Fear of the deep. It is, at once, a very modern form of horror and a very ancient one. In this case it begins more than six miles high in the sky and ends more than three miles beneath the surface of the ocean.
Ten years ago, early in the evening of March 8, 2014, Malaysian Airlines Flight 370 began, like thousands of others every day and night, as a routine part of how the world moves easily between nations and continents, heading from Kuala Lumpur to Beijing.
Around 38 minutes after takeoff the crew talked to air traffic control, confirming their route north toward China. That was the last human contact from the Boeing 777. Three minutes later the transponder, automatically reporting the flight’s position, stopped transmitting.
Nearly six hours later, the jet, with 239 souls aboard, had been flying all that time sealed off from the world around it. No word had been heard from it and during that time nobody knew where it was.
The flight was impersonating a normal one even though it was far from normal. It was cruising steadily at around 35,000 feet at an airspeed of 630 mph and the engines had never missed a beat—there had been no messages from the crew of mechanical or other distress.
The problem was that no jets were ever supposed to be flying over this remote stretch of the southern Indian Ocean, far beyond any radar tracking.
The first few hours of the flight had been in darkness, but after five hours and forty minutes of this “zombie flight” the rim of the curving horizon ahead of the cockpit was turning crimson. The southern hemisphere dawn was breaking.
Suddenly, with an initial hiccup, MH370 began to die. One engine, on the right wing, was being starved of fuel and spooled down with a low whine.
At first, the controls were able to compensate for the effects of this engine failure—the Boeing 777 was designed to be able to fly for hours on one engine, although it would have to descend to denser air at 29,000 feet to maintain its height.
Fifteen minutes later the engine on the left wing began to falter and die; it also had exhausted its fuel. There was no way now that the flight could continue.
According to a computer simulation of the last minutes of the flight carried out by Boeing, the jet’s nose dropped, then pitched up. It was no longer flying within the aerodynamic laws that governed stable flight.
It began a descent to the ocean in what aerodynamicists call a fugoid motion, accelerating swoops, all the time the velocity of the dive rising. At the beginning the 777 was falling at the rate of about 12,000 feet per minute. At the end it reached 25,000 feet per minute, a fearsome speed equal to 284 mph. This meant that when it hit the water, probably as one dipped wing struck first, the airplane was torn apart.
Thus began the greatest mystery of modern aviation, a tragedy that left few clues but a torrent of questions and, inevitably, some bizarre conspiracy theories. We have an impatience with the inexplicable. Where facts are few, fantasy flourishes.
Not long after the story broke, and well before the technical factors involved had been evaluated, suspicion turned to the pilots, Captain Zaharie Ahman Shah and co-pilot Fariq Abdul Hamid.
Blaming foul play by either pilots (or others) in an air crash conveniently serves the industry interests involved—airline, airplane manufacturer, regulating agencies—and is a predictable habit. After the first of two deadly crashes that eventually grounded the world fleet of Boeing 737MAX jets Boeing insinuated that the pilots could have prevented the crash and even later implied that Asian pilots were less proficient than western ones, even though the airplane itself was at fault (in fact, Boeing expected pilots to deal with a design flaw that the pilots were unaware of).
An experienced accident investigator told me in 2014, “There are many interests here, and they don’t necessarily align with 100 percent full and candid disclosure at an early date. In fact, the motivation for full and candid disclosure by all parties hardly ever occurs in serious accidents, for some very important financial, political, liability, and social reasons.”
It’s also a quick and lazy way out for journalists looking for a sensational lede that can’t be instantly refuted because the pilots can’t rebut it.
Indeed, the vilification of Captain Shah began within a week. Malaysian authorities staged a deliberately public raid on his home. Soon afterwards they said that on his personal computer they had found a flight simulator he had been using in a plan to hijack his own airplane.
The hard drives of Shah’s computer were turned over to the FBI for examination in their labs. Nothing incriminating was found. Many pilots use simulators to keep themselves sharp.
Nothing in Shah’s private life suggested any motive. At the age of 53 he had flown more than 18,000 hours on commercial jets, 8,659 of them on the Boeing 777. Hamid, 27 years old, was regarded as one of the airline’s most capable first officers.
Had Shah—or anyone else aboard—carefully planned to make the airplane vanish, they would have acted very differently. Immediately after the loss of contact, Flight MH370 made a left turn. In doing so it headed straight into airspace covered by overlapping radars in Malaysia, Vietnam, Thailand and Indonesia. Had it turned right, over the South China Sea, it would quickly have slipped beyond radar coverage.
The left turn made sense as an emergency measure. It coincided with what was called the terminal primary approach for the nearest airport, at Kotah Bharu, on the eastern coast of Malaysia. Pilots making a sudden emergency response would then immediately have begun a rapid descent from cruise height. That never happened, neither then nor for the rest of the flight.
The charge most frequently made against Shah was that he had somehow devised a way of making the airplane disappear by disabling its two essential links to the ground, the transponder that continually reported its position and a system called ACARS, Aircraft Communication Addressing and Reporting System, that drawing on a multi-layered computer network in the airplane, monitored its vital functions and relayed them every half hour to the airline’s ground control center.
As already stated, the last exchange between the pilots and air traffic controllers was 38 minutes after takeoff and the transponder stopped working three minutes later. Fifteen minutes later the next batch of data from ACARS did not arrive.
These two systems were in the electronics bay, on a floor beneath the flight deck and adjacent to the forward cargo hold. The one remaining working link between the airplane and the ground was the small SDU, the Satellite Data Unit, situated in the roof of the passenger cabin behind the wings, the origin of hourly “pings” that were to prove vital to tracking the final hours and route of the flight. That difference in the locations turned out to be revelatory.
Late in 2015, the Australian Transportation Safety Board, Australia’s equivalent of our NTSB, issued a 22-page report on the duration of the flight, with a timeline, that included what clues they had been able to gather from the SDU signals. Buried in the weeds of the report I spotted a startling detail: “Power loss occurred at some time between these times.”
Power loss? There had never been any hint before of a loss of power in the airplane’s systems. I filed a story for The Daily Beast saying that the pilots appeared to have suddenly faced what could have been a cascading loss of electrical power somewhere in the belly of the jet.
Three days later a spokesman for the ATSB, Dan O’Malley, sent me an urgent email: “I’ve read your article and I fear that the ATSB may not have been clear enough in our report. The ‘power loss’ mentioned was referring to the SDU only. The transponder ceased to function at 17.22 [local time] but it is not known if that was a result of the same factor that caused the SDU not to perform its task. It is not known if any other systems stopped working during this time… we are publishing a clarified version of the report shortly.”
What was not being said here was noisier than what was. I asked for more clarification. The reply was: “It is not known if any other systems operations were affected by the same loss of power that affected the SDU.” (My italics.)
For the first time it was clear that something serious had happened in the electronics bay, serious enough to knock out the power supply to ACARS and the transponder. (The SDU, not directly hit by whatever had happened, had been able to reboot itself in minutes from a backup power source.)
One of the more gymnastic charges against Captain Shah was that he had somehow managed to open a hatch in the floor of the first class cabin, directly behind the cockpit, and gone below to pull the circuit breakers to kill the ACARS.
A veteran 777 pilot told me, “Few pilots would even know how to get down to the lower deck while in flight. And even if they tried, few would be familiar with the locations of avionics components, or be able to find the relevant breakers to pull.”
The single fact of the power loss and its location, without any knowledge of its scale or cause, led to another widely discussed theory, that a fire in the adjacent cargo bay, originating in a large shipment of lithium-ion batteries, had burned through a bulkhead and taken out the avionics.
I spent months investigating this thesis with the help of experts on batteries and on the 777’s fire suppression system. There were two recent cases of fires originating with consignments of lithium-ion batteries in the cargo hold. Both involved the larger Boeing 747, but the freighter version, carrying no passengers. The fires rapidly overwhelmed the pilots and destroyed the airplanes.
Any fire serious enough to penetrate a bulkhead in the 777 would surely have had similarly catastrophic results. Even a smaller fire would have compromised the integrity of the airframe and curtailed the flight. (There was some speculation that the fire could have died from lack of oxygen, leaving toxic fumes that overcame the pilots and passengers.) But without more physical evidence there was nothing conclusive about this theory.
As it turned out, more physical evidence did appear. The 777 had not entirely vanished. In the summer of 2015, a cleaning crew on a beach on the north coast of the small Indian Ocean island of La Reunion spotted a large piece of debris lying at the water’s edge. It was a significant piece of the control surfaces of an airplane wing called a flaperon.
La Reunion is around 2,300 miles from where the search for the Boeing 777 was taking place. Soon, other pieces of an airplane were turning up on beaches on other islands and on the coast of east Africa. Experts at the French air investigation body, the Bureau d’Enquêtes et d’Analyses (BEA), confirmed that they were from Flight MH370. All the debris was of lighter parts of the airplane that had broken away on impact and were buoyant enough to float for a long period.
This was a hugely significant development. First, it destroyed the most fanciful of conspiracy theories, that the airplane had been flown to a remote airfield by hijackers and landed intact, even though no hijackers were ever heard from. More usefully, experts could “read” the debris and assess how the jet had impacted the ocean, with what force and at what angle.
Moreover, there was the possibility that by reverse-plotting the course taken by the debris after the crash, the location of the point of impact, and therefore where the main wreck was likely to be found, could be calculated.
So far, this has turned out to be an expensive and disappointing endeavor.
It rests on a relatively young technology called drift modeling, normally used to predict the path of oil slicks in the ocean to see if they pose threats to coastlines. It involves complex calculations of the interaction of ocean currents, water temperature, winds, weather and the physical characteristics of what is being tracked—in the case of various pieces of an airplane, difficult enough and in this case, covering the span of the Indian Ocean, an untested and serious challenge.
This complexity did not, however, deter or daunt Dr. David Griffin, a highly respected oceanographer at Australia’s Commonwealth, Scientific and Industrial Research Organization (SIRO), based in Hobart, Tasmania.
Griffin and his team, with help from other oceanographers and weather authorities, spent many months using drift modeling trying to determine the crash site. By 2017, they narrowed the target area to between latitudes 40 and 30.5 degrees south intersected by a north-to-south arc representing the flight path of the jet’s final minutes, the arc itself being the product of a separate set of complex calculations.
“We are even more confident,” Griffin said, “that the aircraft is within the new search area.”
The seabed in the southern Indian Ocean, previously unmapped, had been revealed during the search as the stuff of nightmares. There were more than 200 volcanoes, many spouting hot mud. Some, in the path of the search, were higher than Vesuvius. The floor was more than three miles deep, with water pressure 300 times greater than on the surface. Wherever the wreck was, it was a hellish place.
Griffin was passionate, against enormous odds, in his effort to prove the efficacy of drift modeling. But the long search was closed down by the Australian government without a result.
A second search in 2018 came up empty after five months. It was carried out by the Texas-based company Ocean Infinity with state-of-the-art equipment, covering as much of the seabed in five months as the previous Australian-directed search had in three years.
Ocean Infinity deployed eight AUVs, autonomous underwater vehicles, that could closely follow the wild contours of the seabed, directed from a mother ship above. The search was agreed to with the Malaysians on a “no find, no fee” basis. Had it succeeded it was estimated that Ocean Infinity’s reward would have been around $70 million.
If a new Ocean Infinity search happens it will involve another generational leap in technology. This time there will be no mother ship. Instead, on the surface will be a new small fleet of robotic ships, each with its own swarm of AUVs. (To begin with they will have small crews but once proved they will be operated remotely from a mission control base in Southampton, England).
Explaining the system two years ago, Ocean Infinity’s CEO Oliver Plunkett said that a single mother ship “wasn’t the future” and “we needed to change our business.”
Plunkett remains committed to finding the wreck. He told The Daily Beast that he was “actively engaged” in returning to the search. “At this stage we are unable to say definitively when a new search will take place and there is still much work to be done.”
This caution was less evident when Plunkett made another statement over the weekend: “We now feel in a position to be able to return to the search…and have submitted a proposal to the Malaysian government. The search is arguably the most challenging one out there. We’ve been working to continue analyzing the data in the hope of narrowing the search area.”
Delivery of the fleet of robotic ships, built in Vietnam, was delayed by supply chain problems during the pandemic. The first two are undergoing trials and the Southampton mission control center has only recently had its first proof-of-concept rehearsals, using new low-orbit satellite links.
The persisting question remains, how certain can they be that the right area is being searched?
Richard Cole, a British expert in the use of satellite tracking who had helped me on a daily basis to follow the progress of the Ocean Infinity search, thought that Griffin had been overconfident about the accuracy of drift modeling. “Something that sounds to the public like an optimistic number, a 90 percent confidence level in predicting a search area, is statistically not very good. Nobody would cross a road if there was a 10 percent chance of being hit.”
Other scientists had worked to test drift modeling to suggest a search area. A team from the Italy-based Euro-Mediterranean Center on Climate Change had, for example, suggested in a report that a promising area lay between the latitudes 28 and 35 degrees south, overlapping the area that Griffin proposed.
One of the authors of that report, Eric Jansen, told me at the time that Griffin’s work was more thorough than his own team’s. But he is a lot more cautious now. He told me, “I don’t think that it would be possible to do studies that significantly reduce the uncertainties of the results we obtained back then. In order to restart search operations we would need a very significant reduction in uncertainty, far beyond what is possible with drift modeling.”
One reason is that, the longer the time elapsed between an event and when the modeling begins to recreate the ocean conditions at a specific moment, the more uncertain it becomes: “Over time,” Jensen said, “the uncertainty becomes quite significant.” It had taken the flaperon 509 days from the moment of impact to its arrival on the beach at La Reunion; other debris had been months longer in the water.
I asked Griffin, with whom I had had frequent and useful email exchanges during the two searches, what he now thought about the chances of a new search succeeding, but he refused to comment.
What of the airplane itself? The 777 is a creation of Boeing at its best, one of the last airplanes living up to the legend of a company that combined daring vision with rigorous attention to safety (the other is the 787). That legend has been deeply eroded by the new management culture devoted to making money rather than airplanes, a rapacity and cupidity that led to the ill-starred and disreputable record of the 737-MAX.
When the 777 was being designed in the early 1990s more safety features were baked into it than any earlier Boeing jet. Additional pressures were on the company to produce a flawless jet because it challenged bedrock beliefs about the safety of jets flying across large expanses of ocean. Specifically, that it was too risky for a jet with only two engines.
As late as 1980 the head of the FAA, Lynn Helms, told Boeing, “It’ll be a cold day in hell before I let twins [twin-engine jets] fly long-haul over water.”
This skepticism was seized upon by Airbus, the European consortium that had emerged as Boeing’s first serious competitor. Airbus led a chorus that it was too risky to have a jet large enough to carry more than 400 passengers fly on routes as long as 9,000 miles, much of the time over water and hours away from the nearest airport in case of emergency.
But eventually the 777 met and passed all the tests that Boeing had set for it in order to push to the edge on transoceanic flights—in fact, the FAA was ready to certify that it could safely fly on routes where it would be more than five hours’ flying time from the closest runway. Much of this was due to the advances in engine reliability. Airlines loved it because the twin was far more fuel efficient than a four-engine wide body.
A former Boeing test pilot told me: “This is arguably the best and safest airplane ever built, it is simply an amazing jet.” Of flight MH370, he said, “I believe that it continued to fly after the event, even with the incapacitation of the flight crew.”
He believes that the crew and passengers were overcome by a sudden depressurization and loss of oxygen, creating the condition known as hypoxia.
There have been several episodes of hypoxia leading to fatal “zombie flights”, notably in 1999 a private jet with four passengers flying from Orlando to Dallas where, following a loss of pressurization the pilots and passengers succumbed to hypoxia—the accident report by the National Transportation Safety Board said, “A period of as little as eight seconds without supplemental oxygen support following a rapid depressurization may cause a drop in oxygen saturation than can significantly impair cognitive functioning…”
In 2014 the Australian Transportation Safety Board said, somewhat tortuously, announcing the first systematic search for the missing jet, “the final stages of the crew/hypoxia event type appeared to best fit the available evidence for the final period of MH370’s flight.”
In 2016, displaying official amnesia, they claimed, “We have never stated that hypoxia (or any other factor) was the cause of this circumstance.”
Perhaps the best course is to follow the principle of a 14th-century Franciscan friar, William of Ockham, for resolving disputed scientific issues in the absence of conclusive evidence, a principle known as Occam’s Razor: “If there are two or more explanations for an occurrence the simplest, or the one with the fewest assumptions, is likely to be the correct one.”