Rubber-Stamping Flight 800
WASHINGTON, D.C.As the National Transportation Safety Board's recent public meeting to discuss its final report on the July 17, 1996, crash of TWA Flight 800 drew to a close, one could almost hear the thunk. Transforming itself into a gigantic rubber stamp, the bureaucracy validated the theory that a mechanical failure, not a missile, brought down the plane. The NTSB, along with every agency with clear jurisdiction over the crash, and a few others besides, had spoken with one voice. The $50 million investigation, if not the controversy, was over.
It was no surprise that the NTSB's technical staff had decided that the event that tore apart the Boeing 747 and sent 230 people to their deaths over the ocean eight miles south of Long Island was "an explosion of the center wing tank resulting from the ignition of a flammable fuel/air mixture. The source of ignition could not be determined with certainty." Yet the NTSB had not succeeded in papering over the cracks in its investigation. During two days of discussion, at August 22 and 23 board meetings, the NTSB staff . . .
- failed to explain some mysterious dark brown spongy material, called "splatter" in an NTSB report, found in several places, including on top and inside the center wing tank;
- disregarded results of their own tests and paid scant attention to more than 30 years of government research into fuel flammability;
- failed to explain how the pressure from a fuel tank explosion could have broken through the much stronger fuselage skin and split the plane apart;
- accepted only eyewitness accounts that, they said, conformed to their theory of how the plane broke up, and simply dismissed the rest by claiming that witnesses' memories must have played tricks on them;
- failed to share the results of analysis done on the brief, loud sound signature at the end of the cockpit voice recorder tape.
The splatter issue is one more anomaly in an investigation that has shrugged off many. On the 22nd, in answer to a board member's question, NTSB director of aviation safety Dr. Bernard Loeb admitted that some investigators had questioned what could be the source of the brown specks of spongy material. Tests determined that temperatures above 490oF had melted plastic foam insulation from around an air-conditioning duct that runs fore and aft outside the tank, over the top of the left side of the tank roof just below the passenger cabin floor. After extensive analysis, Loeb said, "We determined it was fully consistent with the scenario we have given you [i.e., with the breakup of the center wing tank]."
But the "Splatter Deposits Study," a never released report obtained by the Voice, appears to challenge Loeb's statement. The report says the melted foam was splattered over an eight-foot-long section of the tank roof and inside the shattered air-conditioning duct. The splatter was thrown forward with enough force to break through the roof of the tank, landing on two fragments of the front spar (between the wings and under the seats) that were ejected very early in the breakup sequence.
Splatter was also found on the fuselage and floor structure just in front of the front spar, and on passenger seats above. The investigators "explored possible airplane sources of higher than normal heat incoming to the area of the splatter," and found none.
In introducing its own endeavors, the NTSB in Washington ignored existing research into Jet A aviation kerosene. "Basic information was not available when we began," said Dr. Joseph Kolly, an NTSB engineer.
Kolly's statement was baffling, because in fact there is a wealth of data on the flammability characteristics of aviation fuels, including Jet A, available in studies published by the navy, the air force, and a body called the Coordinating Research Council. Many of these studies are even referenced by the NTSB in its reports. For some reason, though, the NTSB felt compelled in Washington to present its own prolific work on Jet A as groundbreaking.
Some of the claims Kolly made for that work do not stand up under scrutiny. For example, when he displayed a slide illustration showing temperatures recorded at many locations inside the center tank of a 747 used for a so-called TWA 800 Emulation Flight test, it was noticeable that none of the temperatures recorded at the altitude at which the plane exploded were below 110oF. But temperatures that appear in the published version of that illustration are as low as 101oF. Kolly said "the average high temperature" inside the tank at the altitude of the explosion was 120oF. But the average inside the tank, going again by the NTSB's published figures, was actually 111.5oF.
These temperatures are significant because they are central to the NTSB case that the center tank exploded and caused the accident. Lacking any evidence of an ignition source, the Safety Board did the flight test and other tests to show that the atmosphere inside the tank, with possibly 50 gallons of fuel, was flammable.
But even if a flame is produced, that does not necessarily mean an explosion will follow. "The mere existence of a normal flame does not by itself imply the beginning of an explosion," notes a NASA addendum to the NTSB's Systems Report.
A spark is just one link in a chain. For an explosion to happen, a spark has to ignite all the fuel-air mixture in a very rapid burning reaction. But this cannot be counted on, especially in a large container like TWA 800's center tank, where there was so little fuel to begin with, and where temperatures varied widely. In such a situation, with some temperatures hovering down near 100oF, a small spark may simply have puffed and gone out, according to Kurt H. Strauss, a nationally recognized aviation fuel expert. Whether an explosion happened "would depend on total energy released when that ignition goes off," said Strauss.
Given that the tank exploded, the NTSB's rationale connecting the blast to the fuselage damage rests on surprisingly shaky ground. According to Jon Hjelm, a Federal Aviation Administration engineer and member of the Sequencing Group, who contributed pages of stress calculations as a kind of reality check on the deliberations of the group, an enormous force, equivalent to more than the thrust produced by one of the 747's jet engines, acted on the bottom of the tank to produce the cracking that sundered the fuselage in front of the tank. Hjelm said he came up with his figure for this force using assumptions he made about the distribution of the pressure from the explosion.
Hjelm said that in order for his calculations to confirm the breakup sequence, he made another assumption. The force had to remain, pushing down inside the tank, after the explosion had ruptured the front of the tank. For how long? "Maybe some number of seconds. Way more than one second," Hjelm said. What if the the pressure all dissipated in under one second? Hjelm said, "That's a question I feel uneasy to deal with." Yet according to the results of the NTSB's own explosive testing and scientists questioned by the Voice, not to mention the account of the breakup sequence given by Sequencing Group chairman Jim Wildey at the NTSB's Baltimore hearings in 1997, the initial explosion was certainly over within one second.
Wildey said in Baltimore that the Sequencing Group had relied upon Hjelm's calculations, which constitute apparently the only available engineering analysis of the breakup sequence. Boeing (according to its submission to the NTSB) did not complete a project to create a computer model of the breakup.
Dr. David Mayer, who holds a doctorate in applied experimental psychology, drew on studies by psychologists to suggest that influences acting on eyewitnesses in the aftermath of the crashthe chatter of friends, TV reports, even the leading questions of the FBI agents who interviewed themmight have led them to embellish their memories of the crash.
And when they reported seeing the initial explosion, which the NTSB says was contained "inside an intact airplane," and thus could not have been visible to witnesses miles away, that was understood by investigators not as a reason to reexamine their theory, but to assume the witnesses must have been mistaken. Thus the accounts even of seasoned airline pilots who reported to air traffic control (ATC) in the first moments that they saw the plane explode were essentially discounted.
In this regard, there is an unexplained apparent discrepancy between the account of the pilot who first reported an explosion and the official ATC transcript. According to the transcript, Captain David McClaine of Eastwind Airlines reported the explosion at 8:31 and 50 seconds, which is 38 seconds after the NTSB says the plane exploded, at 8:31 and 12 seconds. But McClaine states in a written account he gave the Witness Group that after the explosion he "immediately called Boston ATC and reported an inflight explosion out over the water."
When he was questioned by the group, he said in answer to a question from Mayer that roughly 10 seconds passed after the explosion before he made his first radio call to ATC. Ten seconds seems a reasonable pause between seeing something so dramatic and doing something about it. Thirty-eight seconds appears rather long to wait before making the call to air traffic control. But the apparent time lag could be used to support the notion, suggested by both the NTSB and the CIA, that McClaine, for example, saw only a later stage of the airplane's breakup, not the initial explosion.
As for those witnesses who saw a rising streak of light, Mayer said they probably saw the burning plane climb after the explosion. When the chairman asked him, if the plane did not climb, would that affect his analysis? Mayer said no, it wouldn't. "But we believe it climbed," he said. However, McClaine had been closely questioned about this very issue by the Witness Group, and repeatedly said the plane did not climb; he saw only falling debris after the explosion.
Investigators said, when they first examined the two "black boxes" a week after the crash, that neither device provided clues to explain what what happened. But there is indeed a marked sound signature, lasting less than two-tenths of a second, on the cockpit voice recorder (CVR) tape. That signature has never been explained, leading the Air Line Pilots Association in its submission to the NTSB to bemoan the lack of follow-up after a series of tests done in Bruntingthorpe, England, in 1997. The submission notes that research done at the University of Southampton shows that analysis of a sound signature can yield information about the type of explosion (whether a high explosive detonation or a lower-energy fuel-air explosion) and its location within the fuselage.
Jerry Rekart, a clearly unhappy ALPA chief investigator and a TWA pilot, told the Voice he is still frustrated at the NTSB's failure to publish any details of its sound spectrum analysis, as it is called. A Sound Spectrum Group met once, before Bruntingthorpe, but never since, he said, despite repeated requests from ALPA. The data recorded at Bruntingthorpe was sent to Southampton, the ALPA submission says, but "the Sound Spectrum group has never been briefed regarding the analysis of the data completed by the University of Southampton, nor has the group met to finalize any type of report of its activities in relation to the investigation of TWA 800."
But at last month's board meeting James Cash, the NTSB engineer responsible for that analysis, strove to give the impression that no questions remain about the CVR sound. "We did all we could to glean whatever we could get out of the cockpit voice recorder," he said. Responding to Cash, board member John J. Goglia added, "Every single thought that every single person has on our teams gets explored. We don't overlook anything. I visited you in your lab, and never once did I leave with the impression you haven't turned over every stone."
Not everyone shares Goglia's rosy view of the NTSB's investigation to determine the cause of the crash. Michel Breistroff, whose son was killed on Flight 800, might have been speaking for them all when he said, "What I need is that someone from the team will tell me, 'That's the wire [which caused the explosion].'
"We the families are not technical," he told the Voice. "We need true evidence, not only words."
Get the This Week's Top Stories Newsletter