Home Page

June 25, 2002
Flight 587 Update - Witness Reliability
by Victor Trombettas
The New York Times ran a piece on June 23rd related to the Flight 587 investigation titled "For Air Crash Detectives, Seeing Isn't Believing" making the case that in addition to offering little if any useful information, eyewitnesses offer such contradictory accounts, as to render them useless.
Since TWA Flight 800, it's long been known by those close to the NTSB, those who have read their TWA Flight 800 Final Report, and those who listened to the NTSB at Press Conferences after Flight 587 crashed in Belle Harbor, that witnesses are viewed by the NTSB as "notoriously unreliable".
On June 4th ... the NTSB released statistics categorizing the 349 witness statements they had received from witnesses to the crash of Flight 587.
Item 1: 52% specifically reported seeing a fire while the plane was in the air, with the fuselage being the most often cited location (22%). Other areas cited as a fire location were the left engine, the right engine or an unspecified engine, and the left wing, the right wing or an unspecified wing.
In other words, 181 people saw a fire while the plane was in the air. 168 did not say they saw a fire.
On the surface, this appears to be a discrepancy. But a good investigator would speak to many witnesses in person, at the same location where they spotted the plane, and try to ascertain as best they can the position of the plane and how long the witness observed the plane in flight, etc. Bringing a model Airbus A300 along would enhance the interview experience. A good investigator understands that witness sightings occur at different angles and different times and this can have a significant impact on what people see and do not see. Especially in the bright morning sun. If all this is done diligently, patiently, and faithfully, what emerges is that it is extremely unlikely that a large group of people will say they saw something occur that indeed did not.
As best we can determine, no witnesses have been interviewed at the location where they witnessed the plane in flight. Even more disturbing, most have not been interviewed in-person even via telephone by NTSB investigators.
The NTSB witness statistics released on June 4th are based mostly on a combination of "summaries" taken by local law enforcement officers the week of the crash and responses to a questionnaire the NTSB mailed to witnesses beginning in January. Not an exhaustive look at the witnesses by any means. Especially when you consider that the NTSB written questionnaire consisted of 5 questions. My wife had to answer more questions when she took our cat to the Vet last week.
I have interviewed witnesses who saw fires and those who did not. The first clue to unraveling this apparent discrepancy comes from those who do report fires. The majority of them state the fires did not persist for the entire time they witnessed the flight. In other words, there were times when fires were visible and times when they were not. Also, witness John Power made it clear to me that when he first saw Flight 587, it was trailing black smoke but he could see no fire. Only as the plane started to pass more from his left to right and he could see more of the right and right rear of the craft, did he notice fire. In other words, the fire was visible from certain angles, but not from all. John witnessed at least the last 25 seconds of flight. He witnessed fire and a heavy, steady smoke trail, and the plane fly out of control into a roll and flat spins before any parts separated from the craft.
76 individuals said they saw fires in the fuselage. Other areas cited as a fire location were the left engine, the right engine or an unspecified engine, and the left wing, the right wing or an unspecified wing. Well is it left or right? Again, all depends on angles. Also depends on when they saw the plane. Maybe people are calling their left the plane's left when they really should have said the plane's (or Pilot's) right. You can only begin to make sense of this with one-on-one interviews and a model airplane. A simple and unsophisticated technique.
An unfortunate missing piece from the NTSB's witness statistics is "how many witnesses spotted fires or explosions before the separation of any parts from the plane". That statistic is not there because that question was not asked on their questionnaire.
Item 2 from the NTSB statistics: 28 people saw an explosion. That would mean 321 did not. Does that mean there was no explosion? If there was an explosion, it would only last for a second or two. Flight 587 was in distress up to 38 seconds before impact. If the explosion only lasted for a second ... then 37 of the 38 seconds of distressed flight contained no explosion. It is not difficult to understand then why few witnesses would see an explosion if it did occur.
Item 3: 70 individuals specifically reported seeing no fire at all. That means 279 individuals saw a fire or didn't exclude it.
Item 4: 77 reported observing smoke; 70 reported no smoke.
Item 5: 63 reported observing the airplane in a right turn; another 63 reported observing the airplane in a left turn. After interviewing dozens of witnesses, this is no discrepancy. The plane did turn right and left on it's way down. Having lost altitude, airspeed, parts of the rudder, the tail, both engines, parts of at least one wing, there were periods of unstable flight, to say the least, including times when witnesses say the plane spun or cork-screwed down. I've interviewed witnesses who specifically saw the plane roll right and left. Some refer to a roll as a turn. That does not discredit what they saw.
Item 6: 45 observed the airplane "wobbling," dipping" or in "side to side" motion.
Item 7: 258 observed the airplane descend.
Item 8: 199 reported seeing "something" separate from the airplane; 45 reported observing the right wing, left wing or an undefined wing separate; 31 specifically reported observing no parts separate. The NTSB has clarified this item a bit. There were witnesses who positively identified the tail and engines come off in flight. The NTSB lumped all those into the "something" category. What about the accounts of wings separating? This is not part of the official record, is it?
The vertical stabilizer, after separating from the craft and falling to the Bay could easily be mistaken for a wing. Secondly, at least one of the wings was breaking off in-flight based on numerous witness accounts and more importantly, the debris field in the Bay and on land. Ailerons (control surfaces on the wings that are used to turn the plane) were recovered in the Bay, according to a source within the investigation. The left wingtip was recovered 6 blocks east of the crash site. The Rockaway Diner Sunset Diner, 3,900 feet east of the crash site, had two large panels (each about 2 by 5 feet) land on the roof. Those panels were "finished" on the inside, as Diner owner Steve Good related. They looked like the interior of the plane on one side; a "formica" look. The police officers who picked up the two panels from the Diner two days after the crash told Steve Good they were from the "fuselage". Here's a picture of the left wingtip 6 blocks east of the crash site:
When some of the witnesses saying they saw a wing separate ... they may not be too far from the truth.
Investigators say there is no evidence in the wreckage or on the flight recorders of an in-flight fire or explosion. This implies that all the wreckage was recovered, that every last piece of wreckage was scanned for explosives residue, and that recorders can positively identify every type of in flight fire or explosion.
Was every last piece of wreckage recovered? The NTSB and FBI have not reconstructed the craft in the way they did TWA Flight 800 because the nose-first crash of Flight 587 (and fireball) onto land severely damaged the fuselage. Therefore, it is difficult for them to determine that they have recovered every last piece of wreckage. I had informed the NTSB months ago that I had at least one report of a Rockaway resident "saving" a piece of the wreckage that landed near their home many blocks west of the crash site. One man was arrested for saving pieces of the wreckage. A few months ago, in attempting to determine the location of the debris in the Bay , I interviewed officers at the NYPD Harbor Unit, the group that recovered the plane's wreckage from the Bay and had divers in the water looking for debris with metal detectors. I was aware that the earliest witness accounts of explosions and fires in flight occurred when the plane was still at least two miles, or 3500 yards north of the seawall. Therefore, if there had been an in-flight fire or explosion, the possibility existed that some debris may have settled to the floor of the Bay up to 3500 yards out in the Bay. How many yards out from the seawall did the Harbor Unit scan? Less than 1,000 yards. A very large area was not scanned. When I brought this to the attention of the Harbor Unit's leadership, they were willing to go back out (months later) and scan everywhere. They just needed word from the "Feds" to do that. Apparently, word never came. What apparently did come was a firm rebuke to them by the "Feds" for even speaking with me.
Was every last piece scanned for explosives residue? The NTSB told me they do not know what tests the FBI ran and they (the NTSB) do not run explosive residue scans; but they do know how to identify (visually) explosive damage when they see it. I started asking the FBI in January ... "did you scan the fuselage for explosives residue?" No answer. I'd leave message after message. Finally, this is the answer I received about a month and a half ago: "agents of the New York FBI and Lab were on hand ... they conducted an investigation (including tests by lab personnel) they deemed necessary to reach a conclusion ... but we will not identify the specific tests done". Not exactly a clear answer. In fact, the words "they deemed necessary" could imply that they did less than could have been done because they "deemed" more tests unnecessary. Was every last piece scanned for explosives residue? If the FBI won't answer this simple question in the affirmative, then I have to assume they did not. Added to that the possibility that some wreckage may never have been recovered and I ask ... how can anyone completely rule anything out?
Can the flight recorders identify every type of inflight fire or explosion? No, they cannot. In-flight fires (below the passenger cabin) might not appear on the Flight Data Recorder. The bombing of Pan Am 103 proved that with a small amount of explosives (no more than 1 pound) you could bring down an airliner. Just one month after Flight 587, there was Richard Reid (the Shoe Bomber), with an even smaller amount of explosives in his shoes. How would a low weight explosive appear on the recorders? We really don't know. There was never any consensus that the Pan Am 103 explosion was recorded on the Cockpit Voice Recorder (CVR).
Some (possibly many) of the witness accounts of in-flight fire can be explained by the departure of the engines from the plane just seconds from impact. But there are strong witnesses (retired and active police officers, retired firefighters, Professionals, etc.) who witnessed the flight 20 plus seconds before the engines departed and are adamant that the engines and tail were intact when they noticed fire, smoke, or an explosion. These in-flight fires could be the result of compressor stalls, engine fires that result from too much fuel and not enough air intake. But thus far, word from the investigation is that compressor stalls or surges have not been identified on the flight recorders. What I find absolutely incredible is the Times' assertion they the NTSB investigators have somehow determined that "the idea that the plane caught fire is a trick of memory". How anyone can determine that all 181 witnesses suffered from a trick of memory without being interviewed is beyond me. But more importantly, that comment is beyond the realm of a scientific and objective investigation.
The New York times article quoted Dr. Charles R. Honts, a professor of Psychology at Boise State University and the editor of the Journal of Credibility Assessment and Witness Psychology as saying "Eyewitness memory is reconstructive," said Dr. Honts, who is not associated with the safety board (NTSB). "The biggest mistake you can make is to think about a memory like it's a videotape; there's not a permanent record there." The problem, he said, is that witnesses instinctively try to match events with their past experiences: "How many plane crashes have you witnessed in real life? Probably none. But in the movies? A lot. In the movies, there's always smoke and there's always fire." I practically fell over when I read this. Because 181 witnesses have watched action movies with plane crashes ... 181 will add fire and smoke into an experience when it wasn't there? I've left Dr. Honts a voice mail. I have to hear him tell me this in his own words. What's amazing is how some of those 181 witnesses added the fires in at exactly the same point in the flight path. They must have seen the same movie.
Though the NTSB has tried to pretend they value witness statements, their behavior and comments have clearly indicated otherwise. And the well-respected New York Times Aviation Writer, Matthew Wald, writing from firsthand knowledge reveals, "the safety board (NTSB) generally doesn't place much value on eyewitness reports if data and voice recorders are available."
I have spoken with aviation experts, including engine (powerplants) experts who have told me it is possible a compressor stall fire might not appear on the flight recorders. Even the NTSB has admitted the same. But a "notoriously unreliable" witness can tell you ... "there was fire coming from the engine" (i.e. a possible compressor stall) and give an investigator another piece to the accident puzzle one would think they would like to have access to. So who is more reliable when it comes to identifying compressor stalls? In certain cases, witnesses.
The most scrutinized pieces of wreckage in Flight 587 are the rudder and vertical stabilizer. Can the witnesses offer any information to enhance or even add to the knowledge provided by the flight recorders as to the fate of these pieces? Absolutely. Here's how:
The NTSB tells us the rudder position data in the FDR (Flight Data Recorder) became "unreliable" 2.5 seconds before the FDR died prematurely 10 seconds before the crash. NTSB has stated rather definitively, that the rudder data becoming "unreliable" meant the entire vertical stabilizer and rudder had separated from the plane. Wait just a second here. How do you get from A to Z like that? The NTSB hasn't elaborated what they mean by "the rudder data became unreliable". But a reasonable interpretation is that there was some data there, just not the sort of data you'd usually see. The rudder's position data is transmitted by a sensor in the lower section of the rudder, through the base of the vertical stabilizer. How has the NTSB proven that losing this sensor altogether (which is what would happen when the vertical stabilizer ripped away from the plane) translates into "unreliable data"? We don't know that they have. How can a witness help? John Power stated that after the plane had lost flight control, and started to fall out of the sky, the plane lost a piece about the size of a cargo door while the vertical stabilizer was still attached. It wasn't an engine he saw fall away and flutter down to the Bay. It was a flat piece. Two of the larger pieces of the shredded rudder are about the size of the cargo door. The debris field supports this possibility ... several pieces of the rudder were recovered at least 600 feet from the vertical stabilizer "all over the Bay" according to officers who were on the recovery boats. The possibility exists ... in correlating the FDR's rudder data with witness accounts... that the rudder broke way before the vertical stabilizer separated. The flight recorders can't tell you that. Witnesses might be able to help in that regard. You would think investigators would welcome this information. This could be evidence (along with the photos of the rudder's total destruction and the FDR data) that the rudder was the first structural component to fail.
The New York Times article continues, "The safety board's (NTSB's) skepticism of eyewitness accounts was deepened by the explosion of TWA Flight 800 off Long Island six years ago: hundreds of people saw an upward streak that they assumed was a missile, although investigators said it was the body of the plane itself, streaking upward (thousands of feet: emphasis mine) after the forward portion had fallen off following a fuel tank explosion."
The CIA produced a video that illustrated this aerodynamically-incorrect hypothesis. The NTSB then did their own video. When asked to provide the calculations used to create this "animation" the CIA stated they received the calculations from the NTSB. Boeing said they had no idea how that video was produced. The NTSB said they couldn't release the calculation data because it contained "data proprietary" to Boeing and was, and is, a "trade secret". Pilots and aviation experts have calculated as best they could that the headless-747 climbing thousands of feet like a missile is an aerodynamic impossibility given that the center of gravity shift (since the front portion of the plane had fallen away) would have thrown the plane into an aerodynamic stall in just a few seconds and after only a couple of hundred feet of ascent.
Flight 800 is the incident the NTSB now uses as the proof for deeming witnesses unreliable?? If it wasn't for that lone CIA analyst who came up with this headless-ascent theory, I wonder if the NTSB would have said the hundreds of Flight 800 witnesses had all imagined the ascending missile-like trail because they had suffered from a mind trick. They had seen missile attacks in movies. Many experts believe the CIA video, the NTSB's support of that video, and the NTSB's refusal to release the supporting calculations, raise serious questions as to "reliability", not of witnesses, but of our country's intelligence and investigative agencies.
The Times quotes a current NTSB Board Official as saying, "Can you imagine if we didn't interview the witnesses?" I interpret that as "can you imagine the heat we'd take if we didn't go through the motions of interviewing the witnesses?" The NTSB's own spokesman revealed his agency's prejudice towards witnesses in the article ... "I don't think I'm making any news by saying that eyewitness testimony at a plane crash and probably at many traumatic events is unreliable."
Try and follow this logic: Mr. Berman is quoted in the Times article as saying, "Witness statements can be more valuable in crashes of small planes that don't have flight data recorders or cockpit voice recorders". It was established earlier in the Times article that witness statements are unreliable because memories are corrupted. Their minds play tricks on them. They've watched too many movies. But the witnesses become valuable when we have crashes of small planes that do not have on board recorders? The witnesses intuitively sense, "this plane had no recorders ... I'd better cleanse my mind of all self-trickery and cinematic influence ... the NTSB needs me on this one"?
Am I saying there are no witness discrepancies? Not at all. They do exist. But not to the obscenely exaggerated level that has been suggested by witness critics.
The NTSB's own statistics prove that the Flight 587 witnesses were generally careful to not affirm something they were not sure of. For example, Firefighter Tom Lynch has always been adamant he saw a major explosion or eruption while the tail and engines were attached (he saw two explosions, the first fairly small). But he's not sure what happened after that. He recalls a wing or the tail flying off but he's not 100% certain what it was. The very fact he can say "I'm sure of this, but not sure of that" makes him a credible witness. Police Officer Conrad had the same statement. Certain he saw explosive flames and smoke, uncertain of what debris he saw falling from the plane. Most of the witnesses are similarly careful to only affirm what they are certain of. Diner owner Kenny Good (Steve's brother) was aware that the attention-getting witnesses were those who had seen fires or explosions. But he just doesn't remember seeing fires or smoke. He had a very different view than Lynch or Conrad ... he was 30 to 40 blocks east of them and east of the plane. But he states emphatically that he saw the tail and engines come off just a few seconds from impact ... long after that "cargo-door sized piece" John Power had seen separate when the plane was at a higher altitude.
A New York City Deputy Fire Chief, I'll call him Frank, was at The World Trade Center on September 11th when the Towers collapsed. He was one of the men who carried the crushed, dead body of Father Mychal Judge away from the fallen Tower. This Chief is one of the top fire officials in the City of New York. He was in his Rockaway neighborhood on the morning of November 12, 2001 when at around 9:15 AM he looked up and saw Flight 587. He saw a fireball back of the left wing, with the tail on the plane, but the fireball did not persist. Then he saw stuff blow out (parts of the plane), the plane flipped (or rolled right), and spiraled down into the ground. Anyone care to call him notoriously unreliable?

Related Links:
Flight 800 Witness review by FIRO
Flight 800 Witness Hearing by FIRO

Please jump to the U.S.Read Home Page where you will see more links to Flight 587 stories.

Submit your email to the Editor. Please provide your full name, city, and state. Thanks for your comments!