by Jeff Wise
Investigators had no idea why the SDU rebooted the first time, at 18:25, but they thought they understood why it rebooted at 0:19. At that point the plane was nearly eight hours into what had been scheduled as a six-and-a-half-hour flight. The fuel tanks, then, must have been very close to empty. Once the engines ran out of fuel the generators would have cut out. The entire aircraft would have lost electrical power, including the SDU. At that point an emergency generator called the Alternate Power Unit would have automatically kicked in and restored power to select critical systems. After a minute-long power-up process the SDU would have re-established contact with Inmarsat. Without engine power to provide thrust, however, the plane would have been steadily losing speed and altitude en route to an inevitable crash.
The final BFO values offered a clue as to how fast the plane was descending during its final moments. The first, at 0:19:29, signified a vertical speed of between 4,000 and 14,000 feet per minute—a velocity toward the ground of 45 to 160 mph. Eight seconds later it was plummeting even faster, between 15,000 and 25,000 feet per minute, or 170 to 284 miles per hour. That change is equivalent to an acceleration of two-thirds of what the plane would experience freefalling in a vacuum. Given that the aircraft achieved this without engine thrust, it must have been pointing nearly straight down. For this to be the case, either someone had to have pushed the nose down in a deliberate dive, or there was no one at the controls and the plane had sloughed off into a spiral dive on its own. Either way, when the ATSB conducted simulations in which the plane went into a dive like this, it always hit the ocean within 15 nautical miles.
In other words, the plane should already have been found within the defined search area.
What to do? The investigators had to make a decision. The southern winter was drawing near. Soon high seas and fierce winds would make deploying the towfish too difficult and dangerous. The search would have to be suspended until the southern spring.
On April 16, 2014, officials from Malaysia, Australia, and China met in Kuala Lumpur to hammer out a plan. In the end, they decided that simply giving up was not a politically viable option. They issued a communiqué announcing that they were doubling the search area to 46,000 square miles. The newly defined search area stretched a bit further along the 7th arc and was nearly twice as wide. The officials stated that if the plane weren’t found in this newly expanded area, the search would be called off for good.
But then the story took yet another hard left turn.
Chapter 15
July 2015
At 8.30am on July 29, 2015, on the northeastern shore of Réunion Island, a French territory in the western Indian Ocean, a cleanup crew was working its way along a stretch of pebbly beach when a worker named Johnny Begue spotted an unfamiliar-looking object at the edge of the surf. Roughly rectangular and about six feet long, it somewhat resembled a stubby airplane wing encrusted with marine life. Intrigued, the men lifted the object and carried it higher up the shore.
Soon gendarmes were on the scene, along with local news photographers. The officers put the piece into the back of a Land Rover. Within days it had been packed up, loaded onto an airplane, and flown to France.
News of the find caused a sensation. From photographs circulating on the web, the piece was quickly identified as a flaperon, a part of the wing’s trailing edge. The flaperon’s function combines those of a flap, which droop down to allow a plane to fly more slowly on descent to landing, with those of an aileron, which are raised or lowered to cause a plane to turn. Specifically, this flaperon was identified as coming from the right wing of a Boeing 777. Since the only 777 ever lost at sea was MH370, investigators now had physical evidence to back up what the math had been telling them: the plane had gone into the southern Indian Ocean.
France could have turned the flaperon it over to the Malaysians, but instead decided to keep it and launch their own investigation. It was delivered into the care of the Direction Générale de l’Armement (DGA), France’s weapons development and procurement agency. At a laboratory in Toulouse, investigators rooted around inside the flaperon with an endoscopic probe and found serial numbers that matched records kept by the manufacturer in Spain. Now there was no doubt that the first piece of MH370 had been found.
Up until this point, the spoof scenario had slowly been gaining adherents among journalists and independent researchers. But for many the discovery of debris instantly rendered it a dead letter. The first question everyone had for me now was, “So I guess you were wrong, huh?”
I wasn't so sure. The idea that someone might falsify physical evidence in order to throw investigators off the scent struck many as conspiracy-theory craziness. But I didn’t see it that way. To me it seemed that if MH370 had been hijacked via a hack, then planting debris would be basic tradecraft. Indeed, after the Soviets accidentally shot down Korean Airlines Flight 007 over the Sea of Japan in 1983, they reportedly planted false acoustic pingers on the seabed to thwart U.S. search efforts.
For me, the most important thing to investigate was whether anything about the debris ruled in, or ruled out, the possibility that it had been planted. The French were keeping their investigation under tight wraps, but some intriguing tidbits leaked out. According to a report in the French website LaDepeche.fr, scientists who examined the flaperon determined that it had not floated at the surface but had drifted entre deux eaux—a phrase which literally translates to “between two waters,” but here means fully immersed, at a depth of several meters. This was puzzling. Inanimate objects cannot float neutrally buoyant underwater.
From personal experience as a scuba diver, I knew that when you’re underwater, you’re always either more or less buoyant than the surrounding water. By adjusting your weights and the volume of air in your buoyancy control device, you can come close to neutral and hover almost motionless, but every time you breathe in you will start to rise a little bit, and every time you breathe out you will start to sink. If you see a scuba diver suspended motionless over a coral reef, what they’re really doing is subtly rising and falling.
An inert object cannot adjust its buoyancy and will either sink or rise. “It is very hard to build something that will float slightly below the surface,” David Griffin, an oceanographer with the Commonwealth Scientific and Industrial Research Organisation (CSIRO), wrote me in an email. “The probability that an aircraft part does this is miniscule. The only way it can do this is if some of the object breaks the surface. If it does not break the surface AT ALL it must sink.”
I had no idea how long it would take the French to release a report on their findings, but I realized that I might be able to check the “entre deux eaux” assertion on my own. Using Google and Bing image search, I tracked down photographs of the flaperon taken at every angle by journalists who were on hand immediately after the piece was discovered. I saw barnacles growing on every surface.
I reached out to marine biologists who study these animals and learned that the specimens, commonly known as goose barnacles, belonged to a species called Lepas anatifera. These barnacles live exclusively on debris floating in the open ocean. Their larvae spend the early part of their lives swimming freely, then find an object on which to settle. In general, Lepas barnacles like to spread out and prefer the shade. They generally avoid territory close to the waterline, where the rising and falling waves periodically expose them to the air. “The uppermost centimeters of water are normally a quite harsh environment,” Hans-Georg Herbig of the Institut für Geologie und Mineralogie in Cologne, Germany, told me in an email. Exposed to sun and rain, they experience drastic changes in temperature and salinity as well as intense UV radiation.
Given the right environment, though, Lepas barnacles are notoriously fast-growing. The floating debris on which they have evolved to live is most often organic, and eventually will break apart and sink, so time is of the essence. Whereas a species of goose barnacle that lives on rocks might take five years to reach sexual maturity, Lepas can do it in mere weeks. And they set
tle quickly on any material that winds up in the ocean.
“I’ve picked a paper bag out of the Pacific that had barnacle larvae on it,” said Cynthia Venn, a professor of oceanography and geology at Bloomsburg University in Pennsylvania. As a result, Lepas-colonized flotsam can become extremely crowded in short order.
“Goose barnacles grow spectacularly fast,” Charles Griffiths, an emeritus professor of marine biology at the University of Cape Town, told me via email. “I have seen very large barnacles (as long as my finger) growing on a cable known to have only been in the water for 6 weeks.”
It might be possible, I realized, to infer how long the flaperon had been in the water from the size of its barnacles. But how big were they, exactly?
At first I was stumped. The flaperon was broken and jagged, so there was nothing of known size that I could gauge the shells against. Then as I dug through images online I came upon a photograph of gendarmes loading the flaperon into the back of a Land Rover. I found a diagram with the exact dimensions of this particular Land Rover model and used it to determine the dimensions of the flaperon, and by extension the dimensions of the Lepas shells. I calculated the shells on the biggest ones to be approximately 2.3 cm long.
Next I found a paper by a Japanese researcher named Yoichi Yusa who had studied the growth rate of a related Lepas species, Lepas anserifera. Using his data I calculated that it would take about 109 days, or four months, for Lepas to grow to this size. I also emailed Yusa photographs of the flaperon and asked him to estimate how long they’d been growing, and he answered: “I would guess that they had been there for a short time (between 2 weeks and a few months).”
Cynthia Venn’s seat-of-the-pants estimate was “less than six months.”
Recall that Lepas larvae are widespread throughout the ocean and aggressively colonize any available surface. If the flaperon had been in the water from March 8, 2014 to July 31, 2015, we would expect to find Lepas that were more than a year old.
So the barnacles weren’t the right age for an object that had been in the water for 15 months, and the way they were distributed implied a violation of the laws of buoyancy. Together, these observations suggested that the piece might have come to its final condition and location through something other than natural means.
Chapter 16
November 2015
If search officials shared my concerns, they weren’t letting on. Instead, the first piece of verified debris seemed to have reinvigorated their confidence that the plane was in the southern Indian Ocean. And they were increasingly confident in their math. “The experts are telling us that there is a 97 percent possibility that it is in that area,” Australian Deputy Prime Minister Warren Truss told the Wall Street Journal.
Behind the scenes a team of experts had quietly been developing new analytical tools to more precisely determine where the plane had gone. The effort was led by Dr. Neil Gordon, a scientist at Australia’s Defense Science & Technology Group (DSTG). Gordon’s team devised a mathematical approach based on Bayes’ theorem, a tool for gauging the probability of possible outcomes based on a set of initial parameters. They revealed the details of their work on November 30, 2015 in a paper entitled “Bayesian Methods in the Search for MH370.”
This approach was based entirely on BTO values—the timing data. The experts had come to realize that the BFO values—the frequency data—were shot through with too much uncertainty to add meaningfully to an understanding of the plane’s path.
It involved generating a huge number of random routes flown at different speeds with different numbers of turns and testing them to see which best fit the observed BTO data. It turned out that routes with high speed and no turns (or a few very small turns) fit the BTO data well, while most of those with more significant turns or slower speed did not.
The DSTG plotted the endpoints along the 7th arc. The result was a “heat map” representing the relative probability of the plane’s having ended up in any given location. It looked like a diffuse, elongated blob. To make sure their technique was valid, the scientists performed the same calculations on other flights in which they had both BTO data and the actual route flown. It checked out.
Based on their calculations, the DSTG defined a search area that stretched about 400 miles along the 7th arc, from latitude 35° south to latitude 39° south—a more compact area than the previously defined one. The results only reinforced investigators’ conviction that they’d been searching the right stretch of seabed all along.
The paper attracted widespread attention. But there was a detail buried within that was uniquely interesting to me. Page 85 of the report contains a Google Earth screenshot of the eastern hemisphere overlaid by the routes in the DSTG’s probability calculation. The routes all begin with a flight path running up the Malacca Strait. Then they fork into two subsets.
Recall that routes generated from BTO data alone are inherently ambiguous, in that they create an equally valid mirror image. The BFO data is then used to decide which one is correct. The caption for this image, however, explained that the sets of paths were made “using only BTO measurement weighting (i.e. not using BFO measurements).” On display, in other words, were both the route and its mirror.
One subset takes a hard left around the western tip of Sumatra and plunges into the remote Indian Ocean. This is the famous southern route seen in countless TV and newspaper graphics. The other subset takes a slight turn to the right, passes over the Andaman Islands and crosses the Indian coast near Calcutta. It then tops the ridge of the Himalayas in western Nepal and skirts the border between India and China before flying over Kyrgyzstan and into Kazakhstan.
The official assumption was that the BFO data was valid, and the plane had flown one of the subset of routes that turned south, and the subset that turned north were the spurious mirror images. If my suspicions were correct, however, and the BFO data had been tampered with, then officials had got it exactly backwards. The plane’s actual route lay to the north, and the southern routes were the mirror ones.
In effect, the DSTG had revealed where the plane went if the Inmarsat data had been spoofed.
Chapter 17
Radar
For as long as searchers have been aware of the existence of the Inmarsat data, there has always been one particularly compelling reason to suspect that the plane went south rather than north. If it had gone north, it would have passed through the air-defense radars of multiple countries, including the two most populous on Earth. Yet no one had reported seeing it.
It would seem common sense to assume that most nation's militaries routinely monitor their airspace. That turns out not to be the case, however. Military radar is expensive to build and requires a lot of electricity and manpower to operate. Unless there is a valuable target to defend, and missiles and planes capable of defending it, running a radar station 24/7 isn’t worth the cost. So in most parts of the world coverage is like Swiss Cheese in reverse: the gaps far outnumber the areas under surveillance.
“During the Cold War, we got used to the concept that the radar is constantly on and jets are scrambled if anything unexpected is seen,” Tim Huxley, executive director of the International Institute of Strategic Studies in Asia, told the Wall Street Journal in 2014. “We sort of expect that to be the normal response, but that doesn’t necessarily translate into comprehensive coverage in other parts of the world.”
If MH370 did fly north, what kind of radar environment would it have passed through?
Let’s start with the beginning of the route. Half an hour after leaving the Malacca Strait, according to the DSTG’s calculations, the plane would have passed over the Andaman Islands. The archipelago belongs to India, which maintains a radar station there. But the radar is only turned on when a crisis is looming, which wasn’t the case on March 8. “We operate on an ‘as required’ basis,” the chief of staff of India’s Andamans and Nicobar Command told Reuters.
Next, the plane would have crossed the coast of mainland India west of
Calcutta and passed almost directly over the air force base at Kalaikunda. The Russian-built Sukhoi Su-30 fighters stationed there are guided by the nearby radar installation at Salua. But if the Andaman radar was inactive, Salua likely was as well, since both facilities are geared toward defence of the same area. “Kalaikunda… has been a bridge with the Andamans,” an official told the Times of India in 2011. “The role of the base will grow and aircraft based here will play a vital role in patrolling the skies over the Andamans and the Bay of Bengal.”
The Indian military is notoriously underfunded and poorly equipped, so it's not surprising that it focuses its assets on the one region where it faces and ongoing threat: its border with Pakistan, 1,000 miles to the northwest. Relations between the countries have been tense since they achieved independence from Britain in 1947. In 1971 a surprise attack by Pakistani jets against Indian air bases and radar stations precipitated a 13-day war in which more than 10,000 soldiers died. Numerous unresolved territorial claims continue to foment distrust. On February 27, 2019, Pakistan shot down an Indian MiG-21 fighter that had crossed into its airspace.
Even when tensions are not running high, India routinely intercepts any planes that happen to cross into its airspace from Pakistan without adhering to correct air-traffic procedure. It is less vigilant about the Bay of Bengal. “India has an exceptionally large area to cover, a massive airspace and maritime space,” Huxley told the Wall Street Journal. “Looking toward the south, they wouldn’t have so much reason to expect adversary aircraft.”
Continuing northward, MH370 would have crossed into Nepalese airspace. Nepal is a small, poor country with no air force or air defense radar.
Passing west of Kathmandu, the plane would have traversed the spine of the Himalayas and entered Chinese airspace. For the next two hours it would have hewn to the far western edges of Tibet and Xinjiang, China’s two westernmost provinces.