New Potential MH370 Debris Found on Mauritius — UPDATED x3


The photo above is from an article on a French-language website. It says that the object was found two weeks ago by a French tourist, who gave it to a boat captain, who only gave it to the authorities on Tuesday, May 24. The piece is 80 cm by 40 cm and was discovered on a small island called L’ile aux Bernaches, which lies within the main reef surrounding Mauritius. It is now in the possession of the National Coast Guard, who will pass along photos to the Malaysians and, if they deem it likely to be a part of the missing plane, will send experts to collect it. (According to a second story here.)

The photograph above is the only one that seems to be available so far, and is quite low-res, but it seems to lack any visible barnacles, but has quite a lot of the roughness that barnacles leave behind after they’ve detached, as seen in the Mossel Bay piece. Perhaps worth noting that so far, pieces found on islands (Réunion, Rodrigues) have had substantial goose barnacle populations living on them, while pieces found on the African mainland have been bare. This piece breaks that trend.

Also worth noting, I think, is that all of the objects discovered so far were found by tourists, with the exception of the flaperon, which was found during a beach cleaning of the kind that only happens an tourist destinations. Drift models predict that a lot of the debris should have come ashore on the east coast of Madagascar, but this is not a place that tourists generally frequent. There are also large stretches of the southern African coast that probably see little tourism. All of which is to say that a concerted effort to sweep remote beaches should turn up a lot of MH370 debris.

I haven’t seen any speculation yet as to which part of the plane this latest piece might have come from–any ideas?

UPDATE 5/25/16: In a surprising coincidence, another piece of potential debris has also turned up on Mauritius. According to Ion News, the object was found by a Coast Guard foot patrol along a beach at Gris-Gris, the southernmost point on the island. It was found resting about six meters from the water.


UPDATE 5/26/16: In another surprising turn of events, Australia’s Minister for Infrastructure and Transport, Darren Chesterhas issued a media release in which he “confirmed reports that three new pieces of debris—two in Mauritius and one in Mozambique—have been found and are of interest in connection to the disappearance of Malaysian Airlines flight MH370.”

The release goes on:

“The Malaysian Government is yet to take custody of the items, however as with previous items, Malaysian officials are arranging collection and it is expected the items will be brought to Australia for examination,” Mr Chester said. “These items of debris are of interest and will be examined by experts.”

This means of announcing findings related to MH370 marks a departure for the Australian government, which in the past has provided updates from the ATSB (Australia Transport Safety Board) itself. The items are picture below, courtesy of Kathy Mosesian at VeritasMH370:

Mozambique 3A small Mozambique 3B small


Meanwhile, a reader has provided an image analysis of the second Mauritius fragment in order to provide a sense of scale:

size analysis

He observes: “Some rough scaling puts it at around 14 by 26 inches. Those boulders in the other photo look like pebbles; makes it look the size of one cent piece. Note the increasing curvature left to right; ups the bet on a chunk of flap!”

UPDATE 5/27/16: Another piece turned up yesterday, making it four altogether since Wednesday. I think this qualifies as a “debris storm.” At the rate stuff is turning up, there should be a lot more to come. There hasn’t even been an organized search yet!

The BBC reports:

Luca Kuhn von Burgsdorff contacted the BBC on Thursday to say he found the fragment on the Macaneta peninsula.
The authorities have been notified. The piece must be examined by the official investigation team in Australia.
Experts say it is consistent with where previous pieces of debris from the missing plane have been found.
Mr von Burgsdorff took two photographs of the item on 22 May, and sent them to the BBC after reading a story on Thursday about other debris finds in the region.
He said the pieces were “reasonably light, did not have metal on the outside, and looked extremely similar to photos posted on the internet of other pieces of debris from aeroplanes”.


697 thoughts on “New Potential MH370 Debris Found on Mauritius — UPDATED x3”

  1. @Marc
    I found and submitted to this blog the following MH370 Flaperon report because I had not seen it before and was wondering if it had been seen before.

    The report on the surface sounds very professional with lots of documentation. However: Too many unimportant details are included to sound professional. Give credit for the nice sketches of the missing sections and fitting.

    Based on many years supporting Boeing structural tests (including explosive and fatigue tests), I find this report to be mostly complete nonsense.

    Very poor conclusions. The part ID tag? It is not even the same part / location; features don’t match up; and it’s not important. To use Asiana Air 214 as an example; It hit the sea wall and landed on the runway.

    Any fatigue analysis requires detail review of the stress cracking of the fittings noted in the report. With only these photos to review I find this impossible. I have personally sat through detailed crash analysis reports showing how the fatigue occurred. You need micrographs / close up photos at a minimum for proper analysis.

    To conclude that the Flaperon came off in flight and did not detach upon impact with water is ridiculous. If MH370 hit the water at a speed near power out glide speed the flaperon could be mostly intact or torn to pieces; depending on how it impacts the water and how the other parts of the airplane impact it. Anything can happen in a water landing; and it has; from virtually no damage (e.g. Hudson River landing at stall or complete destruction at high speed with small pieces remaining ( recent crash in Russia on the runway / MS804 when found).

    In My Opinion.

  2. Fig.6 of “Refined MH370 Location Based on ECON Speed Control” by Dr. Ulich shows that all standard FMC speed-control methods (MRC, ECON, LRC) would have propelled MH370 to 7th-arc locations southwest of 37.5S, 89.3E. Similarly, Fig.5 of “MH370 – Flight Path Analysis Update” by the ATSB concurs that “Constrained AP Dynamics” generate candidate flight-paths which all cluster within a degree or so of that same location, 37.5S, 89.3E.

    However, the path favored by Inmarsat’s JoN article, terminates much farther northwest, near 34.7S,93.0E. Similarly, the ATSB’s afore-mentioned Fig.5 concurs that “Data Error Optimization” generates candidate flight-paths veering farther to the northwest, up to 32.5S.

    These considered calculations may imply, that MH370 did not proceed southwards on AP with a FMC-controlled speed.

    This author duplicated the Inmarsat JoN article BFO model, and found that minor tweaks to their proposed path, in terms of ping-ring crossing locations — including adjusting the final crash site to -34S — dramatically reduced the RMS BFO error, from 7Hz to 1Hz. However, these minor modifications, of slewing ping-ring crossings slightly, imply a variable airspeed, fluctuating from 750-880 kph (410-480kts). Note, the changes in speed imply changes in Kinetic Energy (KE) equivalent to the changes in Gravitational Potential Energy (GPE) resulting from altitude fluctuations of almost a km either way vertically (+/- 1000m). Likely, Inmarsat required straighter, leveler, and more constant speeds, in the flight-paths that they considered, and so excluded the more volatile flight-path proposed above, so permitting this author to reduce RMS BFO error, over the ghost-flight phase, from 7Hz to 1Hz, obtained by freely adjusting ping-ring crossings, with an eye only towards the BFO error, without any a priori demands upon either airspeed or heading.

    This author offers that MH370 flew south, on a “ghost-flight” without fully active AP and FMC speed control. Perhaps basic Envelope Protection (EP) stabilized the aircraft sufficiently for protracted more-than-less straight and approximately-level flight. However, the aircraft may have “porpoised” up-and-down vertically by up to +/- 3000 feet, and possibly vascillated latterally by up to a degree. Those fluctuations much exceed those expected for any TRUE / MAGNETIC TRACK / HEADING HOLD configuration of the AP. So, optimizing the MH370 ghost-flight-path to best fit the BFO values, may imply that said ghost-flight occurred without the benefit of any engaged AP or FMC speed control mode.

    However, the Inmarsat flight-path, as well as this author’s own slightly adjusted path, all proceed nearly due southwards, despite high velocity westerly winds aloft, above 30,000 feet. That might mean, that MH370 flew southwards below 30,000 feet, so avoiding the high-speed westerlies at high altitude. Note, the afore-mentioned ATSB update acknowledges valid candidate flight-paths exist at altitudes as low as 20-25,000 feet.

    If MH370 actually flew southwards, at a variable speed of (440 +/- 30)kts, at 25-30K feet, then the aircraft’s flight-path qualitatively resembled the “phugoid” and possibly even “S-turn” maneuvers observed in simulators during uncontrolled flight. Note, the combinations of higher speeds at lower altitudes (480kts down at 25K feet), would put the aircraft near the leading edge of its flight-envelope, where dynamic air-pressure (or “ram pressure”) reaches and exceeds engine thrust.

    The ATSB’s “Data Error Optimized” paths cluster near the maximum of excess calculated fuel endurance, where the circle of maximum range from last radar contact reaches, and extends beyond, the 7th ping-ring arc. And, if the aircraft was flying without benefit of fuel-milage-maximizing computerized guidance from APs and FMCs, then its fuel economy would have been lower — especially if the “cabin disintegrated” partially, rendering the fuselage fractured & rough, so increasing drag. The overlap, of MRC fuel range, and “Data Error Optimized” flight-paths from the ATSB update, spans roughly 33-35S. That also agrees with Henrik Rydberg’s as well as Duncan Steel’s & Richard Godfrey’s interpretation of drift analyses, suggesting that crash-sites near 34S minimize the probability of wreckage reaching western Australia, so being the most consistent with (non-)findings there, whilst simultaneously maximizing the probability of wreckage reaching Reunion & Mauritius.

    If so, then perhaps the ATSB priority wide search area would best include locations all the way up to Broken Ridge, perhaps as far as 32-32.5S.

  3. @Richard: I can’t tell you what the additional coincidences are until you propose an actual path.

    @Ge Rijn: I can’t wrap my head around a scenario in which a pilot would reboot the SDU while still well within radar range, and be CONFIDENT this radar would either be off or malfunction. It just doesn’t ring true to me – how could this person be so confident? Why not wait until clear of all installations, just to be sure?

  4. The ATSB’s updated definition of search areas indicates a focus on “Constrained AP dynamics” ghost-flights towards ~38S +/- a degree or so.

    However, as implied by the Inmarsat JoN article, “Data Error optimization” implies crash-sites farther northeast, ~34S +/- a degree or so.

    Different sites display different graphics, and provide differing discussions, of the ATSB priority search area. The exact locations under investigation is extremely ambiguous and the issue is very confusing. Some graphics imply that the ATSB is searching as far north as 32S, others show the same priority search area extending only to 36S.

    The most obvious interpretation of the following graphic, consistent with the 10 December 2015 ATSB update, is that the search area has been refocused towards 36-40S, with the search area widened perpendicular to the 7th arc, so as to increase the search area back up to 120,000 km^2.

    That implies a complete confidence in “Constrained AP dynamics” ghost-flight scenarios. Failure of such a search to locate the aircraft wreckage would be consistent with a confirmation of “Data Error optimization” flight-paths, like that described within the Inmarsat JoN article. Such paths are also more consistent with Henrik Rydberg, Duncan Steel, and Richard Godfrey’s drift analyses, which all also indicate crash-sites near 34S… farther north, and there ought to have been numerous fragments found, farther south fragments found ought to have washed up around W Australia… only sites near 34S simultaneously vanish most of the debris far out in the middle of nowhere, consistent with few finds, whilst also directing those few finds towards Reunion & Mauritius and away from W Australia.

    So although I’m not completely sure where the ATSB is actually searching, I’m under the impression, that they have seized upon “Constrained AP dynamics” scenarios towards 38S. In contrast, minimizing RMS BFO errors generates candidate flight-paths towards 34S, with RMS BFO error (for the ghost-flight phase) reducible under 1Hz if one also allows for fluctuating airspeeds, e.g. “810 +/- 50 kph” in this author’s personal current best-BFO-fit flight path.

    Perhaps everyone could agree, that imposing AP flight guidance is, at best, only marginally compatible with the actual measured data values ? Perhaps all could acknowledge, that more Inmarsat-like flight-paths, are more consistent with the actual satellite data, and also with current drift analyses ? Is it known that MH370 ghost-flew with APs engaged ? Perhaps the FMT itself indicates some sort of “manual intervention & override”, consistent with a landing attempt by the pilot, or even a hypothetical passengers / copilot retake of the flight deck ?? Such scenarios could leave the aircraft in an over-ridden, AT + EP only flight mode, stable enough to remain airborne for hours, but not necessarily a conventional pilot+AP controlled calm cruise at constant altitude & speed ??

    Avoiding more northerly search areas near Broken Ridge may be very justifiable, imagining an AF447-esque debris field strewn over already-jumbled underwater topography could render searches intrinsically unfeasible from the outset.

  5. Wazir Roslan – If the Najib govt had reason to believe the opposition was in any way implicated I reckon they would need to sit on it for their own national interest. Domestically they could damage the opposition but it might be terminal for MAS. And very destructive for Malaysia’s international reputation. I believe they would need to bite their asses and wear it or people would be boycotting MAS and possibly other Muslim airlines?

  6. Ken Goodwin,

    We have touched the flaperon report earlier. It was posted at Duncan’s blog. I think this report did not cause any major effect.

    I agree with your conclusion. In my opinion it is not possible to conclude how the flaperon was torn away based on photographs. The investigators could be lucky enough to detect residual deformations, but presumably they were not.

  7. @Brock McEwen

    Commented your post but it seems lost.
    Maybe it takes some more time this time.
    I’ll wait and see.

  8. Ge Rijn,

    Re motive or motivation.

    No doubt that right now we can only speculate. But it turned out that assumptions with regard to flight mode are not sufficient due to the “black interval” 18:22-19:41. There are many ways to fit data. For instance Bobby’s flight path. It requires S-shaped turns combined with ascent, not detected by 3 radars, all within 15 minutes. Motive/motivation-wise, I can’t imagine how an actively navigated aircraft could become a ghost flight set in TRK/HDG hold mode in as short as 15 minutes.

    You said: “the assumption of a technical failure would be an accident probably without a motive.”

    I guess the right term in this case is “motivation”. Nearly same as motive. Motivation would be to survive in case of a technical failure.

  9. @Brock McEwen

    I’ll try again for it seems lost.
    There was another server timeout it seems.

    The reboot happened at 18:25.
    Last track from Butterworth was from 18:22.
    So the reboot was outside Butterworth radar range (I assume this data is correct):

    I tryed without succes to give an explanation between the ‘stunning synchronicity’ of the reboot and just being out off radar range, from your suggestion in this statement there must be a relation.
    It could also well be there is no relation between the two at all.
    It could as well be pure coincidence.
    Which would rule out this inconsistency from your list.
    As also the ‘inconsistency’ the plane was not tracked by Butterworth radar station.

    About confidence of the pilot being out of radar range. I assume an also militairy trained pilot like Shah would know the range of Butterworth (and other stations).
    And it was the last and only radar station he would worry about not attracting attention and superstition.

    A passenger plane crossing mainly its own country following waypoints and flight paths till far in the Mallaca Straight will not attrackt much if any attention of militairy radar stations in other country’s.

    Another inconsistency however is imo that MH370 was not tracked or seen by the Atjeh radar station with a range of ~500km as far as I know. Or is this explained in the meantime?

  10. @Brock McEwen @Jeff Wise

    I had another try answering Brock’s comment but again it seem lost.
    Is there a problem in recieving Jeff Wise?

  11. @Matty

    At that point in time MAS was slip sliding away anyway into inconsequence, lumbered with huge losses and diminishing load factors.

    The more pertinent point is Shah a jihadi. By all accounts he was far from being one, his lifestyle and that of his family being pretty indicative. Malaysian Muslims save a few are not overzealous in that regard, you can see that to be the case if you visit their urban areas or consort with them in diaspora. They are not hung up over that stuff, maybe in the more conservative insular east coast but not elsewhere. In fact, those rural folk are not very much distinguishable from the bible belters of middle America , fire and brimstone stuff but nothing more.

    Qardhawi is a spiritual mentor for both sides of the divide and despite claims to the contrary by disaffected people, he is a mild benign version of Osama Awlaki and those of that ilk. It’s a simplification to state that Najib and co would not have milked Shah if they knew. After all politics is a game, a survival of the fittest contest where the end justifies the means . Having seen my link who is to say Shah wouldn’t have been milked for all his worth in this case.

    I am inclined to see MH 370 on a wider geopolitical context , one that pits Washington and Beijing . If ever Shah committed suicide, which is I ver seriously doubt given the technical data, it would have been for personal or family reasons not political.

  12. @Matty-Perth
    you know, christians also thought that theirs view is one and only possible religion and in history was spreading it in the world in some cases “quite assertively”; islam isnt the enemy nor muslims, only radical extremists of ALL kind, which you can find in muslim countries, in Israel, in Russia, in USA, here, everywhere – even like small groups with reasonable impact; I personaly add on the list also the “scientology” network as its not in fact something covered by freedom of religion, as it is NOT religion (at least here) but dangerous manipulative pseudoscientific scam, having potential to be behind all the evils today spread by business focused primarily on maximized profit at any human resources and moral costs; truth is, I am becoming radical here too; its easy, its nemesis 🙁

  13. @Brock
    >I can’t tell you what the additional coincidences are until you propose an actual path.

    I was defending the ATSB/DSTG analysis from the charge of inconsistency on the grounds you listed, so you can pick any/all of the paths defined there. From what you have said so far, I would expect you have a ‘second coincidence’ factor for all of them.

    As you will have understood, I doubt the statistical validity of the ‘coincidence’ approach as there is too much freedom to post-facto define a coincidence that obtains, and ignore the possible coincidences that don’t.

  14. @Oleksandr

    I guess that’s the worth of different assumptions isn’t it.
    If one or more fail to fit new data and evidence they proofed their worth by failing those and can be scraped from the list of possible assumptions.

    One of them could be my ‘fuel dump’ assumption. You are proofed to be right with the mininum remaining fuel after jettison.
    It’s actualy dubbel your amount of 5200kg.
    5200kg in each main tank. The center tank can be jettisoned empty.
    Still there seems to be one or two possibilities left to empthy those main tanks. There is a manual override knob to increase or decrease jettisoning fuel.
    The other maybe could be pumping the remaining fuel of the main tanks into an empty centertank and then jettison the main tank empty. A link:

    Anyway in an all controlled flight the pilot would have control over the amount of fuel to use. If jettisoning would be required in any stage of the flight he could have used it imo.

  15. @Rand

    ‘But why to fly to a point of fuel exhaustion or near exhaustion’?

    I explained before if you assume an all controlled flight with a glide and ditching I assume the pilot would have a calculated specific end point to reach.
    To perform a safest possible ditching with a minimum risk of explosion and fuel traces I assume he wants to get rid of as much fuel as possible by jettisoning or exhaustion by burning or a combination of both.
    Anyway he would have control.

  16. Richard,

    There are several problems with the current search area:

    – Nothing was found there so far;
    – It is anecdotally clear that underlying assumptions, which lead to that area, are inconsistent either with each other, or with some data/observations, or their combination does not make sense;
    – There are quite several ‘symptoms’ that the more likely crash area is between 25 to 30S.

    The above points do not invalidate the current search area. However, they rise a question about effectiveness of ATSB’s search strategy, especially bearing in mind shameful budget constraints.

  17. “Co pilot locked out SDU/satcom shut off, depressurytion of the cabin after IGARI would silence them all within half an hour.”

    This is unlikely. The 15 portable oxygen bottles would have lasted for 2 hours 35 minutes each (twice as long on a one liter setting).

    If isolating the left AC bus means unlocking the cockpit door, it sounds like a rather unrealistic endeavour.

    I think this scenario gained traction when some news outlets reported misleading altitude data, suggesting a steep climb. It has since been established that MH370 descended after the diversion, which is further inconsistent with deliberate depressurisation.

  18. @Rob

    I bet it’s my finger pushed some wrong button..
    Or Jeff Wise is censoring only my comment to Brock McEwen for a still unidentified reason..;-)

  19. Ge Rijn,

    “If jettisoning would be required in any stage of the flight he could have used it imo.”

    If the left bus was powered… I already mentioned that fuel dumping remained a mystery for a while, but this problem appears to be resolved now. What the crew could hypothetically do in such a situation is either burn fuel, or attempt to restore power to the left bus. Perhaps they partially succeeded by 18:22. But with left bus powered, something went wrong; perhaps short circuit or so.

  20. Gysbreght,

    The emergency descent protocol for B737 requires entering MCP altitude of 10,000 ft and SPD 0.82 M / 340 knot (ref smartcockpit). Do you know if the same procedure is applicable to B777?

  21. @ Jeff-@ Susie

    Regarder sous l’eau .Vous devriez voir le contour d’un avion . Déplacer ensuite l’image
    vers le centre de votre écran et observer la
    traînée dessous ainsi qu’a la droite les remous du a l’empennage.
    Le cockpit ce trouve a droite l’empennage a gauche l’avion est légèrement incline vers
    la droite .L’avion est sur le cotée droit.

    OBSERVER BIEN L’IMAGES en bas et droite

    Cette MAP ou image ce situe au nord ouest
    de {Banda Aceh} SUMATRA.

  22. @erik nelson

    In the map referenced below I show my estimates of the search status, based on following the search for some time now. The map includes the following:

    1. The track of the three ships on their current ‘swings’ from Fremantle. Only Equator is currently working, bathymetric scanning inside the 7th arc, just turned south.

    2. The 7th arc +/-40nm search area shown in the DSTG and ATSB reports from December 2015. This excluded the ‘Example Reconstructed Flight Path’ in the Inmarsat JoN paper on the stated basis that the errors in the BFO data did not justify its inclusion, against the Constrained AP model.

    3. My estimate of the possible search area that is currently being worked. The northern boundary has been pulled in by 100km, but the width that will finally be covered is unclear. I have drawn it to cover the total area that is being bathyscanned by Equator (85nm wide to date) but this leaves more area to be tow-fish-scanned that has been stated, so probably won’t all be covered.

    4. A number of points of interest, including the most northern point (on the 7th arc) that has been tow-fish scanned (in this case by Go Phoenix in 2015). Clearly, this point is now well outside the current area of interest.

    On the topic of the possible paths based on the Inmarsat data, the BFO residual error analysis in the DSTG report make use of BFO data for detailed path generation very doubtful.

  23. @Oleksandr:
    I haven’t researched the FCOM but expect the 777 procedure to be similar, except for the speeds which would take the 777 Vmo/Mmo into account.

  24. Having replicated the 2014 Inmarsat JoN model, and having exercised the same, to explore the “parameter space” around their primary candidate flight-path from IGOGU at 18:40 to 34.7S on the 7th ping-ring, I observe, that all similar best-fit trajectories “train together”, such that moving the final crash-site up to 34S (say) also moves all intermediate, as well as the initial 18:40, locations equally far northwards, up to 8.2N (say). Frustratingly, moving the final crash-site up towards 34S “fits the BFO data like an astronaught glove”, dropping the RMS BFO errors down to <1Hz… at the cost of sliding the 18:40 location the better part of 100km north-and-east of IGOGU, well out of the KLIA FIR and Malaysian airspace, and so also requring an almost 600kts airspeed to reach the FMT from an 18:25 location near NILAM.

    I suspect, that Inmarsat noticed all these mathematical effects, two years ago. I suspect, that they would have "wanted" to slide their suggested flight-path trajectory northwards, but noticed that they could not input a 7th-arc crossing location north of 34.7S, without "running the FMT into the wall of the KLIA FIR at IGOGU", at 7.5N. Again, if you input a 7th-arc crossing point at 34S, then the best-fit FMT slides up to 8.2N, implausibly far from NILAM and far off of the military radar track.

    However, there are, of course, a wide range of 18:40 locations, coupled with various air-speeds and (invariably negative) RoC's, which all satisfy the 18:40 measured BFO value. The actual inhibiting factor, is the presumption, that the 18:40 BFO must be assigned, to the same ghost-flight trajectory as all of the ensuing, subsequent, following, BFO values. It is very easy to understand Inmarsat's reasoning on this point, because the 18:40 BFO essentially requires a southbound velocity vector, i.e. MH370 basically "must" have already completed the FMT more-than-less onto its fateful southwards ~180 heading. So its very natural to tie that data point onto the remaining pilot-nonresponsive trajectory.

    Never-the-less, excising the 18:40 data point, from the ghost-flight phase — and re-assigning it, implicitly, to the pilot-controlled flight phase from IGARI to somewhere near IGOGU — "frees up" the ghost-flight trajectory to "slide northwards" to 34S and even perhaps beyond. RMS BFO errors drop as the candidate trajectories appear to "fit into place".

    Perhaps MH370 only had one proverbial foot in its watery grave at 18:40, with the other pilot's foot still actuating the rudder pedals, for some short while afterwards ? If so, then a panoply of numerous implications all align, suggesting a similar scenario. Firstly, pushing the 7th-arc crossing point northwards to 33-34S implies slower, and hence very arguably lower, flight paths. The ATSB update labels such paths with altitudes of 20-25K feet. Nextly, situating the FMT entirely within the previous prior pilot-controlled flight phase, from IGARI towards IGOGU, implies that the FMT southwards was human desired & executed, notably with the coupled deceleration (from 500-510kts to 420-440kts) and descent (from 35K feet to 20-25K feet). Thusly, a "half-beating-hearted" landing attempt, plausibly at an Indonesian airfield very near Banda Aceh, appears to be implied. Intriguingly, such is completely consistent with un-denied allegations of a final faint radio transmission from the aircraft at 18:43 requesting an immediate emergency landing clearance due to the "cabin disintegrating". Reassuringly, an emergency landing attempt around 18:40-45 would imply a negative RoC as marked as -5000fpm, which permits fitting the measured actual 18:40 BFO value, of 88Hz, with these slower speeds and lower altitudes. Finally, pilot non-direct-response to the 18:40 sat-phone call could easily be attributed, to a dire emergency, with frustrated and over-taxed pilots not being immediately willing-and-able to notice and/or instantaneously respond, via effectively functional technology, to one more klaxxon blaring from the flight-deck speakers.

    Moreover, the drift analyses of Duncan Steel, and everyone else he credits profusely, appear by-and-large powerfully persuasive, namely that a crash-site near 34S on the 7th arc, simultaneously vanishes most of the debris, being far from any land, whilst directing a relative few items both away from Australia and towards Reunion & Mauritius. Once again, a 7th-arc crossing of 34S is also actually more consistent with satellite BFO data, even than the primary Inmarsat flight-path towards 34.7S, on all points after 18:40, which point can be separately satisfied, with a plausible combination of reduced airspeeds, southwards flight towards Banda Aceh, and a high but justifiable rate of descent, in an ultimately futile and "aborted" landing attempt. Extremely speculatively, but not impossibly, a successfully initiated, but ultimately failed, landing attempt, could have been undertaken by persons other than the official pilot & co-pilot, perhaps due to their incapacitation, due to hostile environmental conditions aboard the plane, and/or hypothetical violent conflict.

    Additionally, such scenarios might plausibly render the aircraft into a configuration, without engaged APs and FMC-controlled speed modes. Possibly, the a/c veered away from Banda Aceh, onto its ghost-flight trajectory, about 18:43, due to a "cabin disintegrating" scenario not utterly incompatible with violent conflict scenarios, and proceeded onwards, in a semi-stable flight mode, including vertical "phugoid" oscillations with accompanying airspeed fluctuations, and possibly even mild lateral "S-turns", all observed in simulators during uncontrolled simulated flight. A crash-site near the 7th arc somewhere between 32S-35S would fit drift analyses, the satellite data, as well as fuel consumption models, especially given possible increased drag from a partially compromised fuselage, which would imply higher fuel burn rates, lower fuel economy, and points near the maximum computed MRC range near 33-34S.

    Admitting the possibility of pilot inputs for a few minutes after the FMT and 18:40 — consistent with a plausible initiated but failed landing attempt at or near Banda Aceh — appears to permit markedly improving fits to the satellite data.

    Without desiring to get side-tracked or mired in arguments over issues not directly affecting the final crash-site location, the "suddenness" of the change from LNAV to ghost-flight in 15-20min from MEKAR & NILAM to FMT vaguely near IGOGU, could imply a violent conflict scenario… e.g. user FabioF's suggested "retake" of the FD and a hypothetical hijacker detonation of some explosive device.

  25. Gysbreght,

    FCOM was the first thing I checked, but I haven’t found anything similar to B737 on the emergency descent procedure. Numbers are interesting, as well as the input on MCP itself (first officer’s responsibility).

  26. @Oleksandr

    The original point from Brock was internal consistency of (all) models, so you are widening the question.

    >Nothing was found there so far

    Well, that’s obvious. The interesting point would have been (if the search were to be continued) what to do next: widen the search, move up (or down) the 7th arc, ignore the arcs entirely? That’s what most of the discussion has been on this blog. Since a simple straight-line, constant speed course gives a good fit to the BTO data I find it hard to believe that a solution that did _not_ include that would have been chosen first by almost any selection process. A hard-core motive-based supporter (not me) might have argued for a more northern route, but would have had little chance of success against a numbers based theory such as the Constrained AP model.

    > It is anecdotally clear that underlying assumptions, which
    > lead to that area, are inconsistent either with each other,
    > or with some data/observations, or their combination does not make sense

    That’s the same point that Brock made and which I questioned. He may be developing a concise summary. At the moment it is not clear to me, at least.

    >There are quite several ‘symptoms’ that the more likely crash area is between 25 to 30S.

    If you mean drift analysis models, their uncertainties make it difficult to use quantitatively against the validated Constrained AP model. Of course they have only been at all relevant since mid-2015. I briefly addressed that in my original response to Brock.

    Of course questions arise over the effectiveness of ATSB’s search strategy, since it has not been successful. I have not seen any coherent proposal that would have led to a definable search area of reasonable total size and could be validated in a quantitative way (as was the constrained AP model). It is 3500km from the north end of the current search area to the Indonesian coast. The maximum length of any searchable area of reasonable size is around 1000km (500km if the search has to reach the forward glide limit), so defendable constraints have to follow from any model.

    That may be an unfair criticism since ATSB have deployed many staff-years of effort in developing their data analysis and modelling, clearly beyond any non-state actor.

    >…shameful budget constraints

    I am glad I don’t have to worry about such a budget/moral/legal/safety/political (not in order) trade-off.

  27. @Oleksandr:

    Non-Normal Checklists, NNC 2.5, in the Lauda Air FCOM reads as follows:

    Condition: Cabin altitude is excessive.
    # OXYGEN MASKS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . ON B
    # CREW COMMUNICATIONS. . . . . . . . . . . . . ESTABLISH B
    # CABIN ALTITUDE AND RATE. . . . . . . . . . . . . . . CHECK B
    [Confirms pressurization problem.]
    # If cabin altitude uncontrollable:
    # PASSENGER OXYGEN SWITCH . . . . . . . . . . . PUSH F/O
    Push and hold for 1 second.
    [Backs up automatic activation of the passenger oxygen system.]
    # DESCENT . . . . . . . . . . . . . . . . . . . . . . ACCOMPLISH C
    Without delay, close thrust levers, extend
    speedbrakes, and descend at VMO/MMO. Level
    off at lowest safe altitude or 10,000 feet, whichever
    is higher. If structural integrity is in doubt, limit
    airspeed and avoid high maneuvering loads.

  28. @Erik Nelson: Although Ashton et al.’s JON paper makes some over-simplifications, the approach of fitting a flight segment starting at 19:41 and not trying to determine the exact flight path between 18:22 and 19:41 has merit because we have no evidence that there was only a single turn to the south.

    Almost two years ago I studied a path that continued due south after 19:41 and crosses the 7th arc at 34.3S. The path requires a loiter pattern north of Sumatra, which may correspond to a holding pattern, an aborted approach to Banda Aceh, or another diversion. I remain intrigued by this path for a number of reasons, including:

    1. A final waypoint of SouthPole (due south track) can be easily entered by a pilot, and allows a continuous LNAV (Great Circle) path until fuel exhaustion.
    2. The Great Circle (geodesic) is a loxodrome (rhumb line) for due south.
    3. Since the minimum BTO occurs at a time (near 19:41) when the plane is traveling due south, the latitude of the plane at that point should be very near the latitude of the satellite (1.65N) at that point to ensure tangency to the ping arc. Therefore the latitude of the plane at this time can be obtained from inspection.
    4. Once the latitude (1.65N) at 19:41 is known, the longitude (93.7E) at this time can be determined time by the ping arc constraint.
    5. The longitude (93.7E) at which the plane crosses the 7th arc is the same as the longitude at 19:41.
    6. A constant Mach number path matches the BTO and BFO data for 19:41 to 00:11, inclusive.
    7. The longitude of the path (93.7E) aligns well with waypoint BEDAX, which is a Standard Instrument Departure (SID) waypoint for Banda Aceh. As such, the final two waypoints could have been BEDAX-SouthPole.

    The crossing of the 7th arc for this path is near 34.3S,93.7E. Unfortunately, that area was searched by GO Phoenix with no positive result, although the area surrounding this point was not searched as thoroughly as end points further south along the arc. Based on the work of Richard Cole, it looks as though this area will not be further searched.

  29. @Richard: okay, in that case, I pick the latest SW extension, to 85°E. This impact location requires…

    1) an FMT so swift, MH370 had to pass over the tip of Sumatra – a flight path its general public pays vast sums to be able to detect. Undetected. (If it was detected, but this info was shared only with search leadership, then search leadership never would have searched multiple time zones to the NE.)

    2) a spectacularly unfortunate coincidence, by which the Fugro ships were turned around just shy of this point, ostensibly because of a material miscalculation in what should have been a conservatively computed range limit.

    3) a spectacularly unfortunate coincidence, by which towfish equipment – in which search-funding nations invested enormous sums in exchange for equipment that would not miss wreckage – missed wreckage.

    Non-detection by

    – Curtin’s acoustic equipment (made all the more curious by the size of found debris, which implying high-energy impact),
    – contrail analysis, and
    – air and satellite surface debris search which focused on precisely this area March 17-27 (the most intensive campaign of the search)

    …but you only asked for one, so I’ll stop at the enumerated three.

    (Of course, none of these are coincidences if one drops the now thoroughly discredited assumption that the search has been conducted in good faith.)

  30. @Richard

    Yes, I have heard all the arguments about the “sensibility” of a constrained AP mode many times from many people. What those arguments indicate to me is a deep philosophical divide in problem solving methodology. My experience (a lot over some four decades of engineering) is that if you cannot get an “answer” that is within 20% or so with a simple white board discussion, it makes no sense to start Matlab or whatever your tool of choice might be. These days it seems that people belly up to a keyboard rather quickly if they have a set of numbers to type in. I think this approach is wrong, and often starts you on a path that is difficult to abandon.

    Simply asking what might have happened before firing up the analytical tools would have taken the current search area off the table very early. It simply makes no sense relative to ANY plausible causality.

    The DSTG report was the final shoe to drop in the “let’s solve this problem at the keyboard” philosophy. The fact that they were able to model previous flights relatively accurately should not be surprising. Their model, by their own admission, was biased to normal flight dynamics. It would never have been able to model the turn at IGARI or any of anomalous behavior prior to the FMT. Their high probability area is basically no different from what had been derived previously by the IG and others using a “common sense” approach without the Bayesian smoke screen.

    I am not suggesting that anyone did anything wrong. It is more a matter of how different people prioritize how they think about a problem. As more debris is found it will become increasingly difficult to cling to what I would call the “analytical mainstream”.

  31. @VictorI

    Re the path travelling due south from the 19:41 crossing.

    The speed found to be a constant Mach number from 19:41 to 00:11, once you had adjusted for the winds rn route?

    It might sound like a silly question, but I believe everyone should be allowed a silly question occasionally.

  32. @Rob. Yes. Path reconstructions include corrections for temperatures and winds at altitude.

  33. So how strong is that cockpit door, anyway?

    There still seems to be some argument as to whether the passengers were incapacitated.

    A similar door held up for the few minutes the Germain Air pilot had to breach it. Could the cockpit door hold up to an hours long assault by multiple passengers using everything from fire extinguishers to pieces of galley carts to possibly an axe?

    If so, that is one VERY strong door. Or the passengers were not in the fighting mood, so to speak.

  34. @Mikeroo

    An interesting and pertinent question. On that has not been properly addressed, since the plane disappeared.

    Depressurization would have been a chancy thing for, in my opinion, as he would need to survive the extreme cold, even if he did have adequate oxygen.

    Could it turn out that the plane was never depressurized? I think it’s possible it never was.

  35. @Rob

    This is actually one of the biggest mysteries to me, and it seems very key to any explanation.

    If I was a passenger, I could be easily fooled that we had to remain in a holding pattern for an extra time…the difference being about 2 hours (5 hours, 34 minutes vs.
    7 hours, 31 minutes). I would be very annoyed but would probably sleep through it. I would never know if we were on course or not…I’m a passenger in a tin can. Some more knowledgeable passengers might catch on with phone GPS, but might contribute the discrepancy to errors caused by the plane’s structure and brush if off. Looking outside might or might not be useful depending on light conditions and cloud cover.

    But the pilot or co-pilot, if only one was involved in the scheme, would never fall for that — of course.

    You are on point noting that depressurizing the cabin would be very uncomfortable in the cockpit as well.

    So, was everyone overcome by fumes? If slowly and they were aware, then a lot of phones are sitting at the bottom of the ocean with farewell videos…some may be retrievable as most are data is stored in non-volatile IC chips.

    If they were overcome without actually being cognizant of the process, then there may still be videos but fewer and of a different type of content.

    That plane needs to be found.

  36. Richard,

    Re “The interesting point would have been (if the search were to be continued) what to do next: widen the search, move up (or down) the 7th arc, ignore the arcs entirely?”

    That is a thing. The nominal 7th arc can differ +-20 km from the actual arc; on top of it add uncertainty in the altitude and distance due to spiral dive. Everything further than such a concervative estimate requires controlled gliding. If the flight was controlled, it could be any location at the 7th arc, so 38S does not make any sense in such a case. So what can justify widening?

    Re: “I find it hard to believe that a solution that did _not_ include that would have been chosen first by almost any selection process.”

    I agree. But it is time to move on. It makes no sense to me to scan low-probability outskirts of the priority area, while there are more promising areas.

    Re: “A hard-core motive-based supporter (not me) might have argued for a more northern route”.

    Obviously I am not a hard-core motive-based supporter, but rather opposite. However, I am a hard-core supporter of “common sense”. I cannot believe that someone was actively navigating at 18:25 and performing non-trivial tasks (SDU reboot), but then suddenly everyone was incapacitated in just 15 minutes; AP was set in HDG or TRK HOLD mode into the SIO; no attempts to descent; no attempts to communicate between 18:25 (after SDU came back) and 18:41.

    Re: “That’s the same point that Brock made and which I questioned. ”

    No. He questions data, but believes in assumptions. I question assumptions, but believe in data.

    Re: “If you mean drift analysis models, their uncertainties make it difficult to use quantitatively against the validated Constrained AP model.”

    Not only. There are several more ‘symptoms’. I would certainly assign the highest priority to the Curtin boom location (~28.5S). It is well defined small area, the most promising in my opinion.

    Re shameful budget constraints. That is because the search budget did not even exceed the cost of a B777 aircraft, but it has potential to partially recover expenses. Insurance companies, for example, cover the cost of the lost aircraft without a problem. Why 3 involved states and Boeing cannot provide adequate budget, especially bearing in mind initial commitment of some of them to find the aircraft? Obviously this is not ATSB’s problem.

  37. @ROB:

    I’m not sure that depressurization necessarily equates to “extreme cold”. The cabin is well insulated, so it doesn’t cool down rapidly, as long as you don’t let the cold outside air in. Cabin pressure and temperature are separately controlled. When the cabin is de-pressurized by opening the outflow valve(s), the airconditioning system can still provide heated air IMO. The cockpit has separate heaters for the pilot’s shoulders and feet.

  38. @Gysbreght

    Agreed. The pilot-in-command only has to survive for a short time…comfort is not required.

    Depressurization is still a strong theory as to how a couple of hundred passengers were kept at bay.

  39. 180 million dollars. Think about that number!

    That’s roughly 90 million gallons of diesel. Or 3000 people paid 60K each for a year.

    Or some combination thereof. I wonder if they will perform an extensive audit when this is all over.

  40. Gysbreght,

    Correction to my post 10:46 am: MCP is Captain’s responsibility in case of emergency descent.

  41. @Gysbreght

    I’m sure that cannot be the case, with all due respect. Not wishing to be argumentative, for the sake of it.
    I understood that to depressurize the cabin, you would need to both open the outflow valves, and close off the aircon pack valves that supply warmed air into the cabin. And in thus situation, there would be no warm air available in the cockpit.

    The Helios accident is often cited as an example of what happens when the pressurization fails. In actual fact, the cabin was never fully depressurized in this incident. The cabin altitude dropped to about 20,000ft at one stage.

    But if it turns out I’m wrong on this issue, I will gracefully concede.

  42. @Mikeroo

    The plane needs to be found. On that I agree wholeheartedly!

    And there is question, touched on here a couple of weeks ago, concerning the passengers’ mobile phone messages, or more specifically, the almost total absence thereof.

    Apparently, with one possible exception, no mobile phone messages went out from the plane after boarding. This if true, was extremely unusual.

    One possible explanation; Could the Captain have forbidden passengers from using their phones, on some false pretense that the plane’s electronic equipment would be at risk. As the flight progressed, there was indeed a fictitious fault with the electronics, possibly a string of them. passengers were fed one story after another about why the plane couldn’t land, until the inevitable happened.
    This way, the pilot could have avoided resorting to a risky depressurization.

    One problem with this scenario is summed up by the question; but what about the co-pilot? Why couldn’t he intervene?

  43. @DennisW
    I certainly cannot contradict your experience – if that is what you found, fair enough.

    On a point of fact, the DSTG model does find its way around the FMT: its input start point is the last radar return, well before the FMT. So it is capable of dealing with non-standard flight behaviour when the data (in this case the BFO data for the attempted phone-call) demands it. The point of the DSTG model is that it constrains the length of the arc that is consistent with the BTO data (and hence gives a search area). Best fit models with no error analysis (such as the early IG models) didn’t give a quantitative estimate of what the search area should be.

  44. @ROB: “The Helios accident is often cited as an example of what happens when the pressurization fails. In actual fact, the cabin was never fully depressurized in this incident. The cabin altitude dropped to about 20,000ft at one stage.”

    Do you have a reference for that statement? I’m asking because i couldn’t locate the accident report, only an article in the press:

    Learmount, David. “Investigation dispels myths around Helios Airways crash.” Flight International. 17 October 2006:

    “… cabin altitude increased almost in line with the 737-300’s height, as the aircraft climbed from Larnaca, Cyprus.

    This is clear from the FDR download from the 14 August 2005 accident carried out by the French investigation agency BEA for the Hellenic Air Accident Investigation and Aviation Safety Board (AAIASB)’s just-published final report on the crash near Grammatikos, Greece.

    Similarly, when the aircraft ran out of fuel and began its descent after being flown over almost its entire route toward Athens by the pre-programmed flight management system/autopilot, the FDR traces of cabin and aircraft altitude shown in the report also coincided.”

  45. Without the ISAT data of BTO and BFO we wouldn’t know in which direction or distance from last known position to look for the wreckage. The BTO and BFO together allow an accurate path reconstruction, provide one knows or assumes a point in time and space from which to start.

    The BTO by itself doesn’t allow that, unless one specifies a starting point and a speed that is maintained forever after that point. It so happens that normal cruising speed provides a path that can be argued to be compatible with the HDG/TRK modes of an un-attended autopilot, i.e. with a passive human pilot. That path leads to the DSTG ‘hot spot’ and similar end points proposed by the IG and others.

    The “Constrained AP” assumption largely disregards the BFO data, on the ground that those are subject to greater errors of mesurement than the BTO data. Based on analysis of data from 20 flights prior to MH370 with known trajectory (but without providing that analysis), the DSTG concludes to a standard deviation of 25 Hz of the BFO bias error, plus a ‘noise’ standard deviation of 7 Hz (“to be conservative”). Therefore, any difference between calculated and recorded BFO can safely be dismissed as an error of measurement.

    The ‘priority search area’ is based on the assumption of “Constrained AP” in combination with a passive (“unresponsive”) crew. The failure to find anything in that area is a strong indication that these assumptions are wrong. So perhaps it is time to give more weight to the BFO data than has been done hitherto? After all, the estimated error distributions are symmetrical about the zero mean value, so the maximum likelyhood value of the BFO values is the one that has been recorded.

Comments are closed.