Australia Confirms Zaharie Flight-Sim Route to Southern Ocean

In a posting to a section of its website called “Correcting the record,” the Australian Transport Safety Board today confirmed that the FBI found data on MH370 captain Zaharie Shah’s flight simulator hard drives indicating that Zaharie had practiced a one-way flight into the southern Indian Ocean, as I wrote in a story for New York magazine on Friday. Entitled “False and inaccurate media report on the search for MH370,” the post concerns several claims by Australian pilot Byron Bailey in The Australian, including Bailey’s interpretation of the flight-sim data:

Mr Bailey also claims that FBI data from MH370 captain’s home simulator shows that the captain plotted a course to the southern Indian Ocean and that it was a deliberate planned murder/suicide. There is no evidence to support this claim. As Infrastructure and Transport Minister Darren Chester said in a statement, the simulator information shows only the possibility of planning. It does not reveal what happened on the night of its disappearance nor where the aircraft is located. While the FBI data provides a piece of information, the best available evidence of the aircraft’s location is based on what we know from the last satellite communications with the aircraft. This is indeed the consensus of international satellite and aircraft specialists.

While ostensibly rebutting Bailey’s claims, the ATSB tacitly acknowledges the fact that the flight-sim data was in fact found by the FBI.

524 thoughts on “Australia Confirms Zaharie Flight-Sim Route to Southern Ocean”

  1. Jeff Wise posted July 30, 2016 at 9:38 PM: “It might indeed seem strange that a 777 line captain would want to play with a flight simulator, yet Zaharie not only manifestly did but shared his passion on YouTube.”

    His hobby was to build that simulator, to make it as realistic as possible, and to share that passion with others on social media.
    Just like repairing ice makers and window seals.

    He didn’t need a simulator for practicing or for obtaining performance information. As Brian Anderson has pointed out, much more accurate and quicker means were available to him. The manufacturers ‘Product Sheet’ for the B777 data package claims the performance to be accurate within 5%.

  2. @DennisW: “I am not going to play that game.”

    You don’t have to. But if you choose to ignore evidence that runs counter to your preconceived notions, then your voice doesn’t count.

  3. @Jeff Wise

    “turning left and flying to fuel exhaustion in the SIO. Why did he do that?”

    Maybe someone else did and left clues as to who helped with the clues, the boom whereabouts, and where the plane ended.

    Everything that happened with MH370 should be giving away that it was no one onboard that did this. But still everyone is stuck in a groove.

  4. @Dr.BobbyUlich
    you said:
    “As explained in the caption for Figure 2, the path from KLIA to IGARI is based on secondary radar, not primary radar. It is essentially a plot of GPS positions transmitted back from the aircraft. Therefore there is no slant range effect. That is why it is a straight line with no waviness. The remainder of the flight track is based on primary radar and appears to be affected by slant-range error.”

    If that would be the case then we would have an either intentionall mislead or we would look at the raw radar data. Radar presentations with ground or airspace structure overlay have to be range corrected to the overlay, otherwise the presented information would not only be useless, but dangerous.

    There are however other possible system inherent errors who could be cause for your observation.

    The most detailed and understandable documentation about radars I’m familiar with can be found in the following link. Change the link from …Book1.pdf up to …Book5.pdf to access all 5 books.

  5. Zaharie’s hobby was building things, fixing things. Just look at all the tools laying about on his desk.

  6. @DrBobbyUlich. Thank you for the clear, concise and comprehensive review.
    As you know, we need a new search proposal which is both demonstrable and which has good prospects (and, I think we can add, a very localised area). If it should be that a glide and ditching becomes more probable, the area/likely success/benefit/cost will lead to a that’s that. If a high speed crash remains the probability, then with search width theories unlikely to come up with the goods and drift theories unlikely to localise sufficiently, the prospects of a new search depends on confirmation of high speed descent and a new insight about route. Just thinking aloud.

    You were short of fuel in some instances (LRC and ECON 52 and the MRC True Track). Can I take it that your modelling takes earth shape into account? I remember George Tilson’s comment about that. With the closer distance to earth’s centre of about 20, 000 ft at around the search latitude the energy saved is equivalent to about 550 lbs fuel, net, so rather more gross; that is once divided by engine and propulsion efficiencies. The “descent” gradient is around 0.1% and an alternative way of calculating is to suppose an L/D up there of about 10, in which case by proportioning to normal consumption the fuel saved is about 800 lbs (my sketch calcs). But I do assume your model will account for this.

    As to going under budget on fuel, is there a reason why a lower altitude cruise would not fix this while alternatively accommodating fuselage (say) damage?
    How about continuous APU operation after around say, the 18:25 logon?

    For the Magnetic Track and True Holding Routes, partial searching really should not count.

    Just some comments in case of use observing how important a breakthrough with route alternatives may well be needed if the current search fails.

  7. Can anyone make sense of this quote from Bayesian Methods:

    Figure 4.2 shows the derived speed and heading obtained from this filter. [The measurement error was assumed to have a standard deviation of 0.5 nm and the process noise variance was 3:5104 nm2 s3. The process noise was adjusted to minimise the mean squared prediction error.] The speed estimates vary dramatically during the first turn, which is not an accurate representation of the aircraft speed at this time. It is likely due to the mismatch between the assumed linear Kalman filter model and the high acceleration manoeuvre performed by the aircraft.

    As I said earlier a “high acceleration manoeuvre” performed intentionally by Captain Zaharie doesn’t make any sense. VictorI has pointed out that such a manoeuvre implies zooming up, which would only have increased the airplane’s exposure to the radars.

  8. @David

    I never, ever said he simulated the 7th arc, so you definitely understand me Wrong. All I said was I believe he set up the 2nd logon. He would not have known about ping rings. He definitely would not have known that ISAT (BFO in particular) could be used to show he went south – nobody on earth knew this was possible, until Inmarsat actually worked it out! Truly, the ISAT data is a gift, as Dennis said. As for the ELT, nobody has a satisfactory explanation for why it didn’t transmit – the subject has already been discussed on this forum, at great length. The pilot cannot switch off the ELT from the cockpit, it is a self contained unit, with its own battery power.

    I think he set up the 2nd logon to let the world know the plane had run until fuel ⛽ exhaustion. I think he would have known in advance that the SDU went through a re-logon procedure each time it’s power supply was interrupted. The condition of the SATCOM can be monitored on cockpit MFD.

    I don’t think he would have allowed the engines to flame out at altitude, and just rely on the RAT until the APU started up and came on line,a minute later. And he wouldn’t know how long the APU could be expected to run for after LH engine flameout, which is why I think he made sure he had enough fuel ⛽ for a descent to low level,and be able to deploy the flaps before ditching.

    It remains a moot point as to why or indeed if he flew on as far as possible after simulating fuel ⛽ exhaustion. The fact that he hasn’t been found within the priority search zone, strongly suggests he did. However, he may not have glided as far as some have suggested (120Nm) but the condition and nature or the debris items recovered very strongly point to a controlled ditching, rather than a high speed impact. I certainly don’t buy RetF4’s suggestion of an aircraft that didn’t want to crash!

    Then there is the vexed question of the BFO data at 00:19, apparently indicating a high rate of descent. The pilot would not have known about the BFO side of things, and could not have manipulated it in any way. It seems very hard to believe that the aircraft could be descending at 5000fpm at one moment and 8 seconds later, be descending at 15,000fpm! coincidentally, at exactly the same time the 2nd logon was being transmitted! Too much of a coincidence, for my liking. It would have hit the water within 3 minutes. We would be seeing a different kind of debris, if that were the case. To me the most like likely explanation is that the BFO data is spurious, an artifact of the SDU logon, and due to callibration.

  9. A few notes on speeds:

    The lowest speed at which an airplane can develop a lift force equal to its weight is called the 1g-stall speed, Vs1g, in IAS. For a given airplane configuration it increases with increasing weight and, at higher altitudes, with altitude.

    At the weight existing at IGARI the 1g-stall speed Vs1g was at least 170 kt IAS at low altitude, probably higher at FL350. 170 kt IAS at ISA+10°C corresponds to 305 kt TAS at FL350, or 356 kt TAS at FL435. According to the DSTG’s analysis of the primary radar data, the groundspeed dropped below 200 kt in the turn-back past IGARI.

    After the 180° turn, the ground speed varied between 500 and 550 kt until the end of the continuous primary radar data available to the DSTG of “regular estimates of latitude, longitude and altitude at 10 second intervals from 16:42:27 to 18:01:49.”

  10. @DennisW

    “It is far more likely that Shah was acting out his desire to be a patriot, and rebelling against an obviously corrupt Malaysian regime. The facts which work against that theory and favor suicide are:

    1> When Shah knew he would run out of fuel before being able to safely land, he apparently made no attempt to communicate his position to maximize the potential for survival via search and rescue.

    2> Why did he not activate an ELT.”

    those facts could be valid if we were certain that he was the one under control at the end of the flight, but we are not

    we don’t have any clue what happened on the plane nor who was flying it at what time, but from SATCOM reset we can conclude there were some shenanigans going on and it was everything but the ordinary flight (regardless of destination not being original one)

  11. @Gysbreght
    “As I said earlier a “high acceleration manoeuvre” performed intentionally by Captain Zaharie doesn’t make any sense. VictorI has pointed out that such a manoeuvre implies zooming up, which would only have increased the airplane’s exposure to the radars.”

    Let me comment on this a bit.
    I assume the above statement from Victor is in regard to primary radar. At Igari MH370 at FL350 was within range of primary military radar. Increasing altitude or making an accelerated turn would not have increased the exposure, as it would not have changed the radar cross section. Descending from FL 350 below the radar horizon could have made a change though. Problem is, that a tight descending turn from cruising speed would be problematic concerning speed control and would pose the danger of overspeed combined with high descent rate and steep low nose pitch position. The initial climbing part of a high reversal maneuver therefore is to reduce the speed while already turning tight by climbing, followed by a descending part of the turn while increasing speed again. While descending could serve to drop below the radar horizon it would be a sensless try, as the intended path was heading in general direction of the radar position, and staying below the radar horizon was not possible.

    The reversal therefore would not be for reducing the exposure to primary radar, but to change direction as quick as possible and to confuse a controller in front of the radar scope, who would expect to find MH370 in the original direction it has travelled before. Just imagine the scenario when ATC calls the military control center and tells them they’ve lost radar contact of MH370 heading xxx° with speed xxx kts at FL 350 close to position IGARI 3 minutes ago. Where would he start his search for an non squaking primary radar return on his live scope? First in the direction of the original track and at a distance being flown in 3 minutes, then in a ever broadening sector left and right, and then somewhere in an eggshaped area around the last known position derived from known speeds and a possible turnback at standard turn rates.

    I came around on a post on pprune some time ago, and I think it fits in this discussion quite well. Let me thank the unknown poster ( I did not record rhe link and poster, but it might have been IAN-W), which I fully support from my expierience.

    He wrote:

    “Do a few long boring nights in an Air Defence Notification Centre when the system is based on a modern processed radar system. The first time the aircraft track is noted it is categorized. In this case a ‘friendly’ taking off from one of our airports add tag saying friendly. The aircraft position is a track not a radar response. All available surveillance and position reporting aids are correlated by a ‘multi-sensor-tracker’ and a track position displayed tagged friendly. So when all cooperative position reporting systems were turned off (or went off) the remaining primary track would continue with the same position symbol and tag so no change but a friendly turning back all inside local airspace.

    Civil ATC almost certainly had primary selected off – why have it as all their aircraft have SSR/ADS not interested in noise from other aircraft on primary. So the aircraft is on the edge of cover outbound being passed to Vietnam and _after_ it leaves my frequency I cease to be interested, so its secondary code disappears {{Shrug}} ‘someone else’s problem’.

    The one indication that could have been flagged is the departure North up the Malacca Straits, but as a tagged friendly aircraft outbound from the airspace it is not an ADNC threat and ATC in Malaysia cannot see it.

    By the time ‘overdue’ action was taken the aircraft was around the northern tip of Indonesia and turning South. When Air Defence was then asked what they saw, going back to the ‘tapes’, they found the aircraft. Probably similar for the Thai radars but they would be a little edgy giving away what they could see.

    No need to invoke tin-foil-hat theories. Having pulled many overnights on Saturday night to Sunday morning with ADNC it is tedious. If an aircraft is not inbound and it is tagged friendly, it is not a threat – back to cribbage.

    I should add, that the timings and route flown by the aircraft and the position of the cooperative aids going off, and then the altitudes apparently flown indicate that the person controlling the aircraft was aware of the way both ATC and ADNC were likely to react. It was one of the indicators for me that this was no automation run awry in some probability tail failure mode, but was intentional. Something that the authorities now seem to believe too.”

  12. @jeff
    I dont know if it is usual or not, nor if it was already mentioned, but for statictic crunchers, can be somehow interesting that on the passengers list was NO children/youngs between 5-18 years? (5 kids under 4 years, though – and I cant believe all the people ended somewhere in ocean, all the time, until hard evidence)

  13. @RetiredF4


    I hope everyone investigating this looks from every angle with that in their minds. Shah could not pull off something like this, neither would he have wanted to.

  14. @RetiredF4: Nice story, but a quick look at Figure 4.2 and the track that it represents should tell you that your story doesn’t fit the facts.

  15. @Klaus

    “And I have always wondered why there were never reports on what the Captain had been doing on his PC other than posting on Facebook.”

    The Daily Mail once reported that the personal laptop has not been found:

    “Zaharie’s personal laptop was not found, and is thought to have been with him in the cockpit of the plane, the source said.”

  16. @all,

    Ok I’m going to present a slightly different glide theory..

    I watched a bit of 60mins segment with Larry Vance..Who claims the flaperon was deployed which suggests a glide into the water..Looking a the pics from outboard flap, that also suggests a glide into the water..

    A lot of you are saying that MH370 must be much further south from current search area..

    But the drift models doesn’t back up this scenario..

    What if the flaps were extended much earlier and a programmed decent was set on the MCP..

    So the plane would burn more fuel at a slower speed and would not travel as far..Hence a crash impact further north near broken ridge as drift models suggest..

    I still can’t accept a crash much further south as the drift models does not support that.Debris would have washed up somewhere on the west coast of Australia by now..

    I know some of you have said impact would reduce the amount of debris..We have had debris from within the plane be found..So the likely hood of more that debris would increase therefore some of the debris would have found it’s way to west coast of Australia..

    Opinions welcome 🙂

  17. @Gysbreght

    @RetiredF4: Nice story, but a quick look at Figure 4.2 and the track that it represents should tell you that your story doesn’t fit the facts.

    I hope you are not categorizing myself as a story teller or a fiction writer.

    For the benefit of all would you explain what story or part of the story you are talking about and what facts of what Figure 4.2 you refer to? Maybe you missed to post a link?

  18. “60 minutes” Australia, Channel 9 program just shown, main points were that there was a strong case the aircraft had been piloted in a ditching with Peter Foley of the ATSB saying that the flaperon trailing edge separation could have been the result of it being deployed for that. He implied that the French had concluded that this was possible.

  19. @RetiredF4: I’m referring to Figure 4.2 of the book “Bayesian Methods in the Search for MH370” (Pre-publication Draft) by the Defence Science and Technology Group, Australia. The link can be found on the ATSB website. The essentials of what I’m discussing have been posted here at several occasions. The track corresponding to the speeds and track angles has been constructed independently by sk999 and myself and shows that the course reversal wasn’t any quicker than a standard turn would have been.

  20. 60 minutes again. Further to the above, Peter Foley said, “There is a possibility that there was someone in control at the end and we are actively looking for evience to support that”.

Comments are closed.