On the evening of July 17, 1996, TWA 800 fell into the ocean seven miles from the Long Island town of East Moriches. The plane had taken off from New York’s JFK Airport and had been bound for Paris, France. All 230 people on board died. The inquiry into the crash, the most expensive and (in its attention to the plane’s internal systems) the most rigorous inquiry in aviation history, has lasted for four years. A final hearing by the National Transportation Safety Board has been taking place on August 22 and 23, 2000, as this issue of The New York Review goes to press; neither the hearing nor the report that will follow it is expected (according to advance press reports) to “pinpoint” the cause of the crash. The full written report will become available to the public in several months.1
During the first year and a half of the inquiry, investigators from the National Transportation Safety Board and from the FBI concentrated on three matters: the possibility of a mechanical cause, the possibility of a bomb, and the possibility of a missile. In November of 1997, the FBI formally announced its conclusion that neither a bomb nor a missile had caused the accident (at this point, the FBI withdrew from the case). In December of 1997, the Safety Board—which had painstrakingly reconstructed most of the plane and scrutinized all of its internal systems—held a week-long public meeting reviewing the extensive, but inconclusive, evidence it had accumulated.
Investigators had long known that the plane’s central fuel tank exploded, but had at first been uncertain whether that explosion occurred early or late in the sequence of events that brought the plane down. By the time that the December 1997 public meeting took place, the Safety Board had long since concluded that the central fuel tank explosion was an early event in the catastrophe. What still remained was to find the source of the ignition, the cause of the explosion.
Between December of 1997 and August of 2000, the Safety Board continued its search for the ignition source. Included in its inquiry, and emphasized in the August hearings, was the possibility of a short-circuit or some other problem in the plane’s 150 miles of aging wiring. Also included in the inquiry was the possibility of an external ignition source, electromagnetic interference from one, or more than one, of the many military and civilian ships and planes that had been in the vicinity of TWA 800 and that, along with powerful civilian and military transmitters on land, might have produced an adverse electromagnetic environment. This second line of inquiry—electromagnetic interference from a source external to the plane’s own wiring (internal and external sources of ignition are not mutually exclusive:damaged wiring inside a plane can increase its vulnerability to external transmissions)—is the subject of the article that follows.
In the past two years, the National Transportation Safety Board has taken steps to assess the severity of the electromagnetic environment that surrounded TWA 800 on the evening it fell. In July of 1998, the Safety Board promised to enlist the technical expertise of the Joint Spectrum Center, the agency inside the Department of Defense whose laboratories are designed to sort out problems in the electromagnetic spectrum, particularly as those problems affect military craft. The NTSB also promised to call upon the services of the National Aeronautics and Space Administration (NASA), some of whose scientists work directly on problems of electromagnetic interference affecting civilian planes, as well as space shuttles and other craft.2
By August of 1998, the Joint Spectrum Center had begun its work and in January of 1999 it delivered to the NTSB its completed study of the ground, sea, and air transmitters that, according to the NTSB, were known to have been in the TWA 800 environment.3 The $45,000 study designates the signal frequency of each fixed and mobile transmitter and computes the electric field strength each signal would have had at the accident site itself. The Joint Spectrum Center’s report was then forwarded by the Safety Board to NASA so that its scientists could determine whether the electrical field strength at the accident site (as calculated by the Joint Spectrum Center) was strong enough to have provided an ignition source for the explosion suffered by TWA 800. NASA’s assessment was expected to be completed by December 1999, but the complex study required fifteen months rather than the hoped-for twelve.4 In March of 2000 NASA forwarded to the NTSB a report concluding that the signal stengths specified in the Joint Spectrum Center study were not strong enough to have caused the central fuel tank explosion.5
Two aspects of the work so far accomplished need to be looked at: first, the impressive technical strengths of the Joint Spectrum Center and NASAstudies; second, the incompleteness of the studies, the result of the omission from the analysis of certain ships and planes in the area. The Joint Spectrum Center, though headed by officers from the Navy, Air Force, and Army, does not itself have the legal authority to conduct an inquiry into the military craft in the area of the accident. The National Transportation Safety Board does have that legal authority, and the Joint Spectrum Center must rely on the Safety Board to carry out a full reconstruction of the possible external sources of electromagnetic interference and to provide it with a complete list of those sources. The Joint Spectrum Center openly states that its report is only as accurate and as comprehensive as the Safety Board’s list of military craft in the area is accurate and comprehensive. Since NASA’s research, in turn, is based on the figures arrived at by the Joint Spectrum Center, NASA’s conclusions also depend for their validity on the accuracy of the list the NTSB originally provided to the Joint Spectrum Center.
While major steps have been taken to ascertain the role that electromagnetic interference might have played in the fall of TWA 800 (steps whose strengths and weaknesses will be looked at in detail in a later part of this article), it is important to recall that during the time the Joint Spectrum Center and NASA studies were being made, two other drastic events took place in the same geographical region: the fall of Swissair 111 on September 2, 1998, and the fall of EgyptAir 990 on October 31, 1999. When each of the three accidents is looked at in isolation, each has features that suggest that electromagnetic interference is a possible cause—a possible cause that, like all other possible causes, should be carefully scrutinized.
The three accidents viewed together, however, greatly increase the need to assess with precision the possible role of electromagnetic interference: the occurrence of three accidents in a single area increases the chance that something in the external environment is acting as a contributing factor. Studies by the Joint Spectrum Center and by NASA need to be undertaken not only in the case of TWA 800 (where ships and other equipment that could have caused electromagnetnic interference were unfortunately omitted from the study and should now be added) but also in the cases of Swissair 111 and EgyptAir 990. Carrying out such studies will require that the National Transportation Safety Board first undertake a full reconstruction of the military and civilian craft in the area surrounding each plane. Without a comprehensive picture of all transmitters in the vicinity, no precise analysis of the radio and radar environments can be made.
This essay is a request to the NTSB that such inquiries be started as soon as possible.
Like TWA 800, Swissair 111 is believed to have suffered an electrical catastrophe.6 In the case of Swissair 111 as in the case of TWA 800, the originating event remains mysterious: TWA 800’s catastrophe involved the wiring in the central fuel tank and Swissair 111’s catastrophe involved the wiring in the entertainment system located near the cockpit, but in neither case do investigators believe they have yet found the originating cause.7
A third feature shared by the two accidents is location, and a fourth is timing. Each year in the United States planes take off more than 8.25 million times; in two years and two months (the time period that separates the TWA and Swissair accidents) the figure is close to 18 million.8 Two of those 18 million departures led to a mysterious electrical catastrophe. The two flights that suffered the catastrophe could have originated from two different airports anywhere in the country,9 or, for that matter, anywhere in the world. But as it happens, both planes took off from a single airport, New York’s JFK. The two flights could have taken off on any two days of the week and at any two minutes of the day.10 But as it happens, both took off on a Wednesday at 8:19 PM.11
The literature on electromagnetic interference is full of stories about unwanted electrical upsets that recur in the same space at the same time: one company, for example, found that its computers crashed every Friday at 3:00 PM; the cause turned out to be a piece of mowing equipment that was turned on at 3:00 PM each Friday during the summer. 12 Electromagnetic interference often does, of course, happen only a single time; but a one-time-only event often remains an unsolved mystery. Frequently unsolved, too, are problems caused by electromagnetic interference that recur irregularly (one day at 3:00 PM, seventeen days later at 9:00 AM, four hundred days later at 7:15 AM) in widely separate locations. When, in contrast, the event recurs at an exact time and place, it may become clear that—as in the case of the computer and the mowing machine—the problem is arising not from some problem inside the affected piece of equipment but from something outside—something on a regular enough schedule that its responsibility can eventually be tracked down.
Both planes started out on a route that took them east along the southern coast of Long Island. TWA 800 began to fall at 8:31 PM, twelve minutes into the flight. Swissair 111 flew for an hour longer than did TWA 800; it crashed off the coast of Nova Scotia at 9:31 PM Eastern Daylight Time. (See map.)
If both planes crossed paths with some fatal electromagnetic event in their environment that was operating13 on an almost identical schedule on the two evenings, it would seem that Swissair 111 should have begun to have trouble twelve minutes into the flight (the point when TWA 800’s data recorder, transponder, and data box ceased operating). In fact, something close to this seems to have taken place.
Reports about Swissair 111 have left the public with the incorrect impression that the plane managed to make its way, uneventfully, east along Long Island and north along the New England coast before suddenly beginning to encounter trouble when it was sixty miles from Halifax. The record of the difficulty, as widely reported in the press, begins at 9:14 PM, when Swissair 111’s pilot requests permission to make an unscheduled landing at Boston; the air controller in Moncton, Canada, reminds him that he is much closer to the Halifax airport than to Boston, and asks him if he would prefer to land at Halifax. At the moment of the 9:14 call, the cockpit has smoke in it and—as the data recorder would later reveal to Canadian investigators—ten minutes later the computers on board suddenly begin receiving false information. Over the next ninety seconds, the autopilot disconnects; the data grow more and more anomalous; soon the plane’s electrical systems fail altogether.14
Swissair 111 was an MD-11, a type of plane made by McDonnell-Douglas and derived from the DC-10. When the MD-11 first appeared in the 1990s, its “design philosophy” was widely celebrated: 1,500 software engineers (working in consultation with pilots from thirty-seven airlines) had created a plane that could fly smoothly while carrying out tremendous feats of self-repair. Even if the plane were to suffer “multiple electrical faults,” its computers would quickly “reconfigure” the electrical systems, instantly redistributing tasks among the plane’s three electrical systems. What was envisioned was not just that the plane would, under duress, buy time for its pilots and passengers; it would diagnose and eliminate even a severe problem, purring along as though nothing had happened.15 It may well be that Swissair 111’s problems began at 9:14 and that the plane’s systems, far from correcting themselves or buying time, simply surrendered to the ever-accelerating decline. But it is also possible that Swissair 111’s electrical problems began much earlier and that for a time the plane did carry out acts of self-repair.
What can be said with certainty is this: It is not the case that Swissair 111 flew uneventfully up the New England coast. In fact, it suffered a serious problem, and the problem first surfaced in the very zone where TWA 800 fell.
In the early stage of the flight, while Swissair 111 was still traveling east along the southern coast of Long Island, it lost radio contact with the eastern seaboard air controllers for thirteen minutes. TWA 800 had begun its fatal fall (and had lost the use of its radio, transponder, cockpit recorder, and data box recorder) at a clock time of 8:31 PM, less than a minute after a normal radio exchange with the Boston air controllers. Swissair 111 had its last normal exchange with the Boston air controllers at 8:33 PM, after which it lost radio contact with every air controller on the northeast coast for the next thirteen minutes.
Under normal conditions, exchanges between air controllers and pilots occur in pairs: a call initiated by the air controller will be answered by the pilot, who restates what the air controller has just said; or the call may instead be initiated by the pilot, who asks a question (such as permission to climb to an altitude that has less wind turbulence), which the air controller answers, after which the pilot repeats the information to verify that the words have been heard and understood. This pattern of call and recall is not a casual practice; it is a required procedure. While the sentences of pilots and air controllers normally occur in tight pairs, there can be many factors that for a few seconds interrupt the rhythm of the call-and-recall pattern, and necessitate a repetition of the call. But the failure to answer is never taken lightly and if it continues, it may become a matter of grave concern.
For a thirteen-minute period from 8:33 PM until 8:47 PM, no completed act of radio contact took place between Swissair 111 and the Boston area air controllers, whose radars are positioned at Sardi on Long Island, Hampton on Long Island, Cape Cod, Nantucket, and Augusta, Maine.16 As the plane progresses, it is passed along from one controller to the next. In Swissair 111’s last successful exchange at 8:33, the Hampton controller had told the pilot the radio frequency he should now use as he begins to enter the Cape Cod airspace and the pilot had accurately repeated back to him that new frequency:
Hampton Controller: Swissair 111. Boston17 one two eight point seven five.
Swissair 111: One two eight seven five. One Eleven, right.
Hampton Controller: Good bye.18
From this point forward, Swissair 111 should be in communication with the Cape sector. But the Cape controller cannot reach the plane; and so at 8:34 PM, he asks the Hampton controller to try to reach him on the old frequency: “Try him again, thanks.”
Hampton Controller: Swissair 111. Center.19
Swissair 111: [no response]20
The radio21 a commercial pilot uses for communication with air traffic control has a double screen: the frequency used for one sector (in this case, Hampton) is kept in place on the first screen when the new frequency (in this case, Cape) is dialed in on the second screen. That way the pilot can quickly get back to the first frequency, should he discover that he has misheard or misdialed the new frequency. But Swissair 111 can now be reached on neither frequency (though it remains visible on radar22 ). Unable to reach Swissair 111, the Hampton controller goes on to normal exchanges with other planes in the area—he instructs a plane addressed as Echo Charley to descend and maintain a specified altitude (and Echo Charley repeats back the altitude); he instructs a Delta flight to proceed to its destination (and the Delta flight repeats back the instruction).
The clock moves forward to 8:36 and the Cape controller renews his efforts to reach Swissair 111:
Cape Controller: Swissair 111. Climb and maintain flight level three one zero [31,000].
Swissair 111: [no response]23
Cape Controller: Swissair 111. Boston.
Swissair 111: [no response]24
The Cape controller now contacts the associated25 controller at Hampton to enlist his help once more:
Hampton Associated Controller: Hampton.
Cape Controller: Try Swissair 111 again, please.
Hampton Associated Controller: We tried him, he’s not here. We’ll try him again.
Cape Controller: O.K.26
At 8:38, the four-step cycle begins one last time. Step one: the Cape controller tries and fails to reach the plane:
Cape Controller: Swissair 111. Cleared direct [to] Bradd.27
Swissair 111: [no response]
Cape Controller: Swissair 111. Swissair 111. Hear Boston Center. Contact Boston one two eight point seven five, one two eight point seven five. If you hear Boston, ident.
Swissair 111: [no response]
Step two: the Cape associated controller contacts the Hampton associated controller to ask for help:
Hampton Associated Controller: Hampton.
Cape Associated Controller: Yes. This is Cape. Could you try Swissair 111 again off of Kennedy.28
Hampton Associated Controller (speaking to Hampton controller): Try Swissair 111 again, Gary.
Cape Associated Controller: Thanks, Bob.29
Step three: the Hampton air controller now twice tries to reach the plane, once by calling the name of the plane and announcing the radio frequency to be used for contact; then by calling the name and identifying who it is that is attempting to reach him:
Hampton Controller: Swissair 111. One twenty-eight seventy-five.
Swissair 111: [no response]
Hampton Controller: Swissair 111. Center.
Swissair 111: [no response]
Step four: having observed the failed exchange between the Hampton controller and the pilot, the Hampton associated controller now reports the unhappy result to the Cape associated controller:
Cape Associated Controller: Ya. Go ahead.
Hampton Associated Controller: Negative joy on that Swissair.30
Cape Associated Controller: O.K., then. Thanks.31
Although Swissair 111 is still in the air, it has lost radio contact.
Swissair 111—off the air for a total of thirteen minutes—eventually does get back in contact with the air controller. The pilot’s voice first comes through not at the air controller station at Hampton, Cape Cod, or Nantucket but at Augusta, Maine. The Augusta air controller at first believes he is receiving a call from a different Swissair plane (flight 104), one that is flying in the Augusta region airspace; but he quickly corrects himself and swiftly relays to the Swissair 111 pilot the frequency on which he should contact Boston:
Swissair 111: Boston Center, Swissair 111 heavy.32
Augusta Air Controller: Is that Swissair 104?
Swissair 111: Negative. This is Swissair 111…[Here Swissair 111 and Swissair 104 begin to speak simultaneously.]
Augusta Air Controller: Stand by, Swissair 104. Swissair 111, Boston Center.
Swissair 111: Boston Center, Swissair 111. Go ahead.
Augusta Air Controller: Swissair 111, contact Boston Center, one three three point four five. [frequency 133.45]
Swissair 111: Three three four five. Swissair 111.33
The time is 8:47.34 A moment later, at 8:48, Swissair 111 successfully contacts Boston’s Nantucket sector:
Swissair 111: Boston Center. Swissair 111 heavy.
Nantucket Air Controller: I’m sorry. Who was that last call?
Swissair 111: Boston Center, Swissair 111 heavy is calling 133.45.
Nantucket Air Controller: Swissair 111. Boston Center, roger. How do you read?
Swisssair 111: I read you loud and clear. Go ahead.
Nantucket Air Controller: Swissair 111. Climb to flight level two niner zero. Higher shortly.
Swissair 111: Level two niner zero. Swissair 111.35
Other than the spirited inquiry about legibility—“How do you read? I read you loud and clear”—the air controller and pilot do not stop to welcome one another back or to discuss the previous radio blackout. They at once turn to the business at hand, the resumption of the scheduled climb to 33,000 feet that had been interrupted at 27,000 feet when the radio transmissions were suspended. The confident tone and the reassuringly professional procedure of information given (two niner zero) and repeated (two niner zero) continue over a sequence of exchanges about altitude and radio frequency until 8:58, when the Nantucket controller passes the plane on to the Moncton controller in Canada. Radio contact has been restored; a normal flight has been regained; the events of the previous quarter-hour now seem—and may actually only be—an uneventful anomaly, a passing fluke.
But before too many more minutes pass, a lethal set of events—as we now know—will begin to take place; and the possibility exists that the fatal sequence of events is linked to the earlier events, that the electrical and radio systems36 of Swissair 111 were already under strain37 and that although the relatively new plane was able for a time to withstand, or compensate for, whatever was affecting it, eventually it lost the capacity to do so.
Two questions are raised by the thirteen-minute blackout. First, is the blackout related to the final set of catastrophic events? Second, if the blackout is related to the final catastrophe, does that tell us anything about whether the problem originates from a source inside or outside the plane? It tells us that external, as well as internal, causes need to be scrutinized. The fact that Swissair 111 begins to have radio trouble at the time when, and place where, TWA 800 suffered its swift catastrophe increases our obligation to include external agents in the overall inquiry.
Almost as mysterious as the thirteen-minute silence of the pilots is the silence of the FAA and our country’s Safety Board after the accident. What can account for their not having reported to the public the radio troubles suffered by Swissair 111 as it progressed along the southern shore of Long Island (where its sister plane had fallen in an earlier summer) and up the sea lanes running beside New England? Why was it important to confine the accident to Canadian airspace and Canadian waters in the public imagination?38
Common sense presses us to consider the possibility that the thirteen-minute radio blackout may bear on what by 9:14 had become a swiftly accelerating electrical catastrophe. So, too, does certain supplementary information. First, it is highly unlikely to be the case that the pilot simply dialed the wrong frequency when his plane was handed off from the Hampton to the Cape sector: as noted earlier, the pilot reads back the correct Cape frequency when the Hampton controller gives it to him; further, pilots have multiple radio screens and leave one tuned on the old frequency.
Second, it cannot be the case that the pilot, distracted by a meal or a book or a conversation or an extraordinary cloud formation, simply “forgot” that he was in the midst of a climb or that he was passing though one of the busiest corridors in the world. The voice of this pilot, both before the blackout and once his radio returns, is professional, crisp, quick: he consistently responds to each air controller’s call a split second after he receives it, and often recites back the information in the precise order in which it has been given. One Swissair official has said that of 450 pilots who fly the MD-11, this particular pilot was considered one of the top five.39 Captain Urs Zimmermann’s professional reputation and voice signature alone should persuade us that he was incapable of “neglecting” air traffic communication. But if more evidence is needed, there is the MD-11’s sophisticated data recorder which registers whenever a pilot attempts to key into the communication system: it shows that during the thirteen-minute blackout, the pilots of Swissair 111 made repeated attempts to initiate radio contact.40
Third, it cannot be the case that either United States or Swiss flight procedures permit such a blackout to pass with a complacent shrug or a bemused scratch of the head. In 1990, the US House of Representatives held a hearing on Pilot/Air Traffic Controllers Communication Issues. The hearings enumerated twelve kinds of communication error and the potentially fatal consequences of even the most seemingly minor of them (such as two people attempting to speak at the same time for several seconds). The twelve kinds of problem had earlier been outlined by a 1988 aviation industry report entitled A Call to Action. Included are such problems as an air controller using an abbreviation that could stand for two different planes; a pilot’s accent making it hard for the air controller to determine whether he has accurately understood the instruction; and a “blocked” line occurring when one pilot attempts to use a frequency shared with other pilots, one of whom is at that moment already speaking. At several points the report specifies with alarm the duration of a given problem, and it is never close to Swissair 111’s thirteen-minute blackout: frequency blockage “for several minutes” is cited as though a self-evidently alarming situation; a “stuck mike”41 that lasted for “five minutes” is a second example; radio contact lost for twenty miles (two or three minutes in a commercial plane) is a third example. If several minutes or two minutes or five minutes of lost communication requires an industry “call to action” and hearings before the House of Representatives, how can a thirteen-minute gap be considered a problem too minor even to mention in the press in the United States? How can it be too minor to mention when it—unlike the shorter communication blackouts cited in the congressional hearings—was prelude to a fatal plane crash?
TWA 800 and Swissair 111, then, share at least five features: (1) a grave electrical accident, (2) a so far indecipherable cause, (3) a takeoff from the same airport and a route across the same geography, (4) a takeoff on the same minute of the day and day of the week, and (5) the malfunctioning of its radios beginning at almost the same time (somewhere in the three-minute interval between 8:31 and 8:34).42
Should these five features be seen as extraneous, a set of interesting but ultimately insignificant coincidences? Or are they instead features that together expose the cause of the accidents? Either answer could be correct. The only way to learn which is accurate is to investigate the second possibility with the greatest possible rigor and speed. Has the United States investigative team (the American researchers who are assisting the primarily Canadian investigation) under-taken to reconstruct the external environment through which Swissair 111 flew? There is, to date, no public sign of any such reconstruction. If United States investigators wait until every possible internal cause has been explored before they begin to look at the external possibilities, will it be possible to construct an accurate and complete record of that external environment? The memories of air control-lers, pilots in the area, and seamen are clearer today than they will be in two years: their assistance in reconstructing the external environment should therefore be sought today, not two years from today. External explanations need to be pursued for exactly the same reasons that internal explanations are already being urgently pursued: because there is an absolute need to know the cause of these two isolated catastrophes and because there is an absolute need to prevent other planes from crashing.
From what we know about the external environment, a sixth feature shared by TWA 800 and Swissair 111 begins to come into view. The two planes attempted to make their flights on an evening when military craft were in the air or sea below. The route from JFK International Airport east along the southern coast of Long Island and north past the New England shoreline requires any plane on its way to northern Europe to thread its way through a ribbon of air that is skirted on one side or the other by military warning zones. The boundaries of each zone are marked on aviation maps and labeled with the letter “W” followed by a number.43 Where the map has room, a printed sentence appears inside the zone: “Warning: National Defense Operations Area, Operations hazardous to the flight of aircraft conducted within this area.”44
Such military warning zones are, of course, often unused by the military, and during such unused periods can be entered by civilian flights. But the record of scheduled military exercises shows plans for air and sea activities in the week during which Swissair 111 attempted its flight, just as the equivalent record from two years earlier shows planned exercises during the week of TWA 800’s flight. 45 This may be why Swissair 111, like TWA 800 earlier, had been directed onto the Bette route in traveling east out of New York, for this route is assigned when the military exercise zones south and southeast of Long Island, called W-105 and W-106, are in use by the military.46
The list of ships and planes in the external environment (along with the already-known location of all ground transmitters) needs to be reconstructed swiftly and accurately by those who have both the authority and obligation to carry out this task, the United States’ National Transportation Safety Board, and not by isolated citizens working through Freedom of Information inquiries. The Freedom of Information procedure is an inspired United States invention, but it allows only a piecemeal picture to come into place, and that only over many slow months. At present, however, Freedom of Information inquiries appear to be the only path of reconstruction available to us; so let us look at what they let us know about the September 2-September 3, 1998, period.
In or near the warning areas skirting the Swissair 111 flight were three submarines: USS Connecticut (SSN-22), USS Dallas (SSN-700), and USS Billfish (SSN-676). The commander of Submarine Group Two from the Naval Submarine Base in New London, whose office provided the names of the submarines, states that these craft were each acting in isolation and not in exercises with other craft and therefore “will not assist you in your stated goal of discovering ‘which exercises took place.'”47 But at issue here is not armed exercises; at issue instead are electromagnetic transmissions (or any related phenomena such as radar decoys, or chaff): all communications between the submarines and both fixed and mobile transmitters need to be scrutinized, including, for example, the two-million-watt submarine transmitter at Corbett, Maine. Whether any submarine transmitters were close to the plane, and, if so, whether their signals were strong enough to adversely affect the plane are among the many questions that need to be answered.
Potentially important to the inquiry is the record of flights by the Navy P3 planes, which are stationed at Brunswick, Maine. On the earlier night when TWA 800 fell, a P3 had crossed the plane’s path fifteen seconds before TWA 800 lost its transponder, voice recorder, and data recorder. (The P3 itself, though it had a safe flight, reported on its return that it lost the use of various pieces of electrical equipment during the flight.) The record for the evening on which Swissair 111 flew has some similarity. According to written documents provided by the Commander Naval Air Forces, Atlantic Fleet, three P3s were in flight during the hour and fifteen minutes when Swissair 111 made its way east along Long Island and north along the New England coast.48 Two of the three were from Patrol Squadron 26, a squadron called the Tridents (named after the three-pronged spear used by the sea god Neptune). The Navy P3 that flew within one mile of TWA 800 was also from Patrol Squadron 26.49
The third P3 in the air on the night of Swissair 111’s passage was a P3 from Patrol Squadron 10, the Red Lancers, whose insignia is a pair of lightning bolts.50 This type of P3 carries more high-powered transmitters than the Squadron 26 planes, and may be closer to an EP-3 or electronic warfare plane.51 (Its surveillance capacities were widely described during the Kosovo conflict: each time we heard that our planes could identify a plane sitting on the ground from an altitude of 27,000 feet, identify a bicycle leaning against a wall from 20,000 feet, and differentiate species of grass from 10,000 feet, it was Red Lancer P-3s from Squadron 10 in Brunswick, Maine, that were being described.52 )
The relation between the P3s from Squadron 26 and the P3s from Squadron 10 is perhaps illuminated by an incident that occurred on the afternoon of the day Swissair 111 attempted its evening flight. A regular P3 from Squadron 26 flying in military exercise zone W-104 (off the coast of Massachusetts) reports in its mission statement that it had to leave the exercise area one half hour earlier than planned because of the presence of a P3 from Squadron 10 (the warning zone covers hundreds of square miles so it is unclear why it could not accommodate two P3s). The pilot’s mission statement does not explain why the presence of the Red Lancer plane necessitated his own departure, but he does report that his plane “lost” its UHF and VHF radios, as well as an instrument that informs the crew when they are positioned over a particular sonobuoy. How it lost its two radios and its On Top Position Indicator (OTPI), and whether that loss is connected to its having operated near the P3 from Squadron 10, are matters that need to be clarified. It is, of course, the three P3s that were flying at the same time as Swissair 111, rather than these flights that took place a few hours earlier, that are of primary importance.
The possibility of electronic warfare practice is suggested by several items in the Navy record. A document called the “Fleet Area Control and Surveillance Facility,”53 which summarizes the military exercises on the eastern seaboard in the first week of September 1998, explicitly announces Electric Counter Measures and Electric Counter-Counter Measures operations during the week and includes among “Weekly Notes” a memo on the need to “submit a small scale E[lectric]C[ounter]M[easure] Notification” in accordance with Navy rules. One such electronic warfare exercise, for example, is specified for the night of September 2, the night Swissair 111 flew, in military exercise zone W-72 off the coast of Virginia from 10:00 PM to 12:00 AM.54
All three P3s were on several-hour-long missions, and therefore could have been very far from, or, instead, near to, the path of Swissair 111.55 Where they were is one of many questions that need to be answered with precision and care.
I hope the National Transportation Safety Board is already at work to reconstruct the external environment of planes, ships, and ground transmitters56 along the route of Swissair 111; or if it is not already, will undertake such a reconstruction soon; and that it will enlist the assistance of the Joint Spectrum Center and NASA in determining the power levels both at the moment the flight first lost radio contact and at later moments along the route. Relevant, too, will be the record of other flights that have lost radio contact in this same geographical region.
In addition to reconstructing the electromagnetic environment of each plane that suffers a fatal fall, a related form of scrutinizing the environment has been suggested by D.V. Giri, who specializes in applied electromagnetics, including EMP (electromagnetic pulse), HPM (high-powered microwave), UWB (ultrawide band systems), and lightning. 57 Mr. Giri suggests that the NTSB could enlist NASA’s modified F-106B to assess the environment.58 Elaborately equipped to measure lightning strikes and coronas, the airborne laboratory could make a series of test flights from JFK along the Bette route, experimenting with various takeoff times and speeds. The test plane (if suitably modified for sensitivity to radio transmissions rather than lightning) could conceivably discover a single transmitter, or a complex ar-ray of transmitters, that may be producing the adverse electromagnetic environment.
This essay has shown eight features shared by TWA 800 and Swissair 111: (1) they took off from the same airport; (2) they took off on a Wednesday at 8:19; (3) they traveled along the Bette route; (4) they both had their first signs of trouble in the same region of airspace between twelve and fourteen minutes into the flight; (5) they both appear to have suffered an electrical catastrophe; (6) they both suffered a catastrophe whose cause remains mysterious, even after years of rigorous inquiry; (7) they both flew during a week when extensive military exercises were being conducted; (8) they both flew when certain specific transmitters (submarines, the Navy P3s) appear to have been in the region.
These overlaps may implicate the external environment. If two planes suffered a mysterious electrical catastrophe but had taken off from different cities, traveled along different routes at different times, and had their first trouble one after twelve minutes and the other after one hundred minutes, it would still be crucial to reconstruct the electromagnetic environment and include it among the pos-sible causes to be investigated. (Electromagnetic interference, as stated at the outset, can just as easily happen at irregular as at regular times and places, though the irregularity of the external environment makes it harder to discern that the external environment is playing a part). But to have two electrical accidents and also to have them share many key features of time and place should surely accelerate the inquiry into the external environment.
Rusty Yeiser, a retired naval aviator and former commander of the Joint Spectrum Center who oversaw the Center’s analysis of TWA 800’s electromagnetic environment, believes that other accidents would benefit from similar analyses. He recently said:
Aircraft electronic systems (often referred to today as “avionics”) continue to grow more and more complex, and are increasingly relied upon for safety of flight functions (including flight control, navigation and communicating, and data storage and retrieval). Given this, it would seem reasonable that a comprehensive examination of the external electromagnetic environment should become a routine component of commercial aircraft accident investigations conducted by the NTSBor other similar national organizations abroad. The Joint Spectrum Center’s core capability for analysis of electromagnetic interference lends itself directly to this type of task.59
Physical events, as has often been observed, tend to outpace our ability to describe them. While the present essay on TWA 800 and Swissair 111 was being written, a third large passenger plane entered the Atlantic Ocean south of Long Island, a third large passenger plane that—as in the other two cases—fell without any discernible mechanical cause. There are factors which differentiate EgyptAir 990 from the two earlier flights (it did not take off on a Wednesday at 8:19) but also key factors that link them.
The second part of this article will look at EgyptAir 990, as well as at the important, but incomplete, work that has so far been carried out on TWA 800.
—This is the first of two articles.
September 21, 2000
Matthew L. Wald, “Flight 800 Report Not Expected to Pinpoint Crash’s Exact Cause,” The New York Times, August 15, 2000, p. A23; Pete Donohue, “Flt. 800 Report Won’t Cite Cause,” New York Daily News, August 18, 2000; Laurence Zuckerman, “US Reviews Draft Report on T.W.A. Crash,” The New York Times, August 23, 2000, p. A19. ↩
NTSB Chairman Jim Hall outlined these steps in a July 8, 1998, letter, printed in The New York Review, August 13, 1998. The article I wrote on the subject in The New York Review and the correspondence that followed were published in the issues of April 9, July 16, and August 13, 1998. These were forwarded by the Safety Board to the Joint Spectrum Center when the Safety Board engaged the Center’s assistance (conversations with Commander John Mahoney of Joint Spectrum Center, September 1998, November 1998, January 1999). NASA’s formal report opens by briefly describing the New York Review articles and correspondence (J. Ely, T. Nguyen, K. Dudley, S. Scearce, F. Beck, M. Desphande, C. Cockrell, “Investigation of Electromagnetic Field Threat to Fuel Tank Wiring of a Transport Aircraft,” NASA/TP-2000-209867, March 2000, p. 1). ↩
Joint Spectrum Center, Department of Defense, “TWA Flight 800 Electromagnetic Environment,” (Project Engineer, Richard DeSalvo; Consulting Engineers, Martin Macrae, Douglas Hughes), January 1999. In the NTSB’s accident inquiry documents, the report is Docket No. SA-516, Exhibit No. 9A. Addendum 2, “Concerning Electromagnetic Environment.” ↩
During this time, NASA studied not only the external transmitters that could have affected the plane, but passenger-carried devices such as cell phones and computers. As their study points out, such devices are often operating several inches away from wires running through the passenger cabin wall. ↩
NASA scientists studied one possible path of interference-induced ignition: they looked at the way external signals could affect a particular wire (the fuel quantity indicator wire) that runs into the central fuel tank. As will be elaborated at a later point, NASA concluded that the external emitters listed by the Joint Spectrum Center could have at most introduced 0.1 millijoule of energy into this wire, and that a minimum of 0.2 millijoules is required to produce ignition. (J. Ely, et al., “Investigation of Electromagnetic Field Threat to Fuel Tank Wiring of a Transport Aircraft,” p. 39.) ↩
According to Vic Gerden, lead investigator of the Swissair 111 accident for the Transportation Safety Board of Canada, investigators have to date found twenty wires running into the cockpitsome from the entertainment system and some from flight-related systemsthat suffered “arcing” (telephone conversation, June 29, 2000). Arcing is a hot and persistent form of electrical sparking across a gap between two conductors; it can damage both the wire in which it occurs as well as the electrical devices to which it is connected. The Canadian Safety Board is trying to determine whether the arcing took place at an early or a late moment in the sequence of events. It is known that some of Swissair 111’s electrical systems ceased functioning six minutes before the plane crashed into the ocean, but whether all five of the plane’s systems (three main electrical systems and two auxiliary systems) failed has not yet been determined (Gerden, June 29, 2000). Both Vic Gerden and Boeing’s safety spokesman have consistently stated that the accident, though so far indecipherable in origin, appears to be electrical in nature. (For accounts of the electrical nature of the Swissair 111 accident, see also Paul Koring’s 1998 articles in Toronto’s The Globe and Mail, September 8, 10, 11, 16; October 5; and November 21, 1998.) ↩
The two locationsTWA 800’s fuel tank and Swissair 111’s entertainment systemhave both been given much attention in the press. But they have been given different importance by the United States and the Canadian safety boards. The United States National Transportation Safety Board believes, with a high degree of certainty, that the central fuel tank is the site of TWA 800’s trouble; the officials of the Transportation Safety Board of Canada have repeatedly stated that damaged wiring connected to the entertainment system is not, in their view, a special suspect in determining the cause of the crash of Swissair 111. The system’s wiring is not ruled out as a cause, but neither is the wiring from any other source. ↩
This number is based on figures provided in the Department of Commerce’s Statistical Abstract of the United States: National Data Book (1998, 118th edition), chart 1070, “US Scheduled Airline Industry-Summary: 1985 to 1996,” p. 655. The last year for which the number of departures is specified is 1996 when the figure is 8,227,900 departures. ↩
According to the Statistical Abstract of the United States there are 18,292 airports in operation in the United States; 5,129 of them are public (Chart No. 1080, “Civil Flying: Summary 1970 to 1996,” p. 659). ↩
Counting only the minutes between 6:00 AM and 9:00 PM, when most plane departures occur, there were 6,300 different minutes at which the two planes might have taken off in a given week. ↩
The 8:19 PM departure time for Swissair 111 was announced by Terry Benczik, spokesperson for the Port Authority of New York and New Jersey, and was widely cited (The New York Times, September 3, 1998, p. A1; National Public Radio Morning Edition, September 3, 1998). TWA 800 is usually also described as taking off at 8:19. But media reports of both Swissair 111’s departure time and of TWA 800’s departure time vary back and forth between 8:18 and 8:19. They do so in part because the official documents themselves vary: for example, the fall of TWA 800 into the sea at 8:31 is alternately described in NTSB documents as taking place twelve minutes and thirteen minutes “into the flight,” which would place takeoff at 8:19 in the first instance and 8:18 in the second. (See, for example, “Flight Data Recorder [FDR] Group Chairman’s Factual Report: Revision 1” [February 15, 2000], Docket No. 5A-516, Exhibit 10A, p. 2.) ↩
Conversation with Ron Brewer, expert on EMC (electromagnetic compatibility) systems design, October 1998. ↩
It is also, of course, possible that they encountered fatal signals from the same mobile or fixed transmitter but at different times, TWA 800 somewhere between 8:19 and 8:31, when it lost all electrical power; and Swissair 111 somewhere between 8:19 and 9:14, when it announced smoke in the cockpit. ↩
Each of these events is compatible with electromagnetic interference, as well as with other possible causes. (Once the fire was underway, it could itself have been the cause of the generation of anomalous data; at issue is the cause of the fire.) Because the events have features that appear to suggest a powerful electrical event, the Canadian Safety Board (which is investigating many possible causes) has looked closely at the question of a lightning strike, as did investigators in the TWA 800 case, who discovered there had been no lightning strike within a 300-mile radius of the plane (Docket No. SA-516, Exhibit No. 5-A, p. 5). ↩
For accounts of the MD-11’s capacity for self-diagnosis and self-repair, see Aviation Week & Space Technology, October 22, 1990, p. 45; The Washington Post, June 18, 1989, p. A3; and Bulletin of the American Association for the Advancement of Science, June 30, 1989, p. 1532. ↩
The following reconstruction draws together eight separate but overlapping FAA tapes from five different radar sectors in the first forty-five minutes of Swissair 111’s September 2, 1998, flight. The tapes were obtained (May 19, 1999) through a Freedom of Information request originally filed November 1998 and re-requested (after denial) on April 21, 1999. ↩
“Boston” followed by a series of numbers means, “Contact the next Boston controller on the following frequency.” ↩
Minute 8:33 on FAA Tape for Hampton Radar Sector for September 2, 1998, 8:23 to 8:45 Eastern Daylight Time. “Swissair 111 Aircraft incident, 9/3/98 at 0131 UTC ZBW, Sec 31R, A/G, Ch 03, 0023-0045 UTC [12:23 to 12:45 Universal Time Coordinated], ‘Cert Rerec’ [contains the name of an FAA official certifying that the tape is an accurate re-recording], Tp 503, FOIA 1999-001839NE [identification number of Freedom of Information request that prompted the sending of the tape]. The eight FAA tapes all use this form of titling, but will be given below in abbreviated form. ↩
“Center” is a terse form of self-identification: “This is Boston Center calling you.” ↩
Minute 8:34 on FAA Tape for Hampton Radar Sector. ↩
Commercial planes carry multiple radios, only one of which is used for communication with controllers. ↩
There is nothing on the air controller tapes to suggest that at any point Swissair 111 disappears from radar. Had that happened, the disappearance would have been registered in the controllers’ verbal statements to one another and recorded on the voice tapes. At one point, the visibility of the plane on radar is explicitly acknowledged by a Nantucket associated controller (see note 35 below). ↩
Which of the eight FAA tapes is being quoted will usually be clear from the context (here, for example, the FAA tape for the Cape radar sector). But because some of the exchanges occur on two different tapes, citations will continue to be given in the notes. ↩
Minute 8:36 on FAA Tape for Cape Controller (covering the 8:31 to 8:44 period). ↩
At each radar sector there is both an air controller and an associated air controller. The associated air controller monitors the exchanges taking place between the controller and the area pilots; sometimes he or she contacts the controller to give advice based either on observation or on information coming in from other radar sectors. For each FAA tape from a radar sector such as Hampton, Cape, or Nantucket, there is a second taped voice record of the associated controller. These tapes overlap, but are not identical, with the controller’s tape since they contain not only pilot-controller exchanges but conversations between controllers and associated controllers, as well as voices coming in from other radar sectors. ↩
This exchange at 8:36 is recorded on two tapes, the tape for the Hampton associated air controller (covering the interval 8:31-8:45) and the tape for the Cape controller. ↩
Bradd is a map point along the ocean flight path. ↩
The phrase “Swissair 111 off of Kennedy” means “the Swissair plane that a short time ago took off from JFK International Airport.” ↩
This exchange occurs at 8:39 PM on two FAA tapes, that for the Hampton associated controller and that for the Cape associated controller (covering the 8:31 to 8:44 PM interval). ↩
The phrase “negative joy” often means “no success.” In this case, what is “negated” appears to be both the substantive outcome of the undertaking (Swissair 111 was not reached) and the state of pleasure that would have come with reporting a positive outcome. It seems to have a meaning close to the following: “Of the two possible outcomes of my inquiry, the result is not the one hoped for.” ↩
A moment later one hears the voice of the [Cape] associated controller saying to a colleague: “ filing that? Looks like they’re filing that.” The fact that this statement comes almost immediately after the report of the failure to reach Swissair 111 suggests that one of the supervisors may be filing a formal record of the failure; but the exchange is much too abbreviated to be clear, and could have a different meaning altogether. In a letter replying to the author’s inquiry, FAA regional administrator Robert Bartanowicz states that there is no formal filing in the Boston office (July 29, 1999). ↩
It is hard to be certain from the tape whether the pilot uses here the word “heavy,” a term that indicates a large passenger or cargo plane, or the word “heading.” ↩
FAA Tape for Augusta, Maine, Radar Sector (covering the interval 8:41 to 8:53 PM). It is conventional practice for pilots to drop the digit “one” when repeating back the frequency to the air controller, as the Swissair 111 pilot does here in the final line quoted. ↩
Swissair 111 actually first appears on the FAA tape of the Augusta Controller Center one minute earlier at 8:46; but the transmission is not completely clear and perhaps because the plane is not yet in the Augusta region, the air controllerwho at that moment is in conversation with another planedoes not hear the call. ↩
FAA Tape for Nantucket (covering the 8:42 to 9:04 PM Eastern Daylight Time interval). ↩
Since a plane has so many backup radios, the loss of all radios for even three or four minutes is, according to Joe Como, a spokesman for the Airlines Pilots Association, highly unusual, and almost never happens without other things also being wrong (conversation with Joe Como, June 1999, describing radio blackouts in general, not Swissair 111 in particular). ↩
The plane’s systems would be strained (to take one of many possible examples) if some, or all, of the arcing found by the Canadian Transportation Safety Board in twenty wires took place at the moment radio contact was first lost (8:33 PM). It is also possible that the arcing occurred not at the initial moment of loss of radio contact but soon afterward, as attempts were madeeither by the pilots or by the automated repair systems within the plane itselfto reactivate the radios (between 8:34 and 8:46). The arcing may also, of course, have taken place late in the flight, around the time the cockpit began to fill with smoke (9:14), or ten minutes later, when anomalies first appeared on the data recorder (9:24), or ninety seconds later when the radio, transponder, and many electrical systems all simultaneously failed (9:26), six minutes before the plane entered the sea (9:31 PM). ↩
“There’s nothing significant on those tapes. This [the fall of Swissair 111] is a Canadian problem,” said one FAA official to the author. ↩
Captain Christian Stuessi, Swissair’s chief of pilots, quoted in the San Diego Union Tribune, December 18, 1998. ↩
Canadian Transportation Safety Board head Vic Gerden, telephone conversation, June 29, 2000. ↩
The term means that the pilot’s mike accidentally becomes stuck in an “on” position that lets that pilot speak but prevents any other pilot or air controller from using the frequency. (This clearly does not describe what happened in the case of Swissair 111, since other pilots were able to continue communicating and the Swissair 111 pilot was not able to do so.) ↩
The two planes would have been in approximately the same location at the first moment of failed radio contactin the narrow corridor south of Long Island and north of military exercise area W-105 and W-106. Both planes were starting into a climb but were at different altitudes. ↩
When the military warning zone occurs inside US territoryeither on land or on ocean waters directly touching the coastlinethe letter “R” for “Restricted” is used, for example, to denote a circle of airspace around Camp David. ↩
The sentence is printed, for example, inside warning areas W-103, W-104, W-105, and W-506 on the “World Aeronautical Chart” for the New York-New England area (Section map CF-19), (Department of Commerce: Washington, D.C., 1997). ↩
“Fleet Area Control and Surveillance Facility, Virginia Capes (FACSFAC VACAPES) Message, 26 August 1998” (giving day-by-day outline of planned exercises along the Atlantic seaboard for week of August 31- September 5, 1998); and “Fleet Area Control and Surveillance Facility, Virginia Capes (FACSFAC VACAPES) Message, July 15-20, 1996” (again giving day-by-day outline of planned exercises). ↩
Confirmation that Swissair 111 traveled on the Bette route is present in four documents: the Flight Progress Slip, the transcript of conversation between the pilot and “clearance delivery” controller at the airport prior to takeoff, the transcript of conversation between the air controller in the Kennedy tower and the New York air controller in Sector 57, and the transcript of conversation between the pilot and the Kennedy tower departure controller as the plane lifted into the air (Documents and letter responding to author’s Freedom of Information request from Franklin D. Hatfield, Manager, Air Traffic Division, Eastern Region Air Traffic Division, FAA, JFK Airport, June 4, 1999). ↩
Letter to author from W.A. Peters, Captain, United States Navy, Chief of Staff for Commander Submarine Group Two, April 21, 1999. ↩
Letter and accompanying documents from Mark E. Newcomb, Commander, JAGC, US Navy, Force Judge Advocate, by direction of the Commander Naval Air Force, United States Atlantic Fleet, Norfolk, Virginia, February 8, 1999. ↩
Patrol Squadron 26, “Patron Twenty-Six Flight Schedule, Wednesday 02 September 1998,” Event 4 (taking off at 4:50 and landing at 9:55), and Event 5 (taking off at 6:30 and landing at 9:55). ↩
Patrol Squadron Ten, “Patron Ten Flight Schedule, Wednesday 02 September 1998,” Event 8 (taking off from Brunswick, Maine, at 7:20 PM and landing back at Brunswick at 11:30 PM). ↩
The electronic warfare EP-3s for the US East Coast and Europe have their home base in Rota, Spain, but sometimes operate out of Brunswick, Maine. According to a knowledgeable military expert, the surveillance equipment on Squadron 10 P3s may be “passive”and may not actively transmit signals that could be harmful to other planes; this is the kind of question that should be explored. ↩
On the Red Lancer P3s’ “multi-sensor surveillance payload” in Operation “Eagle Eye” in Kosovo see Jane’s Navy International, Sept. 1, 1999; and International Defense Review, November 1, 1999; and on the reception of the Patrol Squadron 10s as returning war heroes, see Portland Press Herald, June 26, 1999, p. 2B, and August 10, 1999, p. 1B. ↩
Virginia Capes (FACSFAC VACAPES) Messages 162000Z August 98, Item E. ↩
The memo specifies the use of “two Lear [jets] with pods.” On the day TWA 800 flew (July 17, 1996), an exercise was also scheduled in “W-72” involving “two Lears.” ↩
The package of documents sent by Commander Newcomb includes a schedule of flights that occurred during the hours Swissair 111 was in the air but not the mission statements from those flights (which often contain brief narratives of what occurred). He did, however, include a mission statement from a flight earlier in the day. ↩
It is primarily the air- and seaborne craft that the NTSB needs to identify by further research, since the Joint Spectrum Center already has on file all the fixed ground transmitters along the American and Canadian seaboards. But because the Air Force and Navy sometimes carry out highly unusual experiments at these antenna sites involving transmitters that are not permanently installed (and therefore are not part of the fixed record), the NTSB should make specific inquiries about events taking place at these fixed sites as well. ↩
At the October 1, 1999, sessions of the Technical Interchange Meeting held at the Nuclear ElectroMagneticPulse Laboratory of the Swiss Defense Procurement Agency in Spiez, Switzerland, Mr. Giri and his colleagues discussed the question of whether the Swissair 111 accident could have been caused by electromagnetic interference of unknown origin. Giri concentrated on the radio blackout and the overlaps in the timing of the Swissair 111 and TWA 800 flights. The consensus of the group was that this was a realistic possibility deserving further investigation. ↩
D.V. Giri, conversations with author, September 1999, March 24, 2000. Although the F-106B is primarily used to test the effects of lightning, it has also been tested for high-altitude EMP (nuclear-electromagnetic pulse) effects. A particular sensor on the nose boom of the F-106B was designed by Mr. Giri (who has also designed microwave antennas and has helped to build EMP simulators in many different countries). ↩
Conversation, August 20, 2000. ↩