Deep Dive: MH370, Remastered

Today’s episode is something of a double-header, as we address two different but related topics: the scientific method, and how we the co-hosts came to be involved in the case. In the first half of the show, Andy and Jeff talk about their personal histories, and how their experiences prepared them for tackling the mystery of MH370.

In the second half, Jeff describes how working on this mystery has shaped his understanding of the scientific method, and in particular how scientist deal with uncertainties in data and in their knowledge of initial conditions. Understanding so-called Bayesian methods is crucial, because it’s the approach that search officials used in defining the the search area on the seabed of the southern Indian Ocean.

Bayes theorem has crept into the public discourse in recent years; you’ll often hear about people “updating their priors” and the like, meaning that they’ve changed their mind based on new information. In this episode Jeff refers to it as an inverse probability estimate, in the sense that rather than using one’s knowledge of a current system state to calculate a future state, you use knowledge of the current state to assess what previous conditions might have given rise to it.

There’s a neat example of this idea in The Book of Why by Judea Pearl and Dana Mackenzie. Imagine that you have a 10-foot-long billiards table. If you fling a cue ball so that it vigorously bounces off multiple sides and effectively winds up in a random location, what is the probability that it will wind up within 1 foot of the far-left hand edge? It’s easy to see that the probability is 10 percent.

Now imagine that you have a billiards table of unknown length and the cue ball is one foot from the left-hand edge. How long is the table? The answer you give depends on your knowledge of possible prior states. For instance, it may be the case that all billiard tables are either 10 or 14 feet long. Or maybe they can be arbitrarily long.

The searchers for MH370 found themselves in an analagous situation when it came to defining a search area using the Inmarsat data. They had to ask themselves, “What is the universe of possible flights that could have given rise to this data, and what is the relative probability of each?” This would then allow them to calculate a probability heat map on the seabed of the southern ocean of where the plane’s wreckage might be located. We know a lot about how they carried out this work because the Australian scientists who carried it out wrote a very detailed and lucid account which they published in 2015 as Bayesian Methods in the Search for MH370.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.deepdivemh370.com/subscribe

[Today’s episode is a remastered version of the 10th show I made with Andy Tarnoff as part of the original Deep Dive: MH370 podcast, in which we broke down the basics of the case step by step.]

Today we tackle our most controversial topic date: the question of whether a backdoor exist in MH370’s satcom system that would have allowed the BFO data to be tampered with. If so, it will have radical implications for what might have happened to the plane.

Back in Episode 7 we explained how the BFO data worked and how it showed conclusively that the plane must have gone south. But then some strange facts started to appear that, taken together, suggested that all might not be as it seemed.

The first was that, as we’ve discussed earlier, the Satellite Data Unit had been turned off and back on again. Officials didn’t let that slip until June 26. Up until that time, we’d all assumed that the satcom had inadvertantly been left on when everything else was turned off. The fact that it was turned on was really hard to explain. In fact to this day, people have had a hard time coming up with a convincing reason why anyone would either know how to do this or want to do this. I hasten to add that is a hugely contentious point. Some people say they have perfectly good explanations, but they all seem pretty daft to me. For instance, Victor Iannello thinks it was done in order to prevent someone in the cabin from using the satphone to call for help, but there are simpler ways to turn off the satphone.

The second surprising fact to emerge was that it turned out that the electronics bay, or E/E bay, the compartment that houses all of the computers that control the plane, is accessible through a hatch in the first-class cabin.

Matt Wuillemin, an Australian former 777 pilot, wrote a master’s thesis on the security vulnerability of the unlocked hatch in June 2013. In his thesis, Wuillemin notes that in addition to the Flight Control Computers, the E/E bay also houses the oxygen cylinders that supply the flight crews’ masks in case of a depressurization event, and the controls for the system that locks the flight deck door. “Information is publicly available online describing the cockpit defences and systems located within this compartment,” Wuillemin notes. “This hatch may therefore be accessible inflight to a knowledgeable and malevolent passenger with catastrophic consequences.”

The last piece of the puzzle fell into place when I had a conversation with an influential independent reseracher named Mike Exner. I asked Mike how the Satellite Data Unit gets the navigational information that it needs to calculate the Doppler precomposition that produces the BFO. Exner replied that it comes through a cable from a unit called the IRS (Inertial Reference System) that’s in the electonics bay.

If you climbed down into the E/E bay and disconnected the SDU cable from the IRS, you could plug it into an electronic device capable of generating false position information. Such gear would have to be manufactured from scratch; “there are certainly no commercial, off-the-shelf boxes like that,” said Exner. In essence, the signal would be lying to the SDU about where the plane was located and how fast it was going, causing the SDU to transmit at an incorrect frequency.

Later, Victor Iannello developed this idea even further. He pointed out that if you changed just one parameter inside the SDU, you would create BFO values that would make the plane look like it was going south when it was really going north. He wrote a paper exploring the idea that you can find here.

Two important points:

a) We don’t know definitely that this vulnerability truly exists. There may be some aspect to the 777’s electronics that we don’t know about that would rule this out. One of the interviewees in the Netflix documentary says that you can’t take over the controls of a 777 from the electronics bay. Others have made the same claim. I don’t know how they would know this, and no one has explained so far, but it’s possible they’re right.

b) Victor later became convinced by other evidence that the vulnerability, even if it did exist, was not used.

But! The mere possibility that the data could have been tampered with to create what amounted to a false tale of breadcrumbs was breathtaking.

Remember: at the time, the authorities were getting ready to launch an incredibly ambitious search of the remote ocean seabed based at this point solely on the basis of a handful of signals the likes of which no plane had ever transmitted before, whose origin they couldn’t explain, and whose meaning they couldn’t verify by any means, whether satellite photos, radar returns, floating debris, or anything else.

Everything was hanging on these signals which everyone just assumed had been produced innocently and inadvertently, with no intention on anyone’s part.

What we’d discovered was that that wasn’t necessarily true. There was another way to create these signals. So now, as I saw it, we’d have to put an asterisk next to every instance where someone says, “We know the plane went into the southern Indian Ocean.” Beause there is another way to interpret that data.

And here’s something else to consider.

For this exploit to have been possible, a whole bunch of things had to line up.

The hijacked plane would have to be a Boeing, not an Airbus. It would have to be a fly-by-wire 777 or a 787, not one of Boeing’s earlier planes which used mechanical rather than electronic control systems. The airline would have to be subscribing to the low-cost version of Inmarsat’s service, called Classic Aero, and not its high-end version, Swift Broadband. Its route after disappearing would have to have been entirely under the footprint of an aging Inmarsat satellite in a decaying orbit. And the implied direction would have to be toward an oceanic basin in which the plane could be “lost.” And the mirror routes to discriminate would have to lie on a north-south axis.

MH370 meets all these criteria.

Grant you, if this hack were carried out it would involved some incredibly sophisticated perps. But not unthinkably so. If you work in the field of electronic warfare (EW), the idea of changing the apparent Doppler shift of a signal is used all the time. (As I discussed in a recent article about UAP sightings and EW for New York magazine.) So for some people, this kind of idea wouldn’t be that wacky at all.

Now, not many people understand how this vulnerability works; a lot of people don’t get past the words “Doppler Precompensation.” And of the ones who do, a great many find it preposterous that this vulnerability was exploited.

But no one, to our knowledge, has demonstrated that it is impossible. And that means that there’s a whole alternate lens that you can use to look at the disappearance of MH370.

Things that seem to be random or to have occurred just by chance, like the reboot of the SDU, now seem to have a very specific purpose.

Anyway, by the end of June, 2014, I had started to wonder if the whole world had gotten MH370 fundamentally wrong.

And I thought that the answer to that question would emerge pretty soon. Because a seabed search was about to start, and the if the vulnerability had been exploited, then it would leave a very specific clue:

The plane’s wreckage would not be found on the seabed floor where the BFO data suggested it had gone.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.deepdivemh370.com/subscribe

In the weeks after the disappearance of MH370, many theories were proposed, but one in particular quickly came to the fore: that one of the pilots had seized control of the plane and flown it on a murder-suicide mission into the southern Indian Ocean. While there have been a handful of known cases where pilots have flown their own plane into the ground, no one had ever before carried out a sophisticated, complicated, and aggressive plan to abscond with an airplane only to spend seven hours patiently waiting to die. But there seemed no other way to easily explain the sequence of events that emerged from the Inmarsat data. What would such a person be like, psychologically? What kind of traces would they leave behind in their social media, in their personal photographs and work records, and in the memories of those who knew them? In today’s episode we turn our attention to Captain Zaharie Ahmad Shah and First Officer Fariq Abdul Hamid and take stock of the evidence.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.deepdivemh370.com/subscribe

From the first day MH370 went missing, it was the subject of an intense surface search. Planes, ships and satellites scoured millions of square kilometers of ocean. Not a single piece was ever spotted. In today’s episode we talk about how it went down, and what we might conclude from it. We also touch on a strange coda to the search, that involved an attempt to find the plane by listening for audible pings from the plane’s black boxes.

As we’ve previously discussed, at first everyone thought that the plane had crashed in the South China sea, under its original route to Beijing. At first this was a Search and Rescue mission, as authorities hoped that the plane might have ditched à la Miracle in the Hudson and some survivors could be rescued. But as time went by hope quickly faded.

At first, planes and ships searched the South China Sea and the Andaman Sea to the West. After the Inmarsat data was discovered, the search area shifted into the South China Sea. The hope was that once debris was spotted, its location would give them a rough idea where the plane had crashed. They would then be able to lower listening devices in order to detect automatic acoustic pings generated by the plane’s black boxes. Once the black boxes were found and retrieved, the data they contained would allow investigators to know exactly why the plane had crashed. The mystery would be solved.

This sequence of events has been followed numerous times, such as with the crash in 2009 of Air France 447, which we’ve discussed several times before.

Over the course of the next month, an armada of ships and planes searched over millions of square kilometers of ocean surface, but no debris from the plane was spotted. Frustrated, the investigators realized that the batteries on the acoustic pingers would soon run out. These pingers have an underwater detection range of one one or two miles, so without any surface debris to provide guidance the chance that you could lower a listening device into any random spot on the ocean and detect a signal are vanishingly small. Yet with time running out searchers figured they might as well give it a shot. At the beginning of April they started to listen. Lo and behold, they detected a signal!

The Australian government, which was now in charge of the search, was very confident that they had solved the mystery of MH370. The Prime Minister, Tony Abbott, declared: "we are very confident that the signals that we are detecting are from the black box on MH370.”

I (Jeff) was very skeptical. I went on CNN and pointed out that, among other things, the acoustic pings that were detected were of the wrong frequency. And I turned out to be right. The authorities sent down submersibles to scan the seabed where the pingers had been heard and found nothing at all.

By late May, officials started to admit that it had been a wild goose chase.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.deepdivemh370.com/subscribe

On March 24, 2014, the Malaysian Prime minister made a shocking announcement: using a new kind of mathematical analysis, scientists at the British satellite communications company Inmarsat had determined conclusively that MH370 had flown into a remote area of the southern Indian Ocean. Because there are no islands in the area, there was no possibility that anyone on the plane could have survived. Therefore, all 239 passengers and crew must be dead. It was a stunningly sweeping conclusion to reach based entirely on a kind of mathematics that no one in the outside world knew the details of. But was it correct?

Up until that time it had seemed to me that the plane more likely went north. It seemed implausible that someone sophisticated and motivated enough to steal a plane so aggressively would do it just to die a protracted death. In fact I’d written two articles for Slate making that case. At the time, most people already thought the plane probably went south, so my editor had only let me write the articles after I’d promised that I would write an apology article if I turned out to be wrong. I thought it was worth it, so I said yes.

After the Prime Minister made his announcement, it looked like I’d lost my bet. But I wasn’t 100 percent convinced yet. What was this math that he was talking about? Until I could see for myself what the basis for his claim was, I wasn’t ready to throw in the towel. But I had to wait, because the Malaysians remained secretive about the data and the process that Inmarsat had used to analyse it. Finally, they released the data, and the various independent investigators began working to figure out the mathematical process. It took some work, because the math was indeed a bit tricky, but before long they figured it out. It made sense. The Malaysian Prime Minister had been right — the Inmarsat data did indeed unequivocally indicate that MH370 flew into the remote southern Indian Ocean.

I wrote an article explaining how the math worked, and then another explicitly apologizing for being wrong. It wasn’t the outcome I’d wanted — for one thing, a flight to the south removed any hope that the passengers might still be alive — but part of science is a willingness to accept new evidence. And I was happy that the mystery of MH370 now seemed very close to being solved.

With the first set of metadata, investigators had been able to specficy that MH370 flew one of two quite narrow sets of possible routes, with a crash location very close to the end of them. With this second set of data, they now knew which route was correct. All they had to do was go to the end of the southern route, scan the seabed using sonar, locate the wreckage, retrieve the black box, and figure out what had happened. Voila. Case closed.

The probability map of where the plane probably ended up looked like this, with the most likely resting place in red:

But of course, when it comes to MH370, nothing is ever quite what it seems.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.deepdivemh370.com/subscribe

The First Law of MH370 is that the closer you look at it, the weirder it gets. A good example of this principle can be found in today’s episode, in which we explore how exactly the satellite communications system, or satcom, came to be turned off and back on again after the plane disappeared from radar. At first, most observers assumed that an inattentive hijacker must simply have left they system on when turning off all the other form of communication.

But careful analysis of the data revealed that that was not the case, and raised the crucial question: what procedure could have been used to turn the system turned off and off again, how much expertise would be required to do it that way, and what does this tell us about the perpetrators? Also, Jeff reveals evidence of a little-noticed turn hidden in the Inmarsat data.

This episode is a remastered version of one I made in 2013 as part of the first season of the podcast, “Deep Dive: MH370,” with co-host Andy Tarnoff.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.deepdivemh370.com/subscribe

As they studied the data that Inmarsat collected from the missing plane, Inmarsat scientists quickly realized that they could use it to narrow down the plane’s location to seven successive arcs. With a little more analysis, they found that these arcs implied that the plane took one of two routes to its end point. One route headed north, and the other south — but which is correct?

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.deepdivemh370.com/subscribe

The mystery deepens as, in the days after MH370's disappearance, the Malaysian authorities first deny, then confirm that their military radar detected the missing 777 as it turned back from its planned route to Beijing just a few seconds after it passed the final waypoint in Malaysian airspace. The plane reversed course, flew back over the Malayan Peninsula, then flew up the middle of the Malacca Strait toward the northwest before disappearing again over the Andaman Sea. We discuss the difference between primary and secondary radar, and why the timing of the turnback seemed suspicious.

Later in the show, we discuss the mysterious death of Putin critic Yevgeny Prigozhin, whose plane had blown up in flight shortly before we recorded the episode. I mention a similarity to the destruction of Malaysia Airlines flight MH17, whose perpetrators turned out to be Russian military intelligence, as I wrote about for New York magazine.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.deepdivemh370.com/subscribe

The mystery begins. Shortly after midnight on March 8, 2014, Malaysia Airlines Flight 370 took off from Kuala Lumpur, Malaysia and headed to the northeast, towards Beijing. For the first 40 minutes of the flight, everything was absolutely routine. But then, at 1.21am local time, the plane vanished from the radar screens of air traffic control. In today's episode, I discuss what it means for a civilian flight to disappear from radar screens, and why the timing of the disappearance might raise a red flag for investigators. I also talk about why it took so long for authorities to realize that something had gone wrong, and why the initial assumption was that the plane most likely had crashed close to its last known position.

This assumption, of course, was wrong. But it was to prove only the first of many times the case would defy expectations.

In the course of the episode, I mention a video about MH370 made by YouTuber MenTourPilot. If you’re curious, you can find it here.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.deepdivemh370.com/subscribe

The disappearance of Malaysia Airlines Flight 370 is a mystery unlike any other — stranger, more intricate, and more technically arcane. To grapple with it, we need to approach it with patience and an open mind. In Episode 1, I lay out my own experience with the case and lay out the expert-centric approach that this podcast will follow. The show’s first guest is Peter Waring, who was a lieutenant in the Australian Navy in 2014 and participated in the search for the wreckage of MH370 with the Australian Transport Safety Bureau. I discuss how best to approach this unsolved case, laying out the journalistic and scientific methods that will be used to peel away the layers of the onion in each upcoming podcast episode.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.deepdivemh370.com/subscribe