French MH370 Investigators Eye “Spoof” Scenario

Interest in MH370 revived earlier this month after next-of-kin Ghislain Wattrelos held a press conference at which he revealed that he had been briefed by French judicial authorities about their investigation into the case. As the UK’s Daily Star reported,

Ghyslain Wattrelos lost his wife Laurence, and two teenage children Hadrien and Ambre when Malaysian Airlines Flight MH370 vanished on March 8, 2014, with 239 people on board.

Mr Wattrelos today revealed he was told by the French Gendamarie Air Transport (GTA) team investigating the jet’s disappearance they had found “inconsistencies” in the Malaysian investigation’s official report.

He claimed experts are investigating if navigation data from the missing plane could have been hacked to disguise the route it took before crashing into the ocean.

He also said he had been told several “curious passengers” warranted further investigation – including a Malaysian aeronautics expert seated directly beneath the satcom.

This was of course enormously interesting to me, as I had publicly pointed out in early 2015 that if the plane wasn’t in the southern Indian Ocean, the only conceivable explanation was that hijackers outside the cockpit had managed to perpetrate an extremely sophisticated hack of the satcom in order to make the signals seem like they were coming from a plane heading south when it was actually heading north. This idea met with widespread ridicule at the time, as most experts believed that the plane would certainly be found in the southern Indian Ocean where the satcom signals indicated it had flown. Subsequently, of course, it wasn’t–nearly a quarter billion dollars was spent on a seabed search that covered an area the size of the UK but turned up nothing.

At last, it seemed, the authorities were willing to take my idea seriously.

The Daily Star contacted me for a follow-up article:

[Wise] told Daily Star Online: “This (hacking lead) is an interesting development, because it’s exactly what I’ve been talking about for the last five years or so.

“While I haven’t looked at this particular passenger, the core of the argument I’ve been trying to make is that the Satellite Data Unit, or SDU, has a vulnerability that could be exploited to make the plane look like it went south when it really went north.”

He added: “What I pointed out is, are there any way these signals could have been tampered with?

“Is there some way that someone with ill-intent could have changed them?

“The answer is yes, there actually is a way that it’s physically possible that a person could get into the electronics bay, or directly access the data unit from ceiling of the cabin.

“And they could alter either the inputs into the SDU itself in such a way it would look like the plane was going south when it was going north.

“Do we have any reason to believe that’s the case? I would say yes.

“I think the main and most obvious one is having searched the seabed, based on signals of where the plane went, the plane is not there.”

Inmarsat data has led investigators to believe the plane crashed into the Indian Ocean west of Australia after running out of fuel.

But he has urged a re-analysis of this information, claiming that it could in actual fact have flown north instead.

The radius of one of the “handshakes” runs through Kazakhstan.

And Wise holds Russia as a suspect because of the shooting down of MH17 by a Russian military missile, and how the Kremlin’s annexation of Crimea fell off the news radar following MH370’s disappearance.

He told us: “What’s the motive? I can tell you this was happening in the context of Russia getting a lot of heat for the annexation of Crimea.

“I was on CNN six times a day, and CNN didn’t talk about Crimea anymore, they only talked about MH370 and so it was possible a diversion, a show of dominance.

“Because if I’m right and Russia did take the plane, they completely fooled, ran circles, the western authorities and experts have been completely bamboozled, with their pants caught down.

“I would say, only one other 777 has ever been lost mid-flight, that was the sistership of MH370.

“It was shot down by an operation carried out by the GRU. If you’re a chicken farmer, and you’ve never lost a chicken in 15 years, then you find one of your chickens murdered, and a week later you see a fox jumping over the fence with a chicken in its mouth, what would you think?

“What would be your primary suspect here? The only known cause of 777s coming to grief.”

All of which I stand by. I think the headline was unfortunately sensationalistic and misleading, however: “Plane ‘HIDDEN in Russian base’ as investigators swoop on new ‘hacking’ lead.” I’ve never said that I thought 9M-MRO is hidden on a Russian, and certainly not in all caps–though I am intrigued by the possibility that the plane might have touched down on the remote airstrip at Yubileyniy within the Baikonur Cosmodrome.


243 thoughts on “French MH370 Investigators Eye “Spoof” Scenario”

  1. He also said he had been told several “curious passengers” warranted further investigation – including a Malaysian aeronautics expert seated directly beneath the satcom.

    This shouldn’t be difficult to find out, anyone already looked this up?

  2. @Gysbreght: “The radar data showed not only speeds that were well above approved limits, but alao speed variations that are not compatible with autopiloted flight.”

    Do you mean autothrust ?

    One explanation for the speeds above approved limits (not possible with autothrust AFAIK) is a fast getaway.
    I would put the speed variations down to measurement errors/jitter.

  3. @David

    On page 2 of your previous post, we see that the Flight ID/ICAO Address is transmitted to the SDU (Satellite Data Unit) ONLY from the Left AIMS Cabinet.

    So was the Left AIMS Cabinet disabled?

    In figure 1.9A of the Factual Information Report, we see that the Air Data for the Left Transponder is from the Left AIMS Cabinet. A disabled Left AIMS Cabinet would also fail the Left Transponder.

    The Engine Health Monitoring (EHM) Report was not received, and according to the Factual Information Report (, is generated by the Aircraft Condition Monitoring System (ACMS).

    Is the ACMS also located in the Left AIMS Cabinet?

  4. @Peter Norton: “Do you mean autothrust ?”

    No, I mean autopilot. With autothrust off and throttles in max position (as for “fast getaway”) the autopilot would maintain constant speed and constant heading/track.

  5. @!Peter Norton:

    P.S. The overspeed protection embedded in the PFCS wouls also prevent those excessive speeds on autopilot, requiring pilot input on the control column.

  6. @Gysbreght: You are saying that the overspeed protection can only be exceeded with both A/P AND autothrust OFF, right ?

    Then we can conclude that after IGARI A/P was …

    • intentionally not used, for a fast getaway (IMO most intuitive answer due to excessive speed)

    • or not available due to an “electrical configurations that result in the loss of SATCOM and A/P” (your idea)

    • or intentionally not used, in reaction to an emergency

    • or not available as a result of an emergency

  7. @Peter Norton:

    Reply to your question: Yes
    First and second bullet: Agreed

    The seven hours of continued flight after the turnback is not compatible in my view with an emergency other than intentional diversion.

  8. @Peter Norton: Do you think someone wanting a “fast getaway” would start by climbing from FL350 to FL400+ ?

  9. @Gysbreght, We all need to remember that primary radar data has large positional uncertainty and is not reliable for determining altitude. It cannot tell us conclusively whether or not the plane flew on autopilot, it cannot be used to infer instantaneous speeds, and it cannot tell us how high the plane was flying. However, by looking at the totality of the data, it is clear that the plane was flying fairly fast (probably close to Long Range Cruise, or LRC), and this in turn implies a fairly high altitude.

  10. @Jeff Wise: Primary radar data does not have “large positional uncertainty”. You are correct that civil PSR’s do not measure elevation (and therefor cannot determine altitude). Although military radars measure elevation you are also correct that the altitude determined has a large margin of uncertainty.

    However, Mike Exners detailed analysis of the available civil primary radar postional data versus time (taking into account the slant range from the radars to the target, in particular close to the radar heads) has shown convincingly that the airplane must have flown at altitudes well above Fl400 (geometrical height ~~43000 ft) and at Mach numbers exceeding 0.89, well above LRC and MMO. Those results then roughly corroborate the military radar speeds and altitudes.

  11. @Jeff Wise: Mike Exner’s analysis is factual. It is more than an opinion. Therefore I don’t agree that you are entitled to your opinion on this point. Have you studied what M.E. has done?

  12. @Jeff Wise: You need to understand the methodology employed to arrive at those results. Then you’ll understand what I mean by ‘factual’. Assuming lower heights/altitudes would result in even higher speeds which are quite unrealistic.

  13. @Gysbreght, When your methodology is more granular than your data, you wind up interpreting noise as signal. In my estimation Mike got lured into a faulty and unnecessary process because he got his hands on a new set of data which at first glance allowed a more detailed understanding of the plane’s behavior. I think he wound up overinterpreting it.

  14. @Jeff Wise: The method was first proposed by Mike Exner, who was joined by Victor Iannello. It took me a long time before I had to admit that there was no other way to match the radar data.

    You are wrong when you describe it as interpreting noise as signal. It is best explained across the ‘cone of silence’ (CoS) when the airplane passes overhead the radar and there is no signal and no noise, but the difference between slant range and horizontal range is large and therefore the calculated groundspeed depends heavily on the assumed altitude. There is only one altitude at which the groundspeed calculated across the CoS matches the speeds before and after the CoS.

    That principle applied to all the PSR data from the civil radars at Kota Bharu and Butterworth results in a reliable estimate of altitude and ground speeds.

  15. @Oxy. ACMS component locations are the airplane condition monitoring function (ACMF) in the AIMS left cabinet, user access being through the maintenance access terminal (in the cockpit) and portable maintenance access terminal (PMAT) in the MEC.

    I appreciate your intent is to substantiate a theory you have propounded for some time. I hope you have enough TM info now for that.

  16. Gysbreght: “@Peter Norton: Do you think someone wanting a “fast getaway” would start by climbing from FL350 to FL400+ ?”

    Setting aside your discussion with Jeff about the radar altitude reliability, if the climb was real, one explanation was cabin depressurization. In case this somewhat inefficient maneuver was put up with as a necessary part of the plan, this wouldn’t be a reason to dawdle afterwards, isn’t it?

  17. Gysbreght: “Although military radars measure elevation you are also correct that the altitude determined has a large margin of uncertainty.”

    It depends on the precision of calibration.

  18. Gysbreght: “There is only one altitude at which the groundspeed calculated across the CoS matches the speeds before and after the CoS.”

    Where can this analysis be read ?

  19. @Peter Norton: “Where can this analysis be read ?”

    On the VI blog. You’ll have to search a bit, the discussion was a few months ago.

  20. @Gysbreght

    Thanks Gysbreght for that, but I remain a little confused.

    If the angle between the radar lobes on entering and leaving the CoS remains the same, then the distance between those two points would be greater at greater slant ranges.
    For any given time interval, the speed across the CoS must increase with increasing slant range.
    It is simply a case of v=d/t.

    That said, if t is constant, an increase of d requires an increass in v.

    In this case however, since the CoS is 3 dimensional, and the aircraft track passes relatively close to the antenna, that slant range “d” is a significant function of both altitude and nadir ground track relative to the antenna site.

    Thus, it adds a significant complication in determining the actual ground track, which itself has to be itteratively assumed.

    So, in fact, to do that, you have to make two simultanious and mutually dependent assumptions, ie, the nadir track, ie, the entry and exit points in the two dimensional ground track projection, and the altitude at entry and exit in the third dimension. That becomes complicated.

    The only way I can see of doing it, is by analysing the inbound and exit tracks from as far out as you can, to minimise the slant range component, ie, to nail down, as best you can, the inbound nadir, and the outbound nadir. This you can only do at the extremities of range, and work inwards (forwards) on the inbound track, and again inwards (backwards) on the outbound track, to meet signal loss at the CoS.

    I assume from your graphs that this is what you did ?

    If so, by definition, a higher altitude places the nadir track closer to the antenna than a lower altitude nadir track, does it not ?

    Thus the “d” for the high altitude CoS interval would be less than for the “d” for the lower altitude CoS interval, assuming that the radial displacement of the nadir track at point of closest approach to the antenna (near to, but not necessarily actualy at the mid point of the 180 second CoS outage) from the radar antenna for both cases is relatively small and a low fraction of the altitude difference.

    Thus, for the same time interval “t”, the “v” for the high altitude CoS interval would also be less than the “v” for the lower altitude CoS interval.

    Your plots seem to suggest the reverse.

    What have I missed ?

  21. The Malaysian 1MDB scandal is the gift that continues to give, and the article below starts to give us a few clues to the brains responsible for it…

    “New Goldman Sachs boss faces 1MDB fallout.”

    ‘Wall Street’s most famous bank finds itself under scrutiny following charges against former employees.’

    I’m still open-minded as to whether or not there is a direct link between 1MDB and the loss of MH370, but the fallout from this, and the MH17 shoot-down, certainly helped some wealth western folk make a killing from the resultant Malaysian Airline crisis fire-sale.

  22. @Ventus45:

    In my simplified example we have the slant ranges at entering and leaving the CoS, denoted as R1 and R2 in the scetch below. The distance AB between the entry and exit locations depends on the assumed height, and increases for lower heights at the same slant ranges.

    In the real case the aircraft track passes the radar at some distance (approx. 5 NM), and for some reason the entry and exit ranges are not equal. Also the speed and altitude is not constant, there is ‘noise’ from measurement errors and quantization, and therefore finding the best ‘fit’ by iteration of assumed height leaves some room for judgment.

    The fact that the track passes the radar at some distance does not cause a major complication, because the X,Y location of the track is defined by slant range and azimuth for each assumed height Z.

    Does that explain it?

  23. @Ventus45: Perhaps I should have added that the CoS is more a phenomenon than a clearly defined 3D space. The CoS is not used in the analysis. The analysis uses the recorded radar data: time, range, and azimuth. For each assumed airplane height the radar data define nadir track and groundspeed. The height is iterated until a best fit is obtained.

    So in the simplified example where the airplane passes directly overhead the radar site at constant altitude, speed and track angle the best fit would be obtained for 40,000 ft height.

    In the real case a good fit is obtained for 43,000 ft height, which corresponds to about 40,500 ft pressure altitude for the ISA+ conditions prevailing on the day of the accident.

  24. @Susie Crowe
    @Jeff Wise

    Further to the links on Russian aviation shenanigans….

    @Boris Tabaksplatt

    I don’t see how your linked story shows Goldman Sachs, which I have otherwise no need to defend, is responsible for the theft of 1MDB. Tim Leissener has pleaded guilty and admitted that he lied to the company’s compliance officers. The firm also previously rejected Jho Low as a private client until Leissener concocted false diligence for him. The firm isn’t being accused of stuffing anyone’s billions in its own pockets. This is just ridiculous.

    And as for a Malaysia Airline System fire sale–and I say this knowing how desperately you try to cast blame on the West and its institutions for MH370–minority share holders received a 12 percent PREMIUM from Khazanah on the 30 percent of shares it didn’t already own–this despite a 20 percent drop in price in the weeks after MG370 and an over all 80 percent drop in the prior FIVE YEARS. Even so, the deal was only worth US$400 million–or just a bit more than a single Boeing 777… So I’m not sure who made the “killing” you refer to in your post.

  25. Dear Mr Jeff Wise,
    I really appreciate your efforts and perspective.
    I would appreciate if you would pass me your contact to communicate my personal views being a person who has been studying the case since almost day one. I have given an interview to IB Times Uk, exchanged more than 50 emails to ATSB / Deputy Prime Minister of Australia ( which they replied promptly). In addition with experts like Top Incharge of Immarsat, Dr Duncan Steel and Air Crash Investigation experts, George Hatcher etc.
    Two points :
    1. I have a strong feeling that the guys in Maldives saw the plane on that day
    2. Maybe the initial leg of the flight up to Vietnam was spoofed. It might have travelled to the direction of Maldives immediately after departure from KL. This justify the manoeuvre above Malaysia up to West of Malcca Straight.

  26. Victor Iannello has ‘contrived’ a dive-glide-dive scenario and presents it as the only way the flight could have ended more than 25 NM from the 7th arc.

    I suggest the following is more likely to have occurred (*):

    After pushing over to near-zero ‘g’ the airplane is in an overspeed condition. The pilot pulls up into a climb to bleed off airspeed. He then has to push over again to prevent the speed from getting too low. He continues the descent but without autopilot his technique is not perfect. His airspeed varies between 180 and 230 kt IAS, and his rate of descent between 1000 and 4000 feet/minute. Twenty minutes after fuel exhaustion and 80 NM past the 7th arc the airplane hits the ocean with an airpeed of 200+ kt IAS and 3000 ft/min rate of descent and breaks up.

    (*) I remind readers that I do not subscribe to the “default mainstream scenario that MH370 was hijacked by its pilot, Zaharie Ahmad Shah”.

  27. @Mohamen Shareef, The guys in the Maldives did not see MH370. And the initial leg up to Vietnam was not only recorded in ACARS and ADS-B transmissions but also observed on the radar systems of more than one country. So, unfortunately, you’re barking up the wrong tree.

  28. @Gysbreght
    Was that VI video accompanied by a text post? I do not recall that. It is similar to my recent skipping crash video post, which I think possibly explains the flaperon appearance. It’s a flaps up hard ditch.

    Not the expert, but I do *not* think that is a survivable crash. I believe the aircraft would break in half like the Lion Air picture of a prior Lion Air crash.

    Here is my recent list of possible reasons for a active descent at end of flight. How would you describe your scenario? (in other words, can I add a new reason to my list?)

  29. @SteveBarratt:
    “…Yes one does wonder about the connection, particularly if 1MDB finances benefited from MH370.” Link to article here…

    Difficult to pin down what happens across the different Malaysian Government funds, they are legally able to slosh the money around several different buckets, and what happens is not so transparent. However, found this snippet, which may be a clue.

    “Najib Razak used money raised from a deal with sovereign wealth fund Khazanah Nasional Berhad to pay for some of the liabilities of state fund 1MDB…”

    The ~$1B Khazanah Nasional Berhad fund owns many large assets, including Malaysia Airlines, but I can’t find any confirmation that the money came directly from MAS.

  30. @TBill: As I recall it, there wasn’t much text accompanying the video. I tried to find it but Victor’s archive only goes back to 2017. Perhaps you could ask him.

    Regarding your “list of possible reasons for a active descent at end of flight” I would say that after fuel exhaustion a descent is inevitable, pilot or no pilot. That it was active follows from the BFO’s which cannot be explained at that time in a passive scenario. A rate of descent of 15,000 fpm is excessive in any piloted scenario, and I interpret it as an inexpert pilot over-reacting to something like a stall or depressurisation warning. That stall itself would be unlikely to occur with a qualified pilot.

    At the top of my list (and yours) is a stall, stall warning or perhaps the warning that the autopilot has disconnected.

    The only other reason I can think of is a fight in the cockpit. After the loss of both engines the cockpit door security system is unpowered so unless the door was mechanically bolted any cabin occupant would have access to the cockpit.

  31. Gysbreght is correct. The KB Civil PSR data analysis conducted by Victor, Bobby, me and others (independently) all found similar results. I believe my modelling and altitude selection logic was the best, but the differences were small compared to the main takeaway…very high and fast…not consistent with any plan to land.

    I used Victor’s spreadsheet for the core math, which is very accurate. No question about the math. I confirmed with people in KL that the KB radar measured range error was OTOO -0.03nm (negligible). I have high confidence in the results. The method was also validated using contemporaneous KB CIVIL PSR and ADS-B data on the flight passing KB earlier this year.

    Using my model, I found that 9M-MRO passed KB at ~43,000 feet (Geo, not pressure) with a true airspeed (TAS) close to 500 kts. The track clearly indicated the plane was being flown by hand (curved path, not straight). The military reported very similar speeds, virtually identical path, but unreliable altitudes (as expected). ATSB passed the analysis along to the working group who found the analysis convincing and consistent with some of their own assumptions.

    The KB Civil PSR data does have some noise, but it is far below the signal level in this case. It results in a very small altitude uncertainty compared to the inherent uncertainty associated with the altitude selection method (~±1000 feet).

    Note that the data also indicates that the TAS at 17:31 (~10 minutes past IGARI) was only 430 kts, which is close to stall. This indicates the plane was almost certainly in a maximum performance climb between 17:21 and 17:31. Between 17:31 and 17:37 (KB), the TAS accelerated from 430 to 500 kts. Between 17:37 and 18:02 (last BW data), the TAS remained ~ 500 kts. From the LIDO Hotel radar data, the speed remained 495-500 kts.

  32. @Gysbreght. “That it was active follows from the BFO’s which cannot be explained at that time in a passive scenario.”

    In the ‘normal’ unmanned configuration it is possible a left engine relight accessing residual left tank fuel could have brought the BFO timing well forward in Boeing simulations. As you know neither that nor APU auto-start were part of those.

    Case 5 (though unusual) was what I had in mind.

  33. @Scott O:

    A few hints here that all is not what it seems at Goldman Sachs, although in fairness the same could probably be said about the other mega finance outfits…

    ‘Goldman Said to Get Subpoena Over Its Role in Financial Crisis’

    “In April 2010, the Securities and Exchange Commission filed a civil fraud suit against the company. Goldman was accused of creating a mortgage product that was intended to fail. The company settled with the S.E.C., agreeing to pay $550 million without admitting or denying guilt.”

    ‘Goldman Sachs Director Tied To Insider-Trading Scandal.’

    “The director [Rajat Gupta of GS] told indicted hedge-fund manager Raj Rajaratnam about a $5 billion investment famed investor Warren Buffet made in the investment bank in 2008 before that information was made public. Rajaratnam has been charged with trading on insider information.” – Gupta later resigned from GS.

    I could go on, but the list is long. However, these two links regarding 1MDB are revealing…

    ‘Goldman Sachs bankers ‘cheated’ Malaysia over 1MDB: PM Mahathir’

    “Obviously we have been cheated through the compliance by Goldman Sachs people,” said PM Mahthir, “The bank’s compliance controls don’t work very well.”

    “Goldman Sachs Stock Slumps as Cloud of 1MDB Fund Scandal Darkens.”

    ‘Investors seem to be growing more worried about the potential fallout from the multibillion-dollar 1MDB fraud scandal on Goldman Sachs Goup, driving its shares down 11% over the last two sessions.

    Shares of Goldman tumbled 9% on Friday, following reports that Lloyd Blankfein, the firm’s current chairman and former CEO, had twice attended meetings with the figure at the center of the scandal, Jho Low.’

    Unfortunately Goldman Sachs will survive come what may, as it is yet another financial megalith which is too big to fail.

  34. @David:

    I think you are entertaining a mix of phantasy and “extraordiary coincidence” (to use Victor’s expression).

    Unless shown to be possible by analysis and/or test I don’t believe that “it is possible a left engine relight accessing residual left tank fuel ” could occur.

    Even if engine relight on residual fuel were possible, I don’t believe it would have brought the BFO so much more forward as you suggest, considering the Boeing simulations in abnormal electrical configuration where the left engine continued to operate at unchanged thrust after autopilot disconnect.

  35. @airlandseaman: according to your records, what was the date and source of the earliest official confirmation of the scheduled routing (i.e. the waypoints/routes) for the MH370 flight which took off on the evening (UTC) of March 7, 2014? Wikipedia is silent on the point, and you have always seemed to have the evidence in hand when it comes to MH370. Thanks in advance for your thoughts.

  36. @airlandseaman, Thanks for this. Is there a place where people can read your complete analysis?

    Also, you write that the plane accelerated to around 500 kts TAS, which seems plausible in the light of earlier discussions. I recall that some analysis of this primary radar data pointed to speeds of 520 knots and more, which seemed somewhat astonishing.

  37. JW: Here is one paper I wrote in April 2018 and updated twice. Page 11 was added last night, although the graph was generated in April 2018.

    The ground speed past KB was about 530 kts, but there was a strong tail wind. TAS was about 500 kts, as indicated on page 11.

Comments are closed.