engi-
neers—what these 244 people
saw. What is the explanation?
If they didn’t see a missile,
what could they have seen? All
244 witnesses saw events that
happened after—after—the
Center fuel tank blew up. So
they did not see a missile
attacking the plane.

—FBI Assistant Director James K.

Kallstrom, The Twentieth Century With
Mike Wallace, 3 October 1998

Assembling Information Eyewitness locations (CIA graphic).

The sources of information used “A lot of people saw things in the sky. And a lot of people saw what we think is the
by CIA analysts ultimately same thing. Those witnesses, they’re good people, and they told us what they saw.”
included FBI summaries of state- —FBI Assistant Director James K. Kallstrom, Dateline NBC, 14 March 1997
ments from the 244 eyewitnesses,
some containing sight and sound
observations, and most contain- and heading at the moment the But we did know the precise loca-
ing information that could be CVR and FDR ceased operating; tion of the plane when it
used to determine the locations of the NTSB observation that an exploded, including its altitude
eyewitnesses at the time of the abrupt sound was recorded just (13,800 feet). As it turned out,
crash; DeLorme Version 4.0 before the CVR ceased operating; this information and the fact that
Street Atlas USA commercial and the NTSB observation that the explosion was extraordinar-
mapping software; two sets of no other unusual activity was
ily loud would prove crucial in
radar tracking data (one from a recorded on either the CVR or
radar at Islip, Long Island, pro- solving the mystery of the eye-
FDR.
viding data sampled once every witness “missile sightings.”
4.6 seconds, the other from a Because Flight 800’s radar tran-
radar at Riverhead, Long Island, sponder, CVR, and FDR ceased
providing data sampled once The Eyewitness Accounts
operating the instant the air-
every 12.0 seconds); meteorologi- craft exploded, the altitude
cal data (winds aloft); infrared Although some eyewitness sum-
history of the aircraft from that
(heat) data from a US military time until it hit the water was mary reports proved more useful
satellite; the precise times at not known.4 It was also unknown than others, all were analyzed in
which the cockpit voice recorder how long this took. Radar data painstaking detail, a process that
(CVR) and flight data recorder showed only the horizontal took over a year and entailed
(FDR) ceased operating; the air- motion of the plane after it more than 2,000 man-hours of
craft’s location, altitude, speed, exploded, and this only crudely. work.

3
TWA Flight 800

“
On the whole, the
eyewitnesses provided
remarkably detailed
When conducting our research, Several eyewitnesses made it
we went to places where large descriptions that were clear that they had seen only the
numbers of eyewitnesses had surprisingly latter stages of the disaster:
been, and also visited specific consistent.
vantage points of important eye- • One observer in a beachfront
witnesses. At several of these
locations, videotape templates
were made to use in our analysis
and to provide accurate back-
drops for selected scenes in our
”
• A patron at a local marina
reported seeing an orange flare
condominium reported that he
saw a fireball, which separated
into two fireballs that fell into
the ocean. Immediately after
the fireballs hit the water, he
video. rise in the sky and then come heard loud sounds.
down a little before exploding
On the whole, the eyewitnesses into “two pieces.” She reported • A woman on the beach reported
provided remarkably detailed that the explosion occurred seeing a fiery object falling in
descriptions that were surpris- about 10 seconds after she first two main pieces to the water’s
ingly consistent. The crash surface, followed about four sec-
saw the flare in the sky.
scenario depicted in the video onds later by a rumbling sound.
was the result of the composite • Two observers near the ocean
• An observer on a small boat
analysis of all the eyewitness reported seeing an ascending
reported seeing what at first
reports provided to us by the flare that exploded into an
appeared to be a “shooting
FBI. orange fireball, then descended star,” moving in a “downward,
in “two large flaming chunks” 45-degree sloping arc.” He then
Most eyewitnesses were about 10 to the ocean’s surface. They realized he was observing a
to 12 miles from where the plane both reported hearing a “deep burning aircraft. He saw one
first exploded. The statements thunderous rumble.” wing separate from the fuse-
varied, but many shared com-
lage, and a large “fire trail” of
mon characteristics: • A passenger on a plane burning fuel erupt. Just as the
reported seeing “a small, flare- wing separated, he heard a loud
• A patron outside a local restau-
like projectile traveling in an “concussion” sound. The burn-
rant reported seeing what ing plane and wing then
east-northeasterly direction”
appeared to be a “shooting star” quickly dropped to the water’s
appear in the sky seemingly out
that grew into an oblong ball of surface. (Of all the known eye-
of nowhere. It seemed to be
flames that fell from the sky. witnesses on the land and
ascending, then culminated in a
After the flames fell out of sight water, this person was the
behind the sand dunes, he small explosion, followed by a
much larger fireball, which closest.)
heard two “bangs” that sounded
like fireworks in the distance. appeared to descend toward the
The most useful reports proved to
water.
be those that related observa-
• A patron at a local yacht club tions to specific events in either
reported seeing what appeared • A passenger on a different
space or time. For example, an
to be a boat flare ascend for sev- plane reported seeing an eyewitness report of a loud sound
eral seconds, expand into a “orange-white flare” ascend, “after the fireballs hit the water”
much larger fireball, then split then “bulge” and descend. She was less useful than one of a loud
into two “smoky trails” which likened the ascent and descent sound “about four seconds after
dropped toward the ocean. She to that of a roller-coaster, with the fireballs hit the water.” An
then heard deep rumbling the ascent ending in a “big eyewitness description of “a flare-
sounds. bang,” followed by the descent. like object rising over the ocean”

4
 TWA Flight 800

“
The key to solving the
mystery of the “missile
sightings” was the use
was less useful than a similar ing it near where the eyewitness
description that referenced iden- of sound-propagation claimed the fireball disappeared,
tifiable landmarks along the analysis… behind the second house.
beach. An airline passenger’s
report of a “flare-like object” was
less useful than a similar passen-
ger report that included the
observation that the flare-like
object appeared in the sky about
”
and disappeared. This observa-
tion lasted about 15 seconds. Two
A third important eyewitness
was the one in the small boat.
Based on his location with
respect to where the plane was
when it first exploded, analysts
10 seconds after a small aircraft or three seconds later, a fireball were able to calculate that the
flew under the plane on which appeared in the sky near where concussion sound took 42 sec-
the observer was traveling. the white light had disappeared. onds to reach him. So we knew
that the wing separation he
A few eyewitness reports proved The eyewitness specified a land- observed took place about 42 sec-
particularly useful. One of the mark—a house near the beach— onds after the explosion that
most valuable was from the eye- over which the white light first ended the onboard recordings.
witness in the beachfront appeared and a second house
condominium, even though his behind which the descending fire- The two distinct trails of fire he
observations began well after ball disappeared. We determined described, produced as the plane
Flight 800 first exploded. His precisely where the eyewitness and wing dropped to the water’s
report of loud sounds just after was standing, and then mea- surface, were observed by many
the fireballs hit the water made sured the line-of-sight angle eyewitnesses and often were
it possible to calculate the between the eyewitness and each described as appearing immedi-
elapsed time from when the of the two houses. ately after the “flare-like object”
plane first exploded to when it considered by some to be a possi-
hit the water. From radar data Calculations based on the flight ble missile. But because the trail
and onboard recordings, we knew path of Flight 800 with respect to of fire from the flare-like object
where the plane was when it first this eyewitness show that when culminated in this second explo-
exploded. Because we also knew the aircraft first exploded, it was sion, the flare-like object cannot
where the observer was located, just passing over the house above have caused the explosion that
it was possible to calculate how which the white light first ended the onboard recordings—
long it took sound to travel from appeared. So the white light the that is, the first explosion. It can-
the explosion to the observer (49 eyewitness described probably not have been, as some have
seconds). So his statement estab- was the aircraft briefly ascend- suggested, a missile attack.
lished that the aircraft exploded ing and arching over after it
about 49 seconds before it hit the exploded, rather than a missile
water. attacking the aircraft. Explaining the “Missile
Sightings”
Another excellent eyewitness on This eyewitness’s visual observa-
the land, whose observations tions were consistent with the The key to solving the mystery of
began early and included several aircraft’s known horizontal the “missile sightings” was the
identifiable reference points, motion during the 49 seconds use of sound-propagation analy-
reported that he watched a white which elapsed after the onboard sis to establish that eyewitnesses
light, perhaps a firework, travel- recordings ended. Radar data who appeared to have seen a mis-
ing upward at a steep angle with show that during this time the sile “streak up” and cause the
respect to the horizon. The light aircraft traveled about 15 plane to explode could not have
“zig-zagged” as it rose, and at the degrees from right to left with seen such an occurrence. Once it
apex of its travel “arched over” respect to this eyewitness —plac- was established that the aircraft

5
TWA Flight 800

exploded about 49 seconds before

it hit the water and that one
wing detached (producing two
discrete fireballs) about seven
seconds before water impact, we
could be confident that eyewit-
nesses describing “streaks” that
began just a few seconds before
the appearance of the two fire-
balls could not be describing a
missile approaching an intact
aircraft.5 The plane had exploded
before their observations began.
What they were seeing was a
trail of burning fuel coming from
the aircraft. Nonetheless, some
eyewitnesses interpreted the
“streak” produced by the burn-
ing fuel as a possible missile.
This interpretation no doubt was
reinforced in their minds when
the streak suddenly culminated Reconstructed wreckage, Calverton, Long Island, 6 February 1997 (CIA photo).
in an explosion.

Other descriptions from eyewit- agrees with the direction that there would be no separate air-
nesses who for whatever reason Flight 800 is known to have craft to see.
did not report hearing sounds been traveling when it exploded.
supported this conclusion. For So the flare-like object he saw
example, a passenger on USAir almost certainly was Flight 800 Meteorological Conditions
Flight 217 reported seeing an just after it exploded, not a
aircraft fly under him 10 sec- missile. Because the speed of sound is
onds before the appearance of “a affected by altitude, wind, and
small, flare-like projectile travel- Another factor corroborating this temperature, we had to correct
ing in an east-northeasterly theory was that eyewitnesses for these meteorological condi-
direction.” Radar tracking of tions. For example, the speed of
who suspected that they had
Flight 217 and the small air- sound at the 13,800-foot altitude
watched a missile destroy an air-
craft—later confirmed to be a where Flight 800’s CVR and FDR
craft were puzzled that they had
Navy P-3 Orion—shows that he ceased operating was 1,062 ft/sec
not actually seen the aircraft
first saw the flare-like object at (without wind). At sea level, it
before the missile hit it. Consid- was 1,116 ft/sec (without wind).
about the time Flight 800’s CVR
detected an onboard explosion. ering the lighting conditions at It averaged 1,089 ft/sec (without
He also specified where the the time—just at sunset—the air- wind). Wind slowed the arrival of
flare-like object first appeared, liner should have been visible to the sound from the initial explo-
which coincided with Flight any observer witnessing a mis- sion as much as 0.9 seconds for
800’s location when it exploded. sile approach it. But if burning some key eyewitnesses compared
And his statement that the fuel from the crippled aircraft to when they would have heard
flare-like object was traveling in was what they were seeing— the sound without wind. Inde-
an east-northeasterly direction rather than a missile streak— pendent wind-speed calculations

6
 TWA Flight 800

were done for each eyewitness boat several miles offshore. Oth- 800 was when its recorders
claiming to have heard sounds ers question whether the CVR ceased operating, and loud
associated with the disaster. Nec- would capture the sound of a enough to shake a 70-ton bridge
essary temperature corrections— warhead detonation, if one had at that distance. Based on
other than those in the altitude- occurred. Still others question descriptions by NTSB investiga-
sensitive calculations mentioned how we could include the obser- tors—including the judgment
above—were negligible. vations of eyewitnesses traveling that the center fuel tank was
in other aircraft in our analysis, sealed and contained hot fumes
because such eyewitnesses would when it exploded, and the judg-
Satellite Data be unable to hear the sounds ment that the power source to
heard by several of the eyewit- the CVR and FDR was severed
An orbiting sensor on a US satel- nesses on the land and water. almost immediately by this enor-
lite detected a heat plume mous explosion—a high-energy
associated with the crash of TWA To analyze accounts from eyewit- explosion in this tank appears to
Flight 800. Combining timing nesses who did not report be the only plausible source for
data derived from the CVR, FDR, hearing sounds, we used eyewit- the initial sound heard by
and this satellite, along with nesses who did hear sounds to eyewitnesses.
sound-propagation analysis help establish when key events
based on observations made by took place, and also to determine Concerning the issue of whether
the eyewitness on a small boat, the approximate time that the sound could have come from
we were able to determine that elapsed between these events. a shoulder-launched surface-to-
the aircraft’s left wing detached We then analyzed other eyewit- air missile, the sound produced
—creating two distinct fireballs— ness statements in the context of by the warhead of such a missile
just before the satellite’s sensor this template, and generally would be roughly equivalent to
detected the large heat plume. found them to be consistent. the sound produced by a
Although the satellite data cor- handgrenade. Because Flight 800
roborated our analysis, it was not The loud sound almost certainly exploded in thin air nine miles or
crucial to our conclusions. was produced by the explosion more from most eyewitnesses,
that severed the power source to the chance that any eyewitness
the CVR and FDR at the begin- would have heard such an explo-
The Loud Sound Controversy ning of the catastrophe. If the sion even if one occurred is
sound had originated near the remote. And it is inconceivable
Some have questioned why we aircraft at an earlier time, it is that such an explosion could cre-
believe that the loud boom heard likely that it would have been ate a concussion sound capable of
by many eyewitnesses corre- recorded by the CVR. If it origi- shaking a 70-ton bridge 11 miles
sponded to the initial explosion of nated later, our estimate that the away.
the center fuel tank, and how aircraft took about 49 seconds to
this sound could be heard by eye- hit the water after the initial Eyewitnesses probably also could
witnesses miles away. They explosion would increase, mak- not have heard the launch of
suggest that these eyewitnesses, ing it even less likely that any such a missile from a boat sev-
rather than hearing the explo- eyewitness saw a missile cause eral miles offshore. The launch of
sion of the center fuel tank, may the initial explosion. a shoulder-fired missile is qui-
have heard the detonation of the eter than the detonation of its
warhead of a shoulder-launched The magnitude of this sound was warhead. If such a short-range
surface-to-air missile at or near enormous—loud enough to be missile were launched close
the position of Flight 800. Or per- described as a “concussion sound” enough to be heard by eyewit-
haps they heard the launch of by eyewitnesses more than 11 nesses on the shore, it would not
such a missile originating from a miles away from where Flight have been able to reach Flight

7
TWA Flight 800

EXCERPTS FROM CONVERSATIONS BETWEEN AN AIR TRAFFIC

CONTROLLER, THE FLIGHT 800 PILOT, AND A PILOT IN THE CRASH
VICINITY—17 JULY 1996 (ALL TIMES APPROXIMATE)

8:30:14 p.m., Boston Air Route Traffic Control Center: TWA eight hundred, climb and maintain one five thousand
[15,000 feet].

8:30:17, TWA Flight 800: TWA’s eight hundred heavy, climb and maintain one five thousand, leaving one three
thousand.

8:31:12: [TWA Flight 800 explodes at an altitude of 13,800 feet, based on post-crash analysis.]

8:31:50, Eastwind Airlines Flight 507: We just saw an explosion out here on Stinger Bee five oh seven.

8:31:51: [Infrared sensor aboard US satellite detects large heat source in the vicinity of Flight 800 crash.]

8:31:57, Boston: Stinger Bee five oh seven, I’m sorry. I missed it. Ah, you’re on eighteen. Did you say something else?

8:32:00: [TWA Flight 800 hits water, based on post-crash analysis.]

8:32:01, Eastwind 507: We just saw an explosion up ahead of us here something [like] about sixteen thousand feet or
something like that. It just went down—to the water.

8:32:56, Boston: TWA eight hundred, [call] Center.

8:33:04, Boston: TWA eight hundred, Center.

8:33:09, Boston: TWA eight hundred, if you hear Center ident[ify].

8:33:17, Boston: Stinger Bee, ah, five zero seven, you reported an explosion, is that correct, sir?

8:33:21, Eastwind 507: Yes sir, about, ah, five miles at my eleven o’clock here.

8:33:48, Eastwind 507: [unintelligible] Stinger Bee, ah [unintelligible] Boston, we are directly over the site where
that airplane or whatever it was just exploded and went into the water. [Then, from a second operator...]
[unintelligible] eighteen, ah, nineteen miles on the two thirty-six radial [unintelligible] Hampton.

8:34:01, Boston: Roger that. Thank you very much, sir, we’re investigating that right now. TWA eight hundred, Cen-
ter. TWA eight zero zero, if you hear Center, ident.

8:35:36, Boston: TWA eight hundred, Center.

8:35:43, Eastwind 507: I think that was him.

8:35:45, Boston: I think so.

8:35:48, Eastwind 507: God bless him.

8:36:57, Boston: Stinger Bee five oh seven, thanks for that report, ah, New York on one three three point zero five
[133.05 MHz]. Good day, sir.

8:37:05, Eastwind 507: Thirty-three oh five, so long Stinger five oh seven. Anything we can do for you before we go?

8
 TWA Flight 800

8:37:11, Boston: Well, I just want to confirm that, ah, that you saw the, ah, splash in the water approximately, ah,
twenty [20 miles] southwest of Hampton, is that right?

8:37:20, Eastwind 507: Ah, yes sir. It, it blew up in the air, and then we saw two fireballs go down to the, to the
water and there was a big [unintelligible] smoke form, ah, coming up from that. Also, ah, there seemed to
be a light. I, I thought it was a landing light [unintelligible] it was coming right at us at, about, I don’t
know, about fifteen thousand feet or something like that, and I pushed my landing lights, ah, you know, so
I saw him, and then it blew.

8:37:40, Boston: Roger that, sir, ah, that was a seven forty-seven out there you had a visual on that. Anything else in
the area when it happened?

8:37:47, Eastwind 507: I didn’t see anything. He seemed to be alone. I thought he had a landing light on. Maybe it
was a fire, I don’t know.

8:37:52, Boston: Stinger Bee five oh seven, ah, roger that. Anything else comes to your mind, ah, you can use your
other radio, come back to this frequency and tell me about it.

8:37:59, Eastwind 507: That’s all I can think of at this time.

The Eastwind Airlines pilot, Capt. David McClaine, first reported an “explosion” about 40 seconds after Flight 800’s
CVR and FDR ceased operating. He most likely was reporting the conflagration produced when Flight
800’s left wing detached—producing two “fireballs”—roughly seven seconds before the burning debris hit
the water. This large heat source was detected by an infrared sensor aboard a US military satellite just
after McClaine made his first report.

The “landing light” he reported seeing earlier probably was a fire produced after the initial explosion and described
by some eyewitnesses on the ground as a “streak of light in the sky” that preceded Flight 800’s “explo-
sion.” It was this “streak of light” that led some people to think that a missile was used to shoot down
Flight 800.

Based on sound-propagation analysis—juxtaposing what eyewitnesses saw with what they heard—CIA analysts
concluded that this “streak of light” was, in fact, a trail of burning fuel from the Boeing 747 after the first
explosion had already occurred, not a missile.

9
TWA Flight 800

800. And the launch sound would more than 42 seconds before the evidently helped many better
not shake a 70-ton bridge, nor be aircraft broke into “two distinct understand the eyewitnesses’
heard by eyewitnesses up and fireballs” and more than 49 sec- observations. In a letter to the
down the coast of Long Island. If onds before the plane hit the Director of Central Intelligence
a loud sound associated with the water. CIA analysts are not dated 17 July 1998—the second
disaster were produced near the aware of any eyewitness who did. anniversary of the crash—Con-
shore, it would be heard almost gressman James A. Traficant,
immediately at high intensity by Jr., a member of the House
the closest eyewitnesses, and Final Observations Transportation and Infrastruc-
much later and at lower inten- ture Subcommittee on Aviation,
sity by eyewitnesses several Analysts often have to make recognized that work:
miles away. This would be incon- judgments based on limited and
sistent with descriptions given by highly conflicting information. I want to extend my gratitude
dispersed eye-witnesses along This was not the case for our to those Central Intelligence
the shore who all reported hear- assessment of the eyewitness Agency employees who worked
ing a loud concussion sound less reports associated with the crash on the analysis of the eyewit-
than five seconds after the burn- of TWA Flight 800. On the whole, ness statements. Their work
ing debris hit the water. the corroborative evidence that was extremely helpful in the
the eyewitnesses saw only the effort to unravel the Flight 800
Concerning the issue of whether burning aircraft without realiz- mystery and in addressing the
the sound of a missile warhead ing it, and not a missile, was controversy surrounding the
detonation would have been cap- extensive and compelling. eyewitness testimony. I think it
tured and recorded by the CVR, would be extremely beneficial if
we cannot be certain that such a Nevertheless, a few people, the CIA undertook an effort to
sound would have been recorded. driven by what they perceive to better educate the American
It also is not relevant to our con- be an overwhelming number of people on the work the Agency
clusion that the eyewitnesses eyewitnesses who “saw” a mis- did on the eyewitness analysis.
saw only the burning aircraft, not sile attack the plane, persist in
a missile. Unless such a war- thinking otherwise. Confident ❖❖❖
head detonation occurred well that so many eyewitnesses can-
before the center fuel tank not be “wrong,” they have
exploded, produced an enor- concluded that the government, Notes
mously loud sound, and created for whatever reason, is covering
warhead fragments that trav- up the true cause of the crash.6 1. All air traffic control communica-
eled faster than the speed of tions cited in this article were
recorded at the Boston Air Route
sound for a long time (so they These people probably are sim- Traffic Control Center on the
caused the center fuel tank to ply reading more into the evening of 17 July 1996 and sub-
explode before the warhead eyewitness descriptions than is sequently released to the public
sound could be recorded by the warranted.7 Nonetheless, they by the Federal Aviation Adminis-
CVR), this scenario would not are likely to cling to their theo- tration. (U)
change CIA’s conclusion that the ries. The crash drew worldwide 2. In mid-January 1997, CIA ana-
eyewitness observations took attention and did not have an lysts decided to document their
place after the aircraft exploded. immediate explanation. Such sit- work in an animated video. The
uations almost always give rise CIA analysts and their line man-
Any eyewitness who thinks he to suspicion and conjecture. agers agreed that such a
production was needed to make a
may have seen a missile shoot convincing case to nontechnical
down Flight 800 needs to have Despite the reservations of skep- audiences that eyewitnesses had
seen something that occurred tics, the work of Agency analysts seen only the burning aircraft. On

10
 TWA Flight 800

11 February 1997, CIA senior 4. Had Flight 800’s CVR and FDR 52 occasions in the past 16 years,
managers gave final approval and continued to operate after the ini- power failure has resulted in lost
allocated the necessary funding. tial explosion, the crash CVR or FDR data. In the case of
investigation would have been TWA Flight 800, power to both
Eventually, this video or seg- simplified considerably. the CVR and FDR was severed
ments of it would be seen by when the center fuel tank
several hundred million people CVRs have been required since exploded. This contributed to the
worldwide. Most responses, 1967 on all commercial airliners. extraordinary difficulty and
including comments from several They typically use a 30-minute, expense of that crash
family members of the crash vic- continuous-loop, 4-track tape that investigation.
tims, were favorable. But some records sounds detected by four
were not. A few people even sug- microphones: one in each pilot’s 5. Had Flight 800 gone down two to
gested that the video was headset, one in the instrument three miles farther offshore, the
fabricated as part of a govern- console, and one in the center of sound of the initial explosion
ment coverup of the true cause of the cockpit designed to monitor would have reached the nearest
the tragedy. overall sounds. The CVR often is observers on shore 10 to 15 sec-
the most useful source of informa- onds later than it did. This would
3. Late on 30 December 1996, CIA tion to determine the general have made the sound-propaga-
analysts first realized that eye- cause of a crash, but rarely tion analysis considerably less
witnesses probably had observed reveals the precise cause. conclusive, because the plane
only the burning aircraft in vari- would have hit the water well
ous stages of crippled flight. The In addition to providing dialogue before the first sound reached the
following morning, this conclu- and comments from the pilot, co- shore.
sion was phoned to the FBI. It pilot, and crew, the CVR serves
took about five more weeks, how- other vital functions. Analysis of
As it turned out, the plane
ever, for CIA analysts to sounds recorded by its multiple
exploded about 10 miles offshore
document a case solid enough to microphones can be used to deter-
and hit the water about 49 sec-
mine if sirens or alarms were
be formally briefed to the FBI. onds later, just as sound from the
activated, or if there was an
This briefing was given in New explosion was reaching the clos-
explosion. If an explosion
York on 6 February 1997. Details est eyewitnesses on the shore.
occurred, CVR analysis can help
were provided in writing on 28 This coincidence permitted sound
locate its source.
March 1997 in a memorandum propagation analysis to be applied
from the CIA’s Deputy Director FDRs have been required since effectively, because analysts could
for Intelligence to FBI Assistant 1958 on all commercial airliners. use it to establish that so-called
Director James Kallstrom. On 18 They typically use a 25-hour con- missile sightings described as
June 1997, CIA analysts briefed tinuous-loop tape that records occurring shortly before the fire-
their results in New York to Kall- precise flight conditions and diag- balls appeared and plunged into
strom and about 30 others from nostic information on the the water took place well after the
the FBI and the Army’s Missile operation of at least 29 of the air- plane had already exploded.
and Space Intelligence Center craft’s critical subsystems. These
(MSIC). They also showed an include thrust from each engine, 6. For example, an NBC Dateline
early version of their video (TWA altitude, air speed, compass head- poll of 503 Americans taken on 14
Flight 800: What Did the Eyewit- ing, vertical speed, horizontal March 1997 showed that 48 per-
nesses See?) at this time. On 22 speed, pitch, and roll. Analysis of cent believed that the government
October 1997, CIA analysts these data is more time-consum- was covering up the real cause of
briefed Kallstrom in New York ing and specialized than analysis the crash.
again and showed him a more fin- of the CVR data, but usually pro-
ished version of the video. He vides vital clues concerning the 7. Although many eyewitnesses
expressed his appreciation for specific reason for a crash. described seeing something akin
CIA’s help and his desire to use to a flare or firework traverse the
the CIA video the following month Because the CVR and FDR are in sky and culminate in an explo-
at his news conference announc- the tail section of the plane (to sion, the vast majority of them did
ing the suspension of the criminal maximize survivability) and the not refer to what they saw as a
investigation. CIA concurred and power source is in the front, “missile.” They repeatedly used
prepared the video for public power can be severed in the event the descriptive terms “flare” and
release. of a breakup of the fuselage. On “firework.”

11