New MH370 Debris Suggests a High Speed Impact

 

Blaine Gibson holds new MH370 debris, accompanied by a fisherman who found the part in Madagascar

Today Blaine Gibson and other MH370 family members delivered to Malaysian authorities five new pieces of debris that are believed to be from the missing aircraft. The parts were found as a result of a campaign led by Blaine and the families to make the residents of Madagascar more aware of debris from MH370 that has drifted across the Indian Ocean and continues to land on the shores of East Africa.

Blaine has provided photographs and descriptions of the five parts, which are made available here in a zipped file. More recent high-resolution photographs are available here. The most recently discovered part (Part 5), pictured above, was found by a local fisherman this past August.

One piece of debris (Part 3) was identified by MH370 Independent Group (IG) member Don Thompson as a shattered piece of the interior floorboard of a Boeing 777. The part’s location in the B777 and the nature of the damage is consistent with a high speed impact, and therefore has probative value.

Fragment of floorboard of MH370

Identification of the Floorboard Piece

IG members Don Thompson and Mike Exner assisted Blaine in identifying the floorboard piece, and Don documented his findings in a report. An important clue was the piece contained a portion of a placard with the identifying characters WPPS61. Don was able to determine that the full placard number is BAC27WPPS61. This type of placard is affixed to high strength panels of material specification BMS4-20, which is used as flooring  material in passenger compartments of commercial aircraft, including the Boeing 777-200ER. Amazingly, Don was able to find a similar placard affixed to the floorboard of wreckage from MH17, which was also a Boeing 777-200ER. This leaves little doubt that the piece recovered from Madagascar is from MH370.

Portion of floorboard placard showing characters WPPS61

Discussion

The new debris gives us additional insight about where and how the aircraft impacted the sea. In light of the past efforts to find the aircraft, there are three main possibilities that remain:

  1. The aircraft impacted the ocean relatively close to the 7th arc, but at a latitude further north than the area previously searched.
  2. The aircraft impacted the ocean at a latitude previously searched, but farther from the 7th arc than previously assumed.
  3. The aircraft debris field was in the subsea area previously scanned by sonar, but was either missed or misidentified.

Sources close to the previous search effort believe (3) is very unlikely, as there was a thorough review of the sonar data by multiple parties with high levels of experience, and because any “points of interest” were scanned multiple times to ensure the resolution was adequate to make a determination with a high level of confidence.

When considering the satellite data, the final two BFO values at 00:19 UTC are consistent with an aircraft at an increasingly high rate of descent. The new debris and some of the previously recovered debris also suggest that the aircraft impacted the ocean at high speed. That means that (2) is possible only if the aircraft first was in a rapid descent (producing the final BFO values), and then the pilot skillfully recovered from the rapid descent and glided some distance away from the 7th arc beyond the width of the subsea search, and then later the aircraft again descended at high speed and impacted the sea (producing the shattered debris). This sequence of dive-glide-dive is considered by many to be a very unlikely sequence of events, although it cannot be completed dismissed.

What is left is possibility (1). This suggests future subsea search efforts should proceed along the 7th arc, starting where the last search ended near 25S latitude, and continuing farther north. (In a previous blog post, I showed that an automated flight ending along the 7th arc at 22S latitude is possible.)

The part recovered in Madagascar in August 2018 was the latest in a series of finds that began with the discovery of the flaperon on Reunion Island in July 2015. Because of the wide range of discovery times, and because there is an undefinable delay between when a part arrives on a beach and when it is discovered, it is difficult to use the timing and location of debris discoveries to precisely pinpoint where to search for MH370.

Finally, the new debris finds illustrate the critical role of independent investigators in the search for MH370, and one investigator in particular. The local communications campaign to educate residents of Madagascar about debris washing ashore was spearheaded by Blaine Gibson with the help of some of the MH370 families. Blaine has also done a commendable job of developing a local network to help recover the debris after discovery. We have to wonder if additional debris is sitting on the shores of other countries like Tanzania, Kenya, and Somalia, where there were no similar campaigns to alert and organize the residents.

Links in this article:

Photographs and description of debris provided by Blaine Gibson

Additional high resolution photographs of debris

Identification of the piece of floorboard by Don Thompson

Reconstructed path ending near 22S latitude by Victor Iannello

473 Responses to “New MH370 Debris Suggests a High Speed Impact”

  1. airlandseaman says:

    Excellent report Victor. The floor board (#3) is aonother interior piece that further confirms that MH370 is close to the 7th arc. The fact that debris continues to show up 4.5 years after the event demonstrates the limitations of drift analysis based on arrival timing.

  2. MH says:

    Could have this BAC27WPPS61 flooring materials be also used in Boeing 767 such as the one that crashed near the Comoro Islands

  3. Peter Norton says:

    @Andrew: as always, thank you
    @DennisW: I appreciate you are not mincing words. I would at least have expected Boeing to be a lot more forthcoming about the very system that lead to the crash. They could answer many questions to clarify the system design, regardless of this specific crash investigation. Instead pilots seem to be kept in the dark, if this is true:

    Pilots say they were ‘in the dark’
    « Previous iterations of 737s would have switched off key automatic control features when the pilot first pulled back the control column, a standard manual override feature in generations of airplanes. Investigators found that the final yank on the control column of Flight 610 registered almost 100 pounds of pressure, suggesting desperation in the cockpit as the plane plummeted.

    Pilots expressed concern about the changing nature of the controls and Boeing’s delayed disclosure of the change. The company sent out its public alert more than a week after the Lion Air crash. Tajer and others said they were aware of no earlier notice of the change in how 737 MAX planes operate compared with their predecessors. »

  4. Peter Norton says:

    good questions the Aviation Herald submitted to the FAA:

    • Why was the MCAS permitted to operate on the base of a single AoA value showing too high angle of attacks? Why does the MCAS not consider the other AoA value?

    • Was the risk assessed that one of the AoA sensors could be damaged by a bird strike, hail strike or similiar and could show a substantially too high angle of attack?

    • Considering the scenario that happened to Airbus twice (the crash in Perignan and the Lufthansa A321 near Bilbao losing 4000 feet), that at least two AoA sensors froze in same positions during climb, was the risk of such a scenario on the 737s assessed, too?

    • Was the risk assessed according to Boeing’s last sentence in the notice to operators: “If the original elevated AOA condition persists, the MCAS function commands another incremental stabilizer nose down command according to current aircraft Mach number at actuation.”, in particular what possibilities existed for that conditions to persist?

    • Did the certification deem not necessary that an “AoA Disagree” message was to be introduced?

    • What should the system response have been in case the AoA values disagree? How would the systems determine which value is plausible and which is erroneous? Is there any such check at all? Would MCAS not need to be prohibited if left and right AoA disagree?

    • Did the certification consider a massive change in the function of the AoA when MCAS (as actor in the flight controls) was introduced in addition to stick shaker (monitoring only) ?

    • What is the reasoning behind the certification permitting to allow a system modify the aircraft’s equilibrium (via trim) in manual flight in a way that the trim could run to the mechanical stop and thus overpower the elevator?

    • Was the AoA input to the MCAS (or in general) ever being cross checked, e.g. by taking into account altitude, IAS, vertical speed to compute TAS via altitude, density and IAS and the angle of the airflow by computing the angle of the flight trajectory with TAS and vertical speed? Could such a crosschecking algorithm not even detect if two or more AoA sensors were frozen/faulty?

    • Is the FAA going to review the certification of the 737 MAX family (and perhaps previous 737 versions) following the findings by the KNKT so far?

    • How [does] the certification deal with spurious faults and spurious functions, in particular during maintenance? The maintenance manuals define a test to be run, then list maintenance steps one by one, the test is to be repeated after each step. If the system is found to be working during the test the maintenance task aborts with the message “You have solved the issue”, which may trigger a wrong analysis and premature end of troubleshooting without removing the fault if the test apparently works correctly by random chance.

    • Why do the FIM procedures for airspeed disagree, altitude disagree, feel difference light, inexplicable stick shaker activation etc. not reference the possibility of an AoA issue although AoA has a crucial influence onto all these error conditions, thus not guiding the AME to verify proper action of this input in each of these error conditions?

  5. Richard Godfrey says:

    @Victor

    Many thanks to you and Don for this excellent report. Blaine and the NOK deserve all recognition for their continuing efforts.

    You conclude that “What is left is possibility (1). This suggests future subsea search efforts should proceed along the 7th arc, starting where the last search ended near 25S latitude, and continuing farther north.”

    As a result of @DrB excellent review of the drift analyses, I posted a comment on November 22, 2018 at 4:57 am stating “I think that is fairly conclusive, that the start latitude range between 20°S and 25°S should be searched to at least a search width of ±22 NM.” I support your conclusion wholeheartedly.

    A simple graphical summary of the satellite data, fuel data, debris drift data and the areas already searched along the 7th Arc, are shown in the link below. The remaining area is from 25°S to 20°S, with a priority from 25°S to 23°S:

    https://www.dropbox.com/s/nxcel9mechchx5h/Clear%20Message.png?dl=0

  6. Peter Norton says:

    re: “Amazingly, Don was able to find a similar placard affixed to the floorboard of wreckage from MH17.”

    @Victor / Don: Could you post this photograph ?

  7. Don Thompson says:

    @Peter Norton askedCould you post this photograph ?

    A URL for the image depicting a floor panel bearing the BAC27WPPS61 placard is provided within my report.

  8. TBill says:

    @DennisW
    “To me it (Lion Air Incident) shows a weak kneed FAA and NTSB response, and a good reason not to get on board an aircraft (any aircraft).”

    I am not sure about including NTSB, but your weak FAA criticism is the same way I feel about MH370, re:FAA letting aircraft transponder being clandestinely turned off by rogue pilots during flight without any remedial measure such as ACARS reporting of such event. Ditto of course for ACARS.

    Although we had a scary U.S. flight about a year ago in a possibly non-airworthy Airbus 320, that the pilot exchanged for a new Airbus after an unplanned stop in Houston, the general defense of “getting on board any aircraft” is the mega-safety of air travel compared to auto death rates. I am not sure that comparison (to autos) is the correct way to look at it, but that’s part of the problem getting public/Congress support for air safety improvements.

  9. Don Thompson says:

    @MH asked “Could have this BAC27WPPS61 flooring materials be also used in Boeing 767 such as the one that crashed near the Comoro Islands

    First, a drift path from ET-AIZ’s crash site in the Comoros Islands to the south-east coast of Madagascar is very unlikely. The prevailing ocean current arriving on the east coast of Madagascar is the Southern Equatorial Current (SEC) flowing westward across the Indian Ocean. An object drifting from the Comoros would be ‘fighting against’ the SEC, at either the north or south of Madagascar, to reach the east coast.

    ET-AIZ crashed 22 years ago, it rolled out of Everett in Sep 1987, carbon-nomex composite floor panels were introduced, progressively, during the ’80s. The Boeing Material Specification, BMS4-20, is now applicable to 767s. I’d expect the condition of a piece of panel, due to weathering after 22 years, to be much worse than evident on this piece.

    With the debris fragment to hand, the Malaysian team can pursue some additional analysis. Two variants of the BMS4-20 specification exist, differentiated by the weight of the Nomex honeycomb in its construction: the weight of the Nomex can be determined during laboratory inspection. Determining that weight would help identify the panel as used in an aisle way or an under seat area. If the MY team is minded, Boeing might answer your question to so as to eliminate any ambiquity for the source, but I don’t regard a find after 22yrs as likely.

  10. DennisW says:

    @Victor

    Nice. I would wonder if the tripartite group of nations is even considering this new information. Is the group still in existence or have they essentially disbanded? Anyone?

  11. Victor Iannello says:

    @DennisW: My understanding is that everybody has been re-assigned to new tasks. Perhaps Malaysia will consider identifying and analyzing the debris. I think that Don has already done the heavy lifting of identifying the shattered floorboard piece, which makes the possibility of a controlled ditching even less probable.

    In my opinion, the main value of the debris is it helps justify a new search along the 7th arc to the north of 25S latitude.

  12. DrB says:

    @airlandseaman,

    You said : “The floor board (#3) is aonother interior piece that further confirms that MH370 is close to the 7th arc.“

    Not really. It only tells us two things: (1) the aircraft interior disintegrated upon impact with the sea, and (2) the impact occurred at a location that brings debris to Madagascar. It does not tell us how far the aircraft crashed from the 7th arc.

  13. TBill says:

    @Victor
    Excellent detective work by the local residents, NoK, Blaine, and Don. Excellent reporting above.

    I hope you are correct about Option-1 though I am getting more towards 1.5 on your scale.

  14. Victor Iannello says:

    @TBill: I was warming up to (3), but I’m told by people with much more domain-specific experience than me that it is very unlikely.

  15. DennisW says:

    @Victor@TBill

    As usual I demand actionable conclusions. (1.5) and (3) – maybe, but so what? An excuse for doing nothing?

  16. airlandseaman says:

    Bobby: Re: “It does not tell us how far the aircraft crashed from the 7th arc.”

    Of course it does not tell us how many miles it is from the 7th arc to 3 significant figures, but it further discounts the various glide fantasies, including glide, dive, glide, etc. IMO, there is a greater chance that MH370 was missed (~3% avg) than the chance it glided further than ~25nm.

    And, it’s not just the piece of cabin floor debris…another interior piece…fractured far worse than MH17 floor pieces. It is also the size and condition of the other 4 pieces that have not been identified yet. One is a 1″ thick piece only 4″ X 4″. Think about the forces required to do that.

    It is also worth noting, except for the flaperon and flap segment, all of the debris recovered, including these 5 pieces, are relatively small. Why haven’t more large pieces been found? Larger pieces would be much easier to find, yet none have been found. I think this is further evidence that the flaperon and flap segment separated in flight.

    Taken all together, the evidence is compelling for a high energy impact NLT 00:21 UTC.

  17. MH says:

    @Don : I read the currents in the IO have seasonable reversal or redirections

  18. Warren Platts says:

    I wouldn’t read too much into the size about these 5 new objects as tending to confirm option 1 above versus option 2. We are pretty sure the aircraft was hijacked by someone. Therefore, it is quite likely that person or persons could have been alive to provide control inputs after fuel exhaustion. Although the weather at the time was not particularly ignorant, the sea level would not have been glassy smooth either.

    Therefore, if a ditching was attempted, imo it likely would have turned out more like that Ethiopian hijacking that ran out of fuel, rather than the Hudson River ditching. The former might be called a “medium energy” event: the entire fuselage was split open, and many small bits of debris consistent the found debris would be produced. Thus the admittedly unlikely scenario of dive-recovery-dive would not be necessary; it would have been a more likely dive-recovery-botched ditching sequence.

    Weighing against the high-speed impact theory (#1 above), besides the negative search results so far, is the fact there were a couple of large pieces, especially the flaperon. In a super high-speed impact, scarcely anything bigger than 1 food across would have been produced, and thus one has to resort to a debatable “flutter” hypothesis to break off the flaperon prior to impact.

    One might also question why MORE debris has not been found? That is, if the entire plane was ground up into 1-foot bits, many of which would float, the number of bits would exceed the number of rubber duckies in an entire container load. There was a book written a while back about just such an event, and the rubber ducks were turning up all over the place. Thus, the relative paucity of found debris weighs against the high-speed impact theory tbqh.

    Agree that not much weight can be placed on drift models. In addition to the usual caveats, I believe a cyclone went through the area not long after the crash that would have caused the actual drift pattern to deviate from models based on average drifts.

  19. Warren Platts says:

    >Why haven’t more large pieces been found?

    That is a good question. The answer, however, would be that if there was an attempted ditching, even one that split open the fuselage like ET961, the vast majority of the plane’s debris would have sank. Even the Hudson River flight 1549 would have sank if left to its own devices. ET961 was visible after the crash only because the water was so shallow.

    Again, the question as to why more SMALL pieces have not been found also needs to be answered.

  20. Victor Iannello says:

    @Warren Platts asked: Again, the question as to why more SMALL pieces have not been found also needs to be answered

    Smaller pieces are less noticeable. I also think the vast majority of pieces have never been reported. The fact that so few pieces that have landed north of Madagascar have been reported (in contradiction to drift model predictions) is strong evidence that the reported debris is a very bad representative sample of the debris produced by the crash.

  21. lkr says:

    @Warren: As you may have seen, most of the regular contributors to this board have been at least open to, if not convinced of, separation of the flap and flaperon in flight. The shattered floor panel certainly raised the likelihood of separation, and makes either dive-glide or dive-glide-dive scenarios much less tenable.

  22. lkr says:

    @Richard Godfrey: [I’m directing to you, but anyone who’s following the drift models may want to weigh in.] Just curious here — assuming there are more pieces all along the western Indian Ocean. What coastline[s] would you most like to see surveyed with an eye to differentiating initial positions along the 7th arc?

    My impression is that Blaine was directed to Madagascar by Pattiaratchee precisely because it would intercept much of the debris from any position on the arc. Most productive, but least informative as to origin, if I understand correctly.

  23. Don Thompson says:

    @MH – just to add to my reply.

    My typical approach, when looking at something held up as potential as 9M-MRO debris, is to consider what else the object might be. Probability is that most flotsam is something else.

    Just as USN ships with SURTASS arrays, spotted off Sumatra, are more likely interested in PLA-N underwater activity than lost airliners.

  24. david Opperman says:

    A date stamp estimate of recovered wreckage would be helpful. Flaperon seems to instigate aircraft was already coming apart before high speed ditch. I would like to see if drift analysis agrees on point of origin of all recovered debris

  25. Richard Godfrey says:

    @lkr

    You asked “What coastline[s] would you most like to see surveyed with an eye to differentiating initial positions along the 7th arc?”

    As you point out, Madagascar is the most likely location for debris to beach and has been very productive, as Blaine Gibson has shown, following the advice from Charitha Pattiaratchi. Every item of debris found can help build the big picture as far as the crash/debris analysis goes. Unfortunately, Madagascar does not help the differentiation of the point of origin along the 7th Arc, as you say.

    We have 4 items of debris reported from South Africa, which are highly likely or almost certainly from MH370.

    With regard to differentiating start locations along the 7th Arc, I would most like to know if there are more debris items to be reported in Tanzania. So far, we only have one item reported, found in a sea cave. It would be interesting if another 3 items were found in Tanzania. Then the next location is Kenya, where nothing has so far been reported. Even if one or two items were reported in Kenya, it would help significantly. In addition, if the arrival time can be assessed, even to a most probable time window, that would be even more helpful.

  26. MH says:

    @Don, isn’t there a southbound current between the African coast and Madagascar.. Wouldn’t this piece get caught up in this North South current since Comoro Islands are at the north end?
    Maybe it then got caught up in the southern currents south of Madagascar?

  27. Warren Platts says:

    >> Why have not more SMALL pieces been found?

    > Smaller pieces are less noticeable and vast majority of bits have not been reported.

    Hi Victor, good points all. But I think the “Friendly Floties” episode may help us to put some numbers to it. Here a link:

    https://en.wikipedia.org/wiki/Friendly_Floatees

    Now the article says that like 2% of debris released into the Pacific Ocean is recovered. However, for our purposes, the debris has to be recovered and recognized.

    Of the friendly floatees, 28,800 of them were released, and 1.4% of them have been found and turned in.

    So let’s say the recovery rate for MH370 is only 1%. So what are we up to now? 32 pieces? That would imply around 3,200 bits of floating debris released.

    But if there was a Silk Air style impact that just shattered everything, there should have been much more released. The floor area in the fuselage is what, 4,000 sq. feet? Wings are also 4,000 sq. feet. Throw in the other control surfaces and wall and ceiling panels, and everything else and we are conservatively looking at what, maybe 32,000 1-square-foot bits-that-could-float conservatively?

    Thus if 32 pieces have been recovered and recognized, that’s a recovery rate of ~0.1%. Is that plausible? Perhaps. Perhaps not..

    But anyways, I think an important question–something some of you all are much more qualified than I to answer–is: What is a rough order of magnitude estimate of the number of recoverable debris objects that would be released under the Silk Air versus the Ethiopian Air crash scenarios? And what should be the expected order of magnitude recovery rate?

    If the “Friendly Floatees” are any guide, that would seem to favor an Ethiopian Air-style of impact. But of course, like everything else in this game, it is possible to concoct all the counterexplanations one could want!

    Too bad we can’t feed everything we know into an AI that would be free from confirmation bias….

  28. Warren Platts says:

    lkr: > most regular contributors…

    Trust me. I have followed this thing very closely from the very beginning. If I do not contribute more posts here than some other people, that is largely because I only type when I believe I have something substantive to say. I am not married to any theory. Therefore, I am open to the flutter hypothesis. But that will remain an open question until and if and when the wreckage is ever found. By all means, if you can raise an extra $100 million to search the northern arc, go for it! Even negative results are helpful..

  29. Don Thompson says:

    @MH

    Current reversals: regionally, yes. For example, across the Maldives prevailing conditions change seasonally, E-W and W-E. I recall Paul S describing that currents in the Madagascar Channel fluctuate seaonally.

    But not the SEC, nor the East Madagascar Current.

  30. Victor Iannello says:

    @Warren Platts: Look at the floating debris field produced by the crash of JT610. The number of parts, the size of the parts, and the size of the floating debris field are much smaller than most would expect.

  31. lkr says:

    @Warren: The wings, aside from control surfaces fairings, and such [which are honeycombed composite], and the fuselage skin are aluminum, and very little would float for more than hours or days. Aside from the tailplane we have a good sampling of the exterior composite structure.

  32. DennisW says:

    @Warren@all

    Funny shit. The debris find conclusions are starting to resemble the estimates of life on other planets – mulitplying an absurd number of weighted guesses to come up with a “conservative” result. The reality is you don’t know anything about the circumstances, the environment, the level of “informed people” in the areas of interest. One thing is consistent. Every time Blaine has looked for debris he has found some. My guess is that not many people in Kenya are reading this blog.

  33. Viking says:

    @airlandseaman

    I looked deeper into the attenuation at GPS frequencies from thunder clouds. It turns out that the dominant term is due to plasma effects in the ionosphere (not the rain). This is created by up-going lightning, in particular from the most powerful thunderstorms in the inter-convergence zone. It can also be created by solar storms. The damping can become so large that it leads to entire loss of GPS signal in particular in the late evening before midnight. Damping of a few dB is more common after midnight (local time). Here is a recent scientific study:

    https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2016SW001439

    It is behind a pay wall so here is a popular summary of the results:

    https://www.dailymail.co.uk/sciencetech/article-3821223/Satellites-suffer-mystery-blackouts-Thunderstorms-edge-space-causing-loss-GPS-signals.html

  34. Peter Norton says:

    ———-
    there are 3 main possibilities that remain:
    (1) further north than the area previously searched
    (2) farther from the 7th arc
    (3) missed or misidentified
    ———-

    The weird thing is that if you follow the arguments, none of them are particularly likely:

    (1) One has to be quite inventive to make the Inmarsat data match (“The challenge in finding automated paths ending further north than 26S is that the reconstructed paths need to curve to the left and decelerate to satisfy the BTO.”) How likely is it that such a path has a perfectly matching straight path cousin just by pure chance ?

    (2) Seems unlikely due to the high+increasing RoD implied by the BFO.

    (3) Unlikely according to estimates by the search crews.

    If you make the assumption that (1)+(2)+(3) = 100% you can come up with relatively high likelihoods for (1) and maybe (2). But looked at either scenario in isolation, none seems to be a good match for the data.

    I also note that for the sake of exhaustiveness a 4th scenario would need mention:

    (4) The crash site is narrowly southwest of the search area.

    This had been discussed quite a while ago but I haven’t seen any mention since. I remember that I had watched Captain Simon Hardy’s 30min 3-part youtube series, which appeared compelling. I don’t remember ever seeing a direct refutation of his technique. The calculated crash site is narrowly outside the search area and at the edge of fuel range (which is a plus if the plan was to maximize distance):

    “Hardy’s calculations put the resting position of MH370 just outside the planned area for the multinational search effort, close to its southern end. So the methodology that the official search team used produced results that are pretty close to the predictions Hardy reached independently.” (source)

  35. Peter Norton says:

    re: scenario 2

    ———-
    Victor Iannello: “Finding reasons for a glide is easier than for the sequence dive-glide-dive.”

    TBill: “I agree passive, uncontrolled descent is one possibility. I was giving the alternate scenario, list of possible reasons for an active pilot descent.”

    Victor Iannello: “Byron Bailey believed an initial high speed descent was necessary for the RAT to produce the required power. He argued it here on this blog. Now, we know that he is mistaken. However, he is a former 777 captain, and his knowledge base might be representative of other 777 captains, so perhaps the MH370 captain had the same mistaken belief.”
    ———-

    In addition to the aforementioned explanations for a dive-recovery scenario (either dive-glide-dive or dive-glide-nodive for which Warren Platts makes good points), let me add another possible explanation:

    Under the assumption that this was a suicide flight, that should get buried in the vast emptiness of the SIO, never to be found, I can understand how a suicidal person could get conflicted once he is actually confronted with this life-ending task and now is confronted with it and has to think about it for 6 long hours.

    The behaviour we see here (dive-glide-…) is very human and observable quite commonly in all parts of life. Example: Spend a hot summer afternoon in a public bath and you will see dozens of children running to the pool’s edge for a running jump, but then aborting, hesitating, and finally jumping without run-up.

    The other alleged cases of pilot suicide have all been very different in so far as the act had to be accomplished quickly (either because the other pilot was in the cockpit and could have prevented it or because people were trying to forcibly open the cockpit door).

    To everyone who says that this is not realistic, I say: suicide might be harder than you think. Nature has build a survival instinct into both animals and humans, so it is quite natural to shy away from it, even if planned methodically.

    We even have an alleged precedent in the case of the Germanwings flight, where the pilot allegedly had already dialed-in a target altitude of 100ft on the previous flight but aborted !

    The bottom line for me is that if this was a suicide flight and the person was still at the controls at the end of the flight after 6 long hours with nothing else to do but to think about it, I could see why the person would hesitate. I don’t think anyone who never has attempted suicide knows what would be “likely” or “unlikely” in that situation (not to mention that peoples’ behaviour can vary substantially from one person to the other). At that point I think a (straight) suicide dive is just as likely/plausible as an aborted suicide from the point of view of human psychology.

  36. DennisW says:

    @Peter

    Yawn.

    The interpretation of the BFO data has been flawed from the get-go. I have no desire to reopen those arguments. My colleagues and I never considered a search to be prudent based on the original ISAT, IG, (that’s the way pilots like to fly airplanes), and Duncan’s Occam’s Razor arguments. You are a relative newbie, without exposure to this history. In spite of that a search was started before any debris was found – an incredibly dumb decision.

    Fast forward to the present. The area from 38S to 25S has been scanned. With that in the bank, it is very obvious that continuing the search by scanning North from 25S is the best course of action. The drift analytics support this choice, and the lack of results below 25S reinforces it. I have no idea where Victor’s head is at relative to a miss below 25S.

  37. Mick Gilbert says:

    @Peter Norton

    Re: ‘I don’t remember ever seeing a direct refutation of his (Captain Hardy’s) technique.’

    Peter, I would have thought that one of the inherent problems with Captain Hardy’s technique is that it assumes a constant ground speed over the final 140-odd minutes of the flight. How likely is that?

    Even if you accept that an assumed constant ground speed is there or thereabouts isn’t he then coming up with a result that is far more ‘precise’ than the data, assumptions and methodology could reasonably produce. I’m pretty sure that I learned something about that in high school science.

    Leaving the methodology and looking at the result, I’m pretty sure that drift analysis largely, if not wholly, excludes a terminus that far south.

  38. Peter Norton says:

    @Mick Gilbert:
    re: ground speed: Oh yes, I missed that. Ok, but that would just introduce some uncertainty/error regarding the crash location, not invalidate the entire methodology, no?

    Some of the error margin would be already covered by the existing search area, given that the calculated crash location is just outside. Another part of the error margin would lie outside the fuel range. Therefore I think the possible area would remain quite small. If Cpt. Hardy’s scenario can’t be refuted, the search area should also be extended a bit to the SW.

    DennisW: “You are a relative newbie”

    No idea what you mean. I have followed this case from day 1 on all sites (DS, JW, here), also actively posting. Although it doesn’t hurt to listen to new voices and to revisit long-held beliefs.

  39. airlandseaman says:

    Viking: Give it up. We have the log with all the C/N0 and BER data for every record. There was no transmission error or loss due to weather. This is old news.

  40. Peter Norton says:

    Ikr: « @Ventus asked a couple of days ago whether a plot of the debris field for Egypt Air 990 is available. […] I was wondering what the seabed record would look like for initially natant debris that “drowned” over a few hours or days. There should be some sorting [by size, material, and whether floating by virtue of large air spaces [eg tanks] or wettable contents [eg, seat cushions and the like] and you’d expect attenuation with time and dispersion. My question was more: If drift carried the initially floating debris across the searched area, wouldn’t there be seabed debris that was initially missed because small/dispersed/etc. that might be recognizable on review with cognizance of precedents like the Egyptair and AirFrance crashes. »

    @Ikr: maybe this might be of interest to you:
    https://sites.math.washington.edu/~morrow/mcm/mcm15/38724paper.pdf
    sections:
    4 Modeling Aircraft Resting Location
    4.3 Forecasting the Sinking Pattern
    4.3.2 Drifting Before Sinking

    In what way would you expect the Egypt Air crash site to differ from the AF447 crash site (for which images are available) in terms of sonar detectability ?

  41. TBill says:

    @Peter
    That idea would come under:
    (4) Pilot elective choice of maneuver
    (c) Indecision or hesitation on planned suicide dive

    https://docs.google.com/document/d/1ELuwGLD0qn7e7ZTEStUEnX3lo3dstYbEkkuOpjaLYO0/edit?usp=sharing

  42. DennisW says:

    @Peter

    No idea what you mean. I have followed this case from day 1 on all sites (DS, JW, here), also actively posting. Although it doesn’t hurt to listen to new voices and to revisit long-held beliefs.

    No intention on my part to degrade your pedigree. I just have no recollection of your participation in the eary issues.

  43. Peter Norton says:

    re: scenario 3 (debris missed or misidentified)

    (A) debris covered by mud

    DennisW: “I find myself wondering if imaging the sea floor is a reasonable model for “detecting” a debris field. Aircraft wreckage should contain quite a large number of specular reflectors as opposed to normal seafloor terrain which is a diffuse reflector. It is hard for me to imagine missing a debris field unless it was masked (not illuminated) in which case an audit would not yield anything new.”

    I was wondering if debris covered by mud would be detectable, but apparently this is possible:
    https://personalpages.manchester.ac.uk/staff/neil.mitchell/sonar/pen.html

      
    (B) debris on rocky seafloor

    Here is an image (MH370 category 2 contact) that shows high contrast of objects on flat seafloor, however I am wondering how aircraft debris could be spotted on the highly chaotic background of rocky seafloor. This image juxtaposes both scenarios:

    http://www.atsb.gov.au/media/5315404/Side Scan Sonar_Cat2_July.jpg

     
    (C) misidentified debris

    Is there an image archive of all category 2+3 contacts for us to review anywhere ?

    Here is an example of a category 2 contact:
    https://worldairlinenews.files.wordpress.com/2015/03/atsb-sonar-2.jpg

  44. lkr says:

    @Peter Norton: Thanks for the reference. What I see relates variables in location of a central debris field [actually, references are almost all to position of the “fuselage” before break-up.

    what I was wondering about was the kind of pattern the fallout from a garbage patch might make, and whether such items, dispersed from a “near-miss” debris field might be recognizable in data from the Fugro and OI surveys.

    Again, Victor’s most recent posts seem to suggest that their methodology should have spotted such fallout.

  45. Peter Norton says:

    fixed link: http://www.atsb.gov.au/media/5315404/Side Scan Sonar_Cat2_July.jpg

  46. Peter Norton says:

    trying again:
    http://www.atsb.gov.au/media/5315404/Side Scan Sonar_Cat2_July.jpg

  47. Peter Norton says:

    last try: cat2 image

  48. haxi says:

    Victor said: I was warming up to (3), but I’m told by people with much more domain-specific experience than me that it is very unlikely.

    What did these people say about the reason why OI missed the submarine the first time?

  49. Peter Norton says:

    @all:
    In reply to Ikr’s and Ventus45’s inquiries quoted above, I would find it useful if we could make a list of sonar images from crash sites.

    Here is what I can contribute:

    Air France 447:
    http://www.isasi.org/Documents/ISASI-2011-Paper-BEA-PHOENIX-WHOI-A4.pdf (p. 6+13+14)

    Egypt Air 990:
    http://www.ibtimes.com/pilot-suicide-when-its-captain-who-crashes-plane-1519756

    Egypt Air 804:
    http://www.deepoceansearch.com/2016/09/09/dos-assist-mh370-search-with-ms804-data
    (Is this supposed to be the crash site? Can’t see any wreckage here.
    BTW, this is what I mean in my previous posting, item B: Would it be possible to identify wreckage on such highly chaotic rocky seafloor?)

    What else do we have ?

  50. Warren Platts says:

    Re: size of debris fields versus number of debris particles

    There is no necessary correlation. It is quite possible for a debris field consisting of 3,200 bits to be the same size as a debris field with 32,000 bits. The latter just has 10X the particle density. Which begs the question: Exactly how many particles are there?

  51. Warren Platts says:

    @ Peter Norton:

    Thank you for the Captain Simon Hardy video references. It is trippy, but my own research using a somewhat separate method landed in the middle of his triangle at the end of the last video.

    I will lay my cards on the table; it will be a fresh take for most here:

    What I did was apply the philosopher Daniel Dennett’s “intentional stance” theory. The idea is to view 9MRO as a so-called “intentional system”. That is, the aircraft itself is viewed as a goal-directed system. It is viewed strictly from a behaviorist viewpoint. There is no attempt to read any minds, or get into any psychology. It is like playing chess online: you do not know if you are playing against an actual human or a computer algo. But it does not really matter. You just hypothesize that whatever it is you are playing against “wants” to win, and so you predict its next move based on that hypothesis.

    Therefore, for our purposes, it does not matter whether MH370 was being piloted by Zaharie, someone else, an evil tiger spirit, an alien, or a computer algorithm installed by hackers. We simply look at the behavior at the beginning of the flight, and project that for the rest of the flight.

    And the behavior at the beginning of the flight is that it was pretty much flying normally, except for the facts that it was not communicating and not flying its official flight plan. That is to say, it was cruising at relatively normal cruise speeds and altitudes, while navigating using established waypoints.

    Thus the hypothesis is that the initial behavior would continue: the plane would continue to fly at normal cruising speeds and altitudes, and continue to use established waypoints.

    And there is only one series of waypoints that fits the criterion of normal cruising speed: POVUS ISBIX MUTMI RUNUT. Interestingly, the track is 189 true–virtually identical to Captain Hardy’s estimate of 188 true derived by independent means.

    Granted, after RUNUT, there is not much out there. To continue the same behavior would require some navigational inputs: that little cover on the TRUE/NORMAL thingy would have to lifted, and 189 true inputted; or another, distant waypoint would be entered, say 69 -69, or perhaps the coordinates for Zhongshan Station in the Antarctic, or the plane could have simply been hand flown. It doesn’t matter which.

    Bottom line is you wind up in the exact same spot that Captain Hardy predicted. And it is not a huge area to be searched: only 6% of what has already been searched.

  52. Warren Platts says:

    >What else do we have ?

    The sonar images from the MH370 searches are amazing. Like the image of the anchor they found on the bottom. It is hard to imagine that they overlooked the wreck of a B777.

    https://www.voanews.com/a/search-missing-malaysia-flight-shipwrecks-underwater-indian-ocean/3177152.html

  53. Warren Platts says:

    @Dennis@all

    >One thing is consistent. Every time Blaine has looked for debris he has found some.

    That is simply not true. He has been looking for years, and while he has found many pieces, if he found pieces every time he went out, he would have hundreds of pieces by now.

    >My guess is that not many people in Kenya are reading this blog.

    You are projecting. People in Kenya are not the hicks you think they are. They have the internet–just like you. They don’t have to read this blog to know about MH370.

  54. Tanmay S says:

    Such a pity that an individual (Blaine Gibson) and the relatives of MH370 victims are themselves funding and organizing debris search along the coast of Africa, whereas Malaysian government is doing absolutely nothing. They can easily co-ordinate with these African countries and carry out a thorough search of the coastlines. I am very sure, there would be many such debris along the entire African shores. Shame on you Malaysian Government.

  55. Mick Gilbert says:

    @TBill

    Bill, your reposting of your paper reminded me that there’s another possible operational reason that you might add to the list. Now I’ll caveat this by saying that I don’t think that this happened (I am firmly in the ‘rapid descent sans pilot inputs end-of-flight sequence’ camp) but for completeness here you go.

    If the intent was a controlled ditching in order to minimise debris then any adequately trained pilot would look to do that with flap to minimise airspeed. Any adequately trained pilot would also look to ditch with power to control the rate of descent but leaving that aside let’s assume the plan was for a dead-stick ditching but with flap. In order to extend the flaps three conditions are required;

    1. You need power for the hydraulics (at least one engine or the APU running),

    2. You need to be at or below the maximum flap extension speed of 250 KIAS, and

    3. You need to be below 20,000 feet.

    The rapid descent may have been an attempt to get below 20,000 feet while the APU was still running to power the hydraulics in order to get the flaps down. That act, however, would seriously limit the distance that the aircraft could then glide.

  56. Mick Gilbert says:

    @TBill

    Just to clarify, when I said ‘That act, however, would seriously limit the distance that the aircraft could then glide‘ the act I was refering to was extending the flaps not the dive per se. Once you wind flap out you’re creating drag so your glide ratio is going to take a hammering.

  57. Victor Iannello says:

    @Peter Norton asked: Is there an image archive of all category 2+3 contacts for us to review anywhere ?

    Yes. There is a report for each contact for the phase of the search that was organized by the ATSB, i.e., the initial 120,000 km2. All data from the search, including the raw data from sensors, has been organized and archived by Geoscience and accessible from is Australia’s NCIS.

    I was particularly interested in contacts identified by the part of the search conducted by GO Phoenix (GP), as the ship operator and the sonar equipment were different (ProSAS vs SSS) than for the Fugro ships, and that part of the arc (around 34S) was interesting to me because of the simplicity of the BEDAX-SouthPole route and because of the excellent fit with the BTO and BFO data.

    I downloaded all of the GP contact reports and merged them into a single document to make it easier to scroll through the contacts. I also identified some of the contacts that I thought were questionable. However, a review with some very knowledgeable people persuaded me that those contacts were very likely not debris from MH370. Also, some of the contacts identified by the towfish were further investigated with a low-altitude pass from either an AUV or ROV. Unfortunately, the contact reports were not consistently updated after these investigations.

    Happy hunting!

  58. Victor Iannello says:

    @haxi asked: What did these people say about the reason why OI missed the submarine the first time?

    It was this sequence of events that made me question whether contacts from the MH370 search were misclassified. I’m told by knowledgeable people close to the search that all contacts were sufficiently investigated until they were eliminated with a very high level of confidence.

  59. DennisW says:

    @Warren

    I remain of the opinion that no conclusions can be drawn based on the number of pieces of debris found.

  60. Victor Iannello says:

    @Warren Platts said: People in Kenya are not the hicks you think they are. They have the internet–just like you. They don’t have to read this blog to know about MH370.

    No recovered debris from MH370 has been reported in Kenya. Yet, drift models suggest some should have reached there. What is your explanation?

  61. DennisW says:

    @Victor

    Relative to your most recent post above I think it is useful to reflect on the history of the search for MH370, and the search decisions previously made.

    As you know, you and I disagreed about starting the initial underwater search. The question posed by the ATSB at that time (before any debris was found) was where should we start an underwater search, not if we should start an underwater search. The IG recommendations at that time were the best available IMO. However, the spreadsheets of the geeks were based on flight path assumptions, and a poor understanding of the limitations of the BFO data. No prudent decision maker would commit to spending a couple of hundred million AU dollars on the information known, and its interpretation, at that time. There had to be significant geopolitical (outside the scope of practice of people here) considerations involved.

    So time goes on. The DSTG report (book?) surfaces supporting the search area selected. Various drift studies are used to further justify the search decisions. The effort is finally suspended, without a happy ending.

    Along comes OI amid new confidence exhibited by the drift modelers. A decision is made to continue the search North from where it left off previously. As above, I disagreed with the decision to start a new search, but did agree with the search area selected. Again, not a happy ending, but the search capability displayed by OI was truly impressive.

    Forward to the present. The combination of the area already searched along with a very realistic Northern constraint based on continued drift modeling (thank you, Richard) makes a new search attractive from a probability of success standpoint. The last statement being true if (and only if) one adopts the “sensible” views:

    1> the wreckage was not missed by previous searches, and

    2> that the aircraft terminated close to the final arc.

    Rejecting either view represents a decision to abandon finding (or even renewing the search for) the wreckage.

  62. Victor Iannello says:

    @DennisW: At this point, I don’t disagree with your proposed way forward.

  63. Warren Platts says:

    >no conclusions can be drawn based on the number of pieces of debris found

    That still begs the question: How many pieces should be produced from a high energy event versus a medium energy event? The former should be an order of magnitude greater. 10’s of thousands versus a few thousand.

    > drift models suggest some should have reached Kenya. What is your explanation?

    It could be a combination of the wreck occurring far to the SW and/or innaccuracy of the drift models. They are, after all, based on average movements of drift buoys. Like I said, a hurricane went through the SIO not long after the crash. We must beware of the illusion of technique. After all, Kenya is loaded with beautiful beaches and luxury resorts. So even if Kenyans are hicks, then what about all the rich tourists that go there? (It is a nice place to go–unlike Somalia–they even speak English there.) Also, it is hard to believe that Blaine has been everywhere in the Indian Ocean except Kenya?

    >Rejecting either view represents a decision to abandon finding (or even renewing the search for) the wreckage.

    There is still the combination of the three lines of position: the 7th arc, the maximum range line, and the 189 true course seemingly implied by the standard cruise speed and available waypoints. This would justify a search in that SW area beyond and to the south of the 25 nm standard band around the 7th arc. ymmv

  64. Don Thompson says:

    @Warren Platts

    a) The mode of impact, low pitch with some forward speed, vs high pitch with little forward speed, is possibly more relevant comparison. For me, anyway. I’ve seen at least 3 small, palm sized, fragments of what looks like interior honeycomb panelling. Paul Smithson found such a fragment on the Tanzanian coast, and NoK filmed during a late 2016 French TV documentaries found such pieces. We’ve now seen a spectrum of sizes from palm sized pieces to the flap and flaperon. Everything’s attached to metal so we can’t expect all buoyant composite fragments to float.

    b) Historical data for the Indian Ocean circulation shows the bifurcation of the Southern Equatorial Current flow at north-east Madagascar isn’t constant all year round. At times, the flow across the north is restricted and directed south instead. I have no information on whether 2014-2015-2016 held to that.

    c) Fugro took the deep tow search out to +90km southeast of the 7th arc, from S37º on the arc to the southern-most extent. Hardy’s spot was covered, I recall he then revised to take a position further out. The deep tow search went out to 52km northwest of the 7th arc along that same segment. So the “mileage was varied”!

  65. TBill says:

    @Mick
    Thank you. I agree the purpose of an intentional descent could be to prepare the aircraft for trimmed level flight when the power is lost, and flaps below 20k feet is a good add, even though I am not particularly thinking flaps were down. I suppose 5% flaps is a possibility. What can be done with RAT power? Will that power A/P or flaps?

    @Warren Platts
    Your path: “POVUS ISBIX MUTMI RUNUT” has merit, but I am thinking it was not RUNUT but instead MUTMI…a path something like ISBIX MUTMI BEBIM (DISCONTINUITY) which basically follows the pilot’s simulator path to about 30 South, such that we should not expect to find NZPG (or 78S67) as a waypoint on the data recorder, rather it might be an overflight of BEMIM towards NZPG. Does that fit your methodology?

    BEBIM looks like an approach to COCOS, so assuming we someday find that crash and data recorder, we will still have the argument of no proven intent for suicide mission. I believe Malaysia specifically asked Australia to surface search in the area of COCOS in the weeks following the loss of MH370 (after Malaysia finally admitted the flight went South).

  66. Victor Iannello says:

    @TBill asked: What can be done with RAT power? Will that power A/P or flaps?

    No.

  67. airlandseaman says:

    Sorry for the repitition, but for new commers:

    On July 30, 2015, I published the following note on the Flaperon: http://bit.ly/2PdLoe6
    Since then, my confidence in that early assessment has increased due to the types, condition and sizes of other debris found, including the 5 new pieces handed over yesterday. Note the following:

    1. We have two distinct groups of debris: (1) 2 relatively large pieces (flaperon and flap segment) mostly intact, and (2) >30 small “shattered” pieces from inside and outside the aircraft that show signs of a very high energy impact.

    2. Large pieces of debris are more easily spotted on a beach compared to the small pieces often found partially buried in the sand. Yet no more large pieces have turned up. Why? AF447, which did not enter the water at very high speed, left a large tail section and other large pieces floating. There should be more large pieces found by now if MH370 broke up following a low speed water landing.

    3. The small fragments are harder to detect on the beach, yet we have >30 now, all indicating a high energy impact consistent with a high water entry speed. There are probably hundreds if not thousands of those small pieces still floating or hiding on the beaches around the WIO.

    4. The 00:19 logon and BFO data indicate FE circa 00:17 (as expected from the fuel analysis) followed by a very high rate of descent (~15,000 ft/min) at 00:19:37. The lack of an IFE logon circa 00:21 indicates likely impact by that time. This 4 minute descent from FE to impact is consistent with those simulator trials where the TAS exceeded air-frame specified limits.

    5. Flight control surfaces are more prone to flutter and possible separation at high speeds if the actuators lose hydraulic pressure, as happens following FE. (The RAT does not provide the pressure needed.)

    6. We know from the ATSB flap segment forensic analysis (AE-2014-054, 2 November 2016) that the flaps were retracted at the time of separation, inconsistent with a water landing attempt.

    7. The lack of any significant leading edge damage to the flaperon or flap segment is consistent with separation in flight, but not consistent with separation at main impact.

    Putting all this together with all the other data and analysis we have makes a strong case for flaperon and flap segment separation in flight, followed soon after by a very high energy impact of the aircraft close to the 7th arc, probably within ~20nm.

  68. Wall says:

    You said, and I quote: ‘there are three main possibilities that remain’.
    I think there are four actually. The fourth one is namely south of 39.6 S. I don’t believe it would be a good idea to look for this plane further north as its average speed would decrease enormously. I think it lies either outside the area searched or it lies south. But I’m curious why you think the fourth one should not be listed in the possibilities.

  69. Victor Iannello says:

    @Wall: Fuel consumption, drift models, past search results. And I already have shown that there is at least one fully automated flight that hits 22S that doesn’t conflict with any of those constraints. Pilot inputs after 19:41 create even more possibilities.

  70. DrB says:

    @airlandseaman,

    You said: “Taken all together, the evidence is compelling for a high energy impact NLT 00:21 UTC.”

    I would say the evidence is compelling for a high-speed impact, but it is only suggestive that it occurred prior to 00:21. You are connecting events separated by at least several minutes with your opinion of how soon the aircraft crashed. I will point out again that NONE of the Boeing EOF simulations succeeded in matching the BFOs and impact at the correct times of 00:19 and 00:21. Can you explain this discrepancy?

  71. airlandseaman says:

    DrB: Maybe Z was flying and caused this scenario. No one knows the answer to your question. Can you explain the lack of IFE logon? If you want to involk the inverted attitude scenario, that only makes my case stronger.

  72. ST says:

    @Victor – Thank you for yet another very interesting summary relative to the new debris finds. It is also helpful that you have provided links to your earlier post regarding possible path with a north terminus.

    Going back to JT 610, it is interesting that with shallow waters, it has been very difficult to find and retrieve the CVR. It speaks to the challenges involved in much deeper waters and difficult terrain for MH370. A new search if initiated should probably look for some of the items that were first located for JT610 such as the landing gear.

  73. Warren Platts says:

    @Don Thompson: >Southern Equatorial Current flow at north-east Madagascar isn’t constant all year round.

    I agree. I happen to be a geologist by training, and I spent a fair amount of time scanning the NE coast of Madagascar on Google Earth looking for likely spots that MH370 debris might accumulate. And it was apparent just from looking at the sand bars that the currents there must seasonally reverse.

    >Fugro took the deep tow search out to +90km southeast of the 7th arc, from S37º on the arc to the southern-most extent.

    That’s what I’m saying. I had that search area marked out on my system, but 90+ km = 50 nm, yet total glide capability is up to 120 nm. (Yes of course a dive-recovery scenario lowers that distance, but still…) The straight and fast path to the extreme southwest remains the most parsimonious flight path imho. Parsimony does not entail truth, however. That I fully admit.

    @TBill: >MUTMI BEBIM (DISCONTINUITY) Does that fit your methodology?

    Not really. That would require a change in the style of behavior. But of course practically anything is possible!

    @airlandseaman: >Sorry for the repitition….

    Don’t get me wrong, sir: you make a very compelling case–about the best one that can be made! The biggest strike against it, however, is that the wreck has not been found despite all the extensive searching..

    By all means, the search should progress further to the north, all the way to Sumatra if necessary! On the other hand, the farther north you go, the closer you get to civilization. Thus one begins to wonder if the crash site was that far north, then why did nobody apparently notice anything?

    @Wall: >The fourth one is namely south of 39.6 S.

    Agree that remains a live possibility. Despite all the theory, your hypothesis has not been empirically refuted.

  74. Peter Norton says:

    DennisW: “No intention on my part to degrade your pedigree. I just have no recollection of your participation in the eary issues.”

    No worries. I’ve been there since the early days though with some ideas and even took a stand for you.

  75. airlandseaman says:

    Warren: Re “The biggest strike against it, however, is that the wreck has not been found despite all the extensive searching”

    I don’t see that fact as relevant. It may have been missed, or it may be further NE. But the evidence supports a POI close to the 7th arc no matter where along the arc. As Victor reports, an expert close to the search has provided reasons to believe it is very unlikely that it was missed. I think the odds are about even that it was missed vs. NE, but it would take much less time to scan virgin territory than several hundred POIs. That suggests the best next place to search is S20 -S25, +/- 22nm.

  76. Niels says:

    @VictorI
    You wrote: “…3. The aircraft debris field was in the subsea area previously scanned by sonar, but was either missed or misidentified.

    Sources close to the previous search effort believe (3) is very unlikely, as there was a thorough review of the sonar data by multiple parties with high levels of experience, and because any “points of interest” were scanned multiple times to ensure the resolution was adequate to make a determination with a high level of confidence.”

    Were these sources refering to the ATSB organized search, or also to the OI effort? My main concern is the coverage in the area around S32.

  77. Mick Gilbert says:

    Always keen to stress the hype in hyperbole News are refering to the recently identified floor panel as a ‘Massive Breakthrough‘.

    In keeping with hoopla the same story refers to Mike Chillit as a ‘mathematician‘!

  78. Peter Norton says:

    @Mick Gilbert:
    re your glide scenario: Could you see any upside for the pilot to wait until fuel exhaustion before extending the flaps ?

  79. Mick Gilbert says:

    @Peter Norton

    Peter, for the avoidance of any and all doubt, let me be clear that it is not my glide scenario; I believe that the aircraft came down without pilot inputs. I was simply providing Bill with another possible (but in my view highly improbable) scenario to add to his list.

    As to waiting till fuel exhaustion to get the flaps down, the former prevents the latter. You need an engine or the APU to be running to power the flaps so it has to be done prior to total fuel exhaustion. You could probably conceive of a scenario where the plane was inadvertently flown till one engine flamed out (maybe the hypothetical perp was engrossed watching The Lego Movie back in business class and lost track of time) and then a hasty Plan B had to be thrown together in order to get the airplane into a ditching configuration.

  80. Victor Iannello says:

    @Niels asked: Were these sources refering to the ATSB organized search, or also to the OI effort? My main concern is the coverage in the area around S32.

    The entire subsea search.

  81. Victor Iannello says:

    @DrB said: I will point out again that NONE of the Boeing EOF simulations succeeded in matching the BFOs and impact at the correct times of 00:19 and 00:21.

    True. However, the timing of the downward accelerating 0.7g descent is very dependent on system interactions after fuel exhaustion, and effects such as fuel slosh, aerodynamic asymmetry, and engine restarts, all of which are either unknown or difficult to accurately model.

    I think the more relevant observation is that in all of the simulations in which a downward acceleration of 0.7g is reached, an impact with the sea occurs within a fairly small radius. With no pilot inputs, that level of downward acceleration is only reached in a steep, banked descent, which would indicate very unstable flight dynamics.

  82. Victor Iannello says:

    @peter McMahon: The readers and contributors of this blog are too smart to have patience for your horseshit. You’ll have better luck on Twitter, Facebook, and in the UK tabloids. You’re banned.

  83. Victor Iannello says:

    @Warren Platts said: Despite all the theory, your [@Wall’s] hypothesis has not been empirically refuted.

    For that matter, any location other than what has been already searched has not be empirically refuted, including the moon. Get back to reality, please.

  84. Peter Norton says:

    Mick Gilbert: “As to waiting till fuel exhaustion to get the flaps down, the former prevents the latter. You need an engine or the APU to be running to power the flaps so it has to be done prior to total fuel exhaustion.”

    Yes. I’m confused now. If I understand your scenario correctly, it takes place after main engine fuel exhaustion but with APU still running, no ?

    Mick Gilbert: “let’s assume the plan was for a dead-stick ditching but with flap. In order to extend the flaps 3 conditions are required […] The rapid descent may have been an attempt to get below 20,000 feet while the APU was still running to power the hydraulics in order to get the flaps down.”

    I didn’t say anything other than what your scenario describes. Or did I ?

  85. Peter Norton says:

    Warren Platts: “The sonar images from the MH370 searches are amazing. Like the image of the anchor they found on the bottom. It is hard to imagine that they overlooked the wreck of a B777.”

    @Warren: The anchor you linked to is a photography, not a sonar image.
    You can’t take ROV photography as a measure for what they might or might not miss in regular AUV surveys.

    Warren Platts: “it was apparent just from looking at the sand bars that the currents there must seasonally reverse”

    How do the sand bars look like to permit this conclusion ?

  86. Mick Gilbert says:

    @Peter Norton

    Peter, IT IS NOT MY SCENARIO. I have nothing further to say on this matter.

  87. Peter Norton says:

    @Mick Gilbert:
    Please calm down. I never alleged you put a high likelihood on the scenario or that you endorsed it. But you wrote it, you are the author. I don’t see why you take issue with that. You added unmistakable disclaimers twice. We all heard you. I don’t know what the problem is. Can we please just talk about the facts?

    I would just like to know, what I got wrong about the scenario ?

    I had the impression that in this scenario main engine fuel exhaustion occurs and then the pilot dives below 20000ft while the APU is still running to deploy the flaps. Or have I misunderstood the scenario? My question was if there would be an upside to wait until fuel exhaustion? I was just interested. (My thought was that the pilot could have extended the flaps before fuel exhaustion to avoid the uncertainties attached to doing so afterwards.)

  88. Peter Norton says:

    By uncertainties I mean all the stuff that comes along with fuel exhaustion – things the pilot knows in theory, but can’t try or practice in the real aircraft.

  89. DrB says:

    @airlandseaman,

    You said: “Can you explain the lack of IFE logon?”

    Of course, as it has been previously discussed several times here. The simplest explanation for the lack of IFE log on at 00:21, assuming the aircraft had not yet crashed, is the APU running out of fuel in the line.

  90. DrB says:

    @Victor Iannello,

    You said: “I think the more relevant observation is that in all of the simulations in which a downward acceleration of 0.7g is reached, an impact with the sea occurs within a fairly small radius.”

    Yes, the Boeing simulations showed that impact occurred soon after high-G descent in the cases simulated, all of which assumed no pilot inputs. Boeing did no simulations of what might happen if pilot inputs were applied after the descent began. So that nearby crash result is based on the assumption of no pilot inputs. That could be correct or it might not be. At this point in time, nobody knows for sure.

    I’m not comfortable with sweeping the timing difference under the rug. It’s telling us something. More simulator runs might or might not clarify the discrepancy.

  91. Victor Iannello says:

    @DrB: We know what could occur with pilot inputs: You would lose several thousand feet in the dive/pull-out/climb sequence, and then potential glide efficiently after that.

    For the plane to have impacted more than 22 NM from the 7th arc, one of two things occurred:

    1) With no pilot inputs, the plane in the steep, banked descent would have to roll to reduce the bank and then glide some distance. No simulations demonstrated this behavior once a downward acceleration of 0.7g was reached. Rather, the simulations showed evidence of a spiral dive, as demonstrated by increasing bank, descent rate, and airspeed.

    2) With pilot inputs, the plane would first enter the increasingly steep descent, recover with skillful pilot inputs, glide some distance, and then enter the water at high speed. The glide could have added a distance of 100+ NM from the 7th arc if the initial steep descent began at 40,000 ft.

    I think that scenario (1) is unlikely, and to search based on scenario (2) results in an unmanageably large search area.

  92. David says:

    @ALSM. Re the new evidence indicating a high descent rate and speed, I agree it adds to that case but it is not as clear cut as you and the press make it sound (e.g. it, “further confirms that MH370 is close to the 7th arc”, and this was, “a very high energy impact of the aircraft close to the 7th arc, probably within ~20nm).

    As to other than the flaperon and flap you say the rest were, “….small “shattered” pieces from inside and outside the aircraft”, but in fact some are relatively large compared to these recent items, including large parts of an engine cowling (item 6), a flap fairing (item 2) and a nose wheel door part (item 18) 3ft long.

    As to no large parts like the vertical stabiliser (as per AF447) and the horizontal stabiliser being evident, they were not detached in the Comoros ditching so that is not evidence there wasn’t a ditching.

    There were parts torn off MH370’s, though that might have been from impact from others parts flying off as per Comoros, bearing in mind that a flaps-up ditching could have been faster than that and into rough seas (Comoros was in a lee, without swell). Also a wing could have dug in.

    You agree that the aircraft might have been piloted. I would go further and say that was most likely unless a relight could bring the BFO timing a long way forward. However even then there was just one of six ‘normal’ configuration simulations that reached the BFO descent rates and acceleration. Even including the ‘abnormals’ (I think they can be discounted for reasons I have argued elsewhere) the point is that shifting the timing only allows a chance of a simulation with a relight being compatible.

    As to distance from the arc to me it is likely that he held the nose down. However, whether he was an intruder, cabin crew member or one of the pilots we know not. Whoever it was the aircraft might have been glided a distance, straight, before impact.

    @Dr B has indicated that IFE non-connection could be from APU fuel shortage.

    Another point you raise is that the flaperons would have been more subject to flutter when hydraulically depowered. Between APU, PMGs, RAT and PSA batteries providing control power, with engine windmilling, APU and/or RAT providing hydraulics, 3 flaperon hydraulic actuators should be operative if not all 4.

    You say, “We know from the ATSB flap segment forensic analysis (AE-2014-054, 2 November 2016) that the flaps were retracted at the time of separation.” I do not think we do. Looking at the evidence I believe the most the ATSB could say is that they were retracted when they collided. I note also that they qualified what they did claim as ‘most likely’.

    It could be speculated that the flaperon could have separated when its designed flutter speed was exceeded. However in my earlier estimation its trailing edge would separate first and when the rest (immediately) followed it would have rotated down so would not be separating from the retracted position. An alternative is the wing being bent back and up, in which case the two might have collided but without detachment right at that point.
    Besides, bear in mind that they were not flush when they collided. The bottom of the flaperon dug into the top of the flap.

    Then you mention, “The lack of any significant leading edge damage to the flaperon or flap segment is consistent with separation in flight, but not consistent with separation at main impact.”

    Aside from the crush damage to the flaperon’s outer leading edge the French report on it points out that there are 4 unexplained vertical cracks along the leading edge, previously undisclosed. Also the flaperon suffered torsion damage, evident particularly at the leading edge.
    The outer flap part had leading edge damage and holes and of course trailing edge damage. On the other hand MH17’s part outer flap recovered after MH17’s crash on land was in surprisingly good nick.

    Even if some of the above proves to be pedantic my point is that future search priorities should be weighed impartially.

  93. DennisW says:

    @DrB

    I don’t have a quarrel with your end of flight concerns, but I think we are trying to address different questions which are:

    1> What actually happened at MH370 end of flight?

    2> What is the best way to continue the search if it is continued?

    In an ideal world the answers should be consistent, but I think this is one of those cases when they might not be.

  94. Wall says:

    @Vicor Ianello

    After looking for information last night, I now believe a position farther south is less realistic than the far north hypothesis. But you should keep in mind, even your 22S calculation doesn’t match the other findings of the ATSB and CSIRO. I think for some reason that you are right about it (the 22S place), but the location is still inconsistent with drift models and/or satcom data. Look, I find it very cool that you guys made it possible to ask questions to each other and to find and share some useful information, but I would advise to respect everyone’s ideas. Your idea seems logic, but it still is inconsistent with the other findings. You’ve advised the ATSB to look for this plane in many places (keep changing), but the reason they kind of ignored the advice was that the places did not match their calculations. I didn’t mean to offend you by the way, but don’t tell people to get back to reality because you believe they are incorrect. You have no proof whatsoever that the aircraft lies in that place. Till that moment, we can only speculate and hope that it will soon be found.

  95. airlandseaman says:

    David said: “As to other than the flaperon and flap you say the rest were, “….small “shattered” pieces from inside and outside the aircraft”, but in fact some are relatively large compared to these recent items, including large parts of an engine cowling (item 6), a flap fairing (item 2) and a nose wheel door part (item 18) 3ft.”

    I’m sure you know, or should know, I meant #6 and #15 were small compaared to the flap and flaperon.. by an order of magnitude. I was not comparing them to the new debris. There clearly two groups as I noted.

  96. Don Thompson says:

    Glides’n’that.

    120nm from nowhere is still nowhere.

  97. Victor Iannello says:

    @Wall: Please read what I wrote. I very specifically commented on the statement that a theory should be seriously considered (any theory, for the matter) on the basis that it has not been empirically refuted. In this matter, people have used that argument to justify all kinds of theories, including the stream of ridiculous stories that appear in the Daily Star.

  98. Peter Norton says:

    @Victor Iannello:
    Many thanks for the links. I have studied all the images in your combined pdf file in-depth.
    My takeaways:

    • If the MH370 crash site looks anything like AF447 or MS990 (posted above), I don’t think it’s possible it could have been missed in non-dataholiday-areas (based on what I have seen thus far, but I have yet to see more category 2 images as your pdf contains only cat 3 sightings). However the impact of these 2 crashes was of lower energy than what many here suspect for MH370. Hence my question if we have sonar images of other crash sites than I posted above ?

    • If only the 2 engines are still in one piece and everything else was shredded into “confetti” with no large sections remaining in 1 piece, than I think it would be hard to distinguish from rocks and geological formations.

     
    Victor Iannello: “I also identified some of the contacts that I thought were questionable.”

    Which contact IDs do you find questionable? I would be quite interested to know after having looked at all of them.

    BTW, do you know what happened to GP-001, GP-032 and GP-042 ?

     
    Concerning the sonar shadows (e.g. the entire right half of the image in GP-013) I am wondering if they had been illuminated in another pass or left in the dark …
    Debris beyond the ridge in the right half of GP-013 would not be detectable I think.

    I’m also a bit worried how the form of objects can be diagnosed (as aircraft debris vs non-aircraft-debris) if the form substantially changes from image to image, e.g.
    compare “cropped image” to “other image” in GP-020 and GP-021.

    “Also of concern was the fact some things appeared in one picture and did not appear in the other.” (see link below)

    But I don’t think the answers to that will substantially change my 2 takeaways above. (I’ll update them once I have seen more cat 2 images).

     
    Here is a critical view of the underwater search reliability. Although it probably should be read with a grain of salt, given that the critic seems to come from a competitor:
    http://www.news.com.au/travel/travel-updates/us-firm-warns-atsb-may-have-mistaken-mh370-debris-for-rocks-in-their-deep-ocean-search/news-story/b7fe40a415d223d738023c0ef98b4f2d

  99. DennisW says:

    @Peter

    So Peter. What do you recommend doing next? That is the question. Not your BS history of what has been observed. Stop the nonsense, and make a recommendation.

  100. TBill says:

    @DennisW
    If there were prospects of a continuing MH370 search in a few months, then we need to stop analyzing and make a recommendation. Which normally @Victor is good about asking particpants for inputs at those times.

    You seem to be fiercely defending the proposed 20-25 South +/-22nm search zone, and trying to discount other ideas becuase that may risk losing the clarity of a solid recommendation.

    But I for one am really looking beyond that 20-25 South search phase, if we do start to look wider, where should that wider search be? I submit the only actual flight path plan we *might* know is the simulator path, or a path closely related to that. Which ends up being somewhat close to DrB’s path and Nederland’s paths, crossing Arc7 in the 30S region.

  101. DennisW says:

    @TBill

    Great. Your answer is typical bullshit – no specific area. My area is specific 25S to 20S +/- 25km. Quit trying to sound smart, and make a specific reommendation.

  102. Donald says:

    @Don Thompson

    Re: Glides’n’that.

    Presuming that Z was a rational actor is to be very uninformed and unenlightened about potential states of mind in suicidal/homicidal individuals. What would be pointless and nonsensical to most may make all the sense in the world to an individual ‘suffering’ from psychosis and possibly other ailments. One simply need to take a moment or two to reflect on the sheer ‘insanity’ of MH370 to hopefully understand this in a clearer light.

    I continue to see the applications of ‘rationale’ and ‘logic’ to explain and attempt to justify the most likely EOF scenarios (impact ‘attitude’, glide/no glide, mode of death– hypoxia/high speed dive/drowning/other). This is a fools errand.

    IMO Z was alive until EOF and simply pointing the nose down and being done with it all would have been a much too mundane final act…and very much beneath him. But who really knows?

    A frequent poster on here has even used the APPARENT absence of mental health issues in Z’s past to claim he most probably wasn’t suicidal!!! That’s just ignorant, at best. (looking at you, Dennis).

    There appears to be a growing confidence that the plane will be found very near to the arc. I dare say I feel it atm to be misplaced but hopefully more debris is discovered and more forensic analysis appears from the ether.

  103. Richard Godfrey says:

    @Viking

    Please find below 2 links to my drift analysis from your end point at 13.279°S 106.964°E.

    I previously published the simulated debris path, excluding the effect of Tropical Cyclone Gillian, which ends after 266 days near Mombasa in Southern Kenya.

    The track passes Reunion after 208 days, around 1,256 km north of St. Andre, where the flaperon was found.

    https://www.dropbox.com/s/7odxz0b64wiflas/Drift%20Map%2013.2790S%20106.9640E%204.8228S%2039.8585E%20265d.png?dl=0

    The simulated debris path, including the effect of Tropical Cyclone Gillian, ends after 272 days near Zanzibar in Tanzania, 204 km further south. In the linked graphic I have marked the path of Tropical Cyclone Gillian. Gillian affected the simulated debris path for 6 days between 21st March 2014 and 26th March 2014, which I have marked in dark blue on the track depicted in the link. The result was that the simulated debris was blown around 1° of latitude further south during the Tropical Cyclone Gillian and ends up around 1.8° of latitude further south on the eastern coast of mainland Africa. Simulated debris also took 6 days longer to arrive in Africa because of the Tropical Cyclone Gillian.

    The track passes Reunion after 223 days, around 1,106 km north of St. Andre, where the flaperon was found.

    https://www.dropbox.com/s/p4qic989vrkisby/Drift%20Map%2013.2790S%20106.9640E%206.5895S%2039.3779E%20271d.png?dl=0

  104. Richard Godfrey says:

    @Viking

    I previously published a link to my MH370 Flight Path Model V17.0 with your route.

    I missed the fact that you asked “Could you tell me if the 26.1 Hz error you find refers to my table 3 or to the numbers in the Note added in proof?”

    The answer is that I was referring to my MH370 Flight Path Model V17.0 and not your table 3, nor to the numbers in the note added in proof.

    As I stated previously, in Column T you will see the “”Error Calibrated BFO (Hz)” increases steadily to 26.1 Hz as the flight progresses, despite including the “Aircraft Compensation Doppler (Hz)” in column DN and the “EAFC Effect” and “Eclipse Effect” in columns DQ and DR. I am referring to the Columns in the attached Excel representing my MH370 Flight Path Model V17.0. In other words, I was confirming the point that Victor has already made, that your route does not fit the BFO data. Sorry for the confusion.

    https://www.dropbox.com/s/5oghhdzpkj2bkng/MH370%20Flight%20Path%20Model%20V17.0%20MK.xlsx?dl=0

    For your information, my current MH370 Flight Path Model was first published on 9th December 2015, on Duncan Steel’s website: http://www.duncansteel.com/archives/2102

  105. Sid Bennett says:

    I have not commented for some time, but in reading the latest discussion, the balance of opinion seems to be that there was a pilot input at the EOF. That said, the only real argument against a glide is that the search area would be so vastly increased as to be impractical with existing technology.

    Please cast your mind back to the early days of the analysis of the satcom data where the IG came to a clear consensus of an impact region that was conceptually different from the official
    conclusion.

    Rather than a long stripe along the 7th arc, it was an approximate point with a small radius determined by a final descent without pilot input. Only when the first search failed and the drift analysis was used to justify a second search did the various alternative flight paths begin to arise, and they needed multiple path changes to fit.

    Surely we should return to the approximate IG original location and do a 120km radius search without reference to any other search to the North. (None of the Boeing simulations have pilot input..it would be politically incorrect!)

    I express my appreciation to the members of the IG and to those of you who have provided such thorough analysis of that little data that is available.

    Sid

  106. Warren Platts says:

    @Victor > “The Moon.”

    Ah the Moon is indeed my bailiwick! 😉 Yes of course the mere fact that theory T has not been empirically refuted does not entail that T is likely or even possible. I think we can all agree the 9-MRO is not in Russia.

    The question is whether we can rule out a far SW crash site on theoretical grounds, given the possibility of a glide scenario. And at least a partial answer to that seems to be that glide scenarios result in an unmanagably large search area.

    Now that is a highly debatable principle of epistemology! It is like trying to guess a password, but limiting yourself to words in a dictionary because random strings would produce too big of a search space.

    For MH370 glide scenarios, that would be the case if there was an equal probability of crossing the 7th arc at all points. But that need not be the case. Let us remind ourselves why the search has been progressing to the NE: because in the beginning it was thought that paths more to the SW were more likely in the Bayesian sense.

    So the thing to do imo is revisit those reasons and decide on a likely scenario, and then search farther to the south for that area in case there was a glide.

    Personally, I still favor Captain Hardy’s assessment of a crossing around 39S: it has the virtues of being where the max range and 7th arcs cross, a straight path at crusing speed leads right to it, there is a nice line of waypoints that point right to it, and behavioristically, it is my claim that such a path would be most consistent with the aircraft’s behavior of the path we know about through radar.

    But maybe DrB’s crossing at 30S is more likely. That is for the experts to decide. Either way, to search the area at 39S or 30S for glide scenarios is not unmanageably large.

    Indeed, one could argue that getting married to the <20 nm out-of-control spiral hypothesis is what causes an unmanageably large search area. Why? Because if you do not find the wreck on your favorite path then your only option is to keep searching the 7th arc itself all the way to Indonesia. But maybe, just maybe by searching 100 or 120 nm on the most likely path might yield successful results.

  107. Warren Platts says:

    What Sid said! (I was typing my response when he posted his–no plagiarism I swear!)

    @ Peter re: Madagascar currents: Sand on beaches tends to move in the same direction as the prevailing currents. So if you are on an east coast and the current is moving north to south, sand (and other debris) will tend to pile up on the north side of any promontories that stick out. It just seemed to me that by looking at different times of the year the patterns were varying. I didn’t keep any notes, but that was my general impression that the current reversed itself seasonally.

    @Donald >This is a fools errand.
    Agree that it is a fool’s errand to try and read the mind of who or whatever was piloting the aircraft. Nonetheless, one can attempt a behaviorist (in B.F. Skinner’s strong sense where you make zero assumptions about states of mind) analysis and make future predictions of the aircraft’s flight based on the earlier observed behavior.

    In our case, the behavior seemed to be a strategy of hiding in plain sight, as it were. I imagine that on any given night there are any number of military planes flying that have their secondary radars turned off. But if the plane is flying at normal speeds and altitudes, is flying along established corridors, and doesn’t violate anyone’s sovereign airspace, if it were picked up by a primary radar somewhere, that might not ring any alarm bells. Hence the approach-avoidance behavior demonstrated by the going around of Sumatra instead of straight across it.

    Thus in the latter part of its flight, if the aircraft followed an established corridor (say POVUS MUTMI RUNUT) at normal cruising speeds and altitudes, it would look like a military transport on its way to maybe New Amsterdam, or the Russian or Chinese bases in the Antarctic that are in line with that path in case it happened to get picked up by the primary radar of a passing navy ship out there.

  108. Victor Iannello says:

    @Warren Platts: I said in a message to @Wall, an impact at 39S latitude is contradicted by fuel consumption, drift models, past search results.

  109. DennisW says:

    @Warren

    The Inmarsat data cannot be used to refine the flight path to a particular location aong the 7th arc.

    The Inmarsat data can be used to:

    1> Infer the plane went South after the FMT

    2> Infer the plane impacted very near the 7th arc based on the BFO data at 00:19:XX.

  110. Warren Platts says:

    @Victor Iannello > “fuel consumption, drift models, past search results”

    OK wait a sec. I just watched that Simon Hardy video, and he was saying the maximum range line crossed the 7th arc at approximately 39S. That matches my recollection of when I worked on this stuff years ago. And I believe Richard Godfrey was also producing paths in the exact same area. Granted, it is at the edge of what is possible; it is not, however, like saying the aircraft went all the way to RERAB or something, which would be theoretically impossible, even if it hasn’t be empirically refuted. If you disagree the 9-MRO could make it as far south as 39S, that’s cool: in which case I ask where you think the maximum range line crosses the 7th arc.

    As for the search results in that area: evidently, the people in charge of the actual search did in fact search the vicinity of 39S, but only went about 50 nm beyond the 7th arc, so if there was an extended glide, then it is theoretically possible they very well could have missed the wreckage, assuming the plane actually traveled that 188/189 corridor. So the search does not rule out a SW wreck.

    Finally, there are the drift models. The argument is that numerical simulations predict that a crash at 39S should result in debris washing up in Australia, a place where the people are not hicks, or at least they are well educated hicks who regularly clean their beaches and were on the lookout for such debris.

    That is the strongest argument against a SW crash. However, the creators of these numerical simulations are the first to admit their limitations. For example, and this is a big one, they don’t take into account major storms. Yet I am pretty sure a major typhoon went through the area not long after the crash; correct me if I’m wrong. So who’s to say why debris was not detected in Australia? Has there ever been a systematic search of the Great Bight? Obviously not. It is too rugged.

    But by all means, make your recommendations to search 25S to 20S. Airlandseaman said their was an 80% probability of finding the wreck in the Ocean Infinity zone; now DennisW thinks there is a 80% chance the wreckage is in the 25S to 20S zone. Sounds good to me. That guess is as good as any. Since it is such a sure bet I am sure he will pony up of few of his many multi-multi-millions, organize an expedition, and garner everlasting world fame when he finds it, and maybe even a tidy profit! 🙂

  111. Warren Platts says:

    @DennisW: > Inmarsat data cannot be used to refine the flight path to a particular location aong the 7th arc.

    I agree. That is my point exactly. That is why it is a good idea to bring other pieces of information into the puzzle. Like the maximum range line. And like how a path down 189T, that happens to correspond with a line waypoints, is about the only straight-line, more-or-less constant velocity path. Especially at cruising speed, which the aircraft was more or less traveling at in the dash up the Malacca Strait. All those extra bits of info point to the SW, as opposed to the NE.

    imho ymmv

  112. DennisW says:

    @Warren

    Airlandseaman is right.

  113. DennisW says:

    @Warren

    Our posts crossed. More to say later.

  114. Victor Iannello says:

    @Warren Platts: In rough numbers, the estimated timing of fuel exhaustion (around 00:15 UTC) corresponds to fuel flows from the LRC table plus about 1.8% to spare. However, the average engine PDA is 1.5%, and the extra fuel flow due to the ISA+10K temperature is 3.4%, which means at LRC speed, the fuel quantity is short by about 3.4% + 1.5% – 1.8% = 3.1%. Now, to reach 39S latitude requires a speed of about M0.84 at FL350, which is substantially faster than the average LRC speed, which means the fuel flows are even higher. In a nutshell, there is not nearly enough fuel to reach 39S latitude.

    As for drift results, independent of the debris reaching Western Australia, an impact at 39S would result in debris reaching East Africa much later than observed. That includes the predictions by David Griffin, who used a drift model with sea and air conditions that existed at the time.

    It’s true that there might have been a dive-glide-dive scenario that took the plane further than 50 NM from the 7th arc. I view this as possible, but unlikely. Together with the fuel and drift results, I believe the prospects of MH370 crossing the 7th arc at 39S latitude and impacting at 50+ NM from the 7th arc as very unlikely, but not impossible.

  115. Peter Norton says:

    Don Thompson: “Hardy […] then revised to take a position further out. […] So the “mileage was varied”!”

    Is the moving-goal-posts criticism fair, considering how much the IG recommendation has moved over the years? 10 times as much? 100 times?

  116. MH says:

    what might be the latest on B737MAX’s air worthiness and workarounds for stall issues?

  117. Peter Norton says:

    > Victor Iannello:
    > the average engine PDA is 1.5%, and the extra fuel flow due to the
    > ISA+10K temperature is 3.4%, which means at LRC speed, the fuel
    > quantity is short by about 3.4% + 1.5% – 1.8% = 3.1%.

    If we assume
    • the most favorable conditions for southern latitude (e.g. LRC, gradual climb, FMT as early as possible, etc.)
    • the pilot pulled out all possible range-extending stops and tricks (examples)
    • stay very conservative (i.e. include safety margins) for all values including fuel consumption, wind, etc.

    … then what is the southernmost latitude on the 7th arc that could theoretically have been reached ?

  118. Peter Norton says:

    ———-
    Victor Iannello: “there are 3 main possibilities that remain”

    Wall: “I think there are 4 actually. The 4th one is namely south of 39.6 S. I don’t believe it would be a good idea to look for this plane further north as its average speed would decrease enormously. I think it lies either outside the area searched or it lies south. But I’m curious why you think the 4th one should not be listed in the possibilities.”

    Warren Platts: re: “The fourth one is namely south of 39.6 S.” “Agree that remains a live possibility.”

    Victor Iannello: “I believe the prospects of MH370 crossing the 7th arc at 39S latitude and impacting at 50+ NM from the 7th arc as very unlikely, but not impossible.”
    ———-

    I agree with others here that this 4th scenario should be included as one that merits thought. At any rate by the standard that many here (I think including you, Victor) deem scenarios 2 and 3 highly unlikely.

  119. DennisW says:

    @Warren

    @Warren

    Our posts crossed. More to say later.

    My guess is that you are living with your parents. Am I right?

  120. David says:

    @Mick Gilbert. Lion Air. Thank you for your fine summary of November 28th at 10:39 am. I regret it has taken a while but I have prepared some thoughts, as below:

    https://www.dropbox.com/s/5pvrcq2y8kuzufx/Lion%20Air%20accident.%20Review%20of%20some%20Preliminary%20Report%20aspects.docx?dl=0

  121. Mick Gilbert says:

    @David

    As you might reasonably infer from that extraordinarily high number of hours and cycles per days in service, Lion were flying the paint off that airplane. There are actually many flights ‘missing’ between those AFML entries. The few days before the accident looked like this:

    October 25
    JT2656 Tianjin Denpasar
    JT3931 Denpasar Surabaya
    JT3930 Surabaya Denpasar
    JT2623 Denpasar Shanghai

    October 26
    JT2622 Shanghai Denpasar
    JT776 Denpasar Manado
    JT2749 Manado Tianjin

    October 27
    JT2748 Tianjin Manado
    JT775 Manado Denpasar
    JT828 Denpasar Lombok
    JT829 Lombok Denpasar
    JT776 Denpasar Manado

    October 28
    JT775 Manado Denpasar
    JT775 Denpasar Jakarta

  122. Peter Norton says:

    Don Thompson: “Fugro took the deep tow search out to +90km southeast of the 7th arc, from S37º on the arc to the southern-most extent. Hardy’s spot was covered.”

    To what exact latitude has the 7th arc been searched on the southern end ?

    Hardy’s area extends up to 120nm from the 7th arc, so based on that alone it cannot have been fully covered by the numbers you cite.

    https://i.imgur.com/LZB5bZ9.jpg
    https://i.imgur.com/bwDqRBA.jpg

  123. Peter Norton says:

    ———-
    Warren Platts: “Let us remind ourselves why the search has been progressing to the NE: because in the beginning it was thought that paths more to the SW were more likely in the Bayesian sense. So the thing to do imo is revisit those reasons and decide on a likely scenario, and then search farther to the south for that area in case there was a glide.
    Personally, I still favor Captain Hardy’s assessment […]: it has the virtues of being where the max range and 7th arcs cross, a straight path at crusing speed leads right to it, there is a nice line of waypoints that point right to it, and behavioristically, it is my claim that such a path would be most consistent with the aircraft’s behavior of the path we know about through radar.”

    DennisW: “Inmarsat data cannot be used to refine the flight path to a particular location aong the 7th arc.”

    Warren Platts: “I agree. That is my point exactly. That is why it is a good idea to bring other pieces of information into the puzzle. Like the maximum range line. And like how a path down [188T], that happens to correspond with a line waypoints, is about the only straight-line, more-or-less constant velocity path. Especially at cruising speed, which the aircraft was more or less traveling at in the dash up the Malacca Strait. All those extra bits of info point to the SW, as opposed to the NE.”
    ———-

    I find myself in agreement with many points Warren Platt makes.
    The fact that the “Hardy area” is at the edge of MH370’s range makes it more probable rather than less, in my opinion, given that this could very well have been the plan.

    I’m very uncomfortable with the idea that MH370 might have been narrowly missed in the south and I would welcome this small part being checked.

  124. Victor Iannello says:

    @Peter Norton said:The fact that the “Hardy area” is at the edge of MH370’s range makes it more probable rather than less…

    Based on our best estimates, there was not enough fuel by a fairly significant amount.

  125. Peter Norton says:

    > DennisW: “So Peter. What do you recommend doing next?”

    I deem a failure any search which takes longer than a “stupid” brute force attack (=starting at LKP and spiraling outward). To my knowledge, this includes AF447 and SanJuan, as DrB pointed out.

    As far as MH370 is concerned, I have already stated being in agreement with you, Dennis, in so far as since 2014 I had been recommending searching the 7th arc by sequentially expanding the width at a ratio of 1:4 inside:outside absent another considerably more promising strategy.

    Given that the search was done differently, what I would suggest NOW (in this order):

    1. First, search the Hardy area which includes the Goldfinch fracture zone. This is only 7000 km² and, if Don is right, already partially covered.

    2. Keep trying to solve the MH370 mystery. Doing so might yield a faster strategy than a stupid brute force attack.

    3. If (2) has not succeeded and thus no better strategy is available at the moment when the search is about to resume, then brute force attack the remainder of the 7th arc up to whatever reasonably data-defined northern limit by sequentially expanding the width at a ratio of 1:2 inside:outside until 12nm:24nm is reached.

    4. I never really understood why MH370 could have travelled just as far inside the 7th arc as outside (explanations welcome!), but if that is the case then search the remaining 12nm on the inner side (to get to ±24nm on both sides).

    5. If (2) still has not produced any better strategy, then search selected small arc segments to 120 nm glide range (e.g. DrB’s path, etc.)

  126. Peter Norton says:

    > Victor Iannello: “Based on our best estimates, there was not enough fuel by a fairly significant amount.”

    What is the southernmost latitude on the 7th arc that
    (a) was searched
    (b) could theoretically have been reached with the available fuel ?

  127. Victor Iannello says:

    @Peter Norton: It is not trivial to determine the farthest point south that is reachable from fuel considerations because it depends on how the plane was flown between 18:22 and 19:41. Because a path at LRC speed does not have sufficient fuel endurance, significant portions of the flight have to be at lower speed, which may include some combination of a descent and slow down to holding speed before 19:41, or slower than LRC speeds after 19:41. It becomes more difficult to find an acceptable combination of speed/altitude changes for latitudes farther south than around 35S because the speed after 19:41 needs to be greater than LRC to satisfy the BTOs, and there is not enough time for a descent and/or slow down before 19:41 to balance the fuel required after 19:41.

    All that said, our best modeling says there is not enough fuel to reach Hardy’s 39S, and there should be enough fuel to reach Bobby’s 31.6S.

    If we are going to search wider at latitudes already searched, I think it is difficult to select a practical latitude range. I have always been fond of 34.3S because of the simplicity of the BEDAX-SouthPole great circle path and because the BTO and BFO fits are excellent (which is why Inmarsat chose a very similar path). However, if there were pilot inputs after 00:19, there is no reason to assume there were not pilot inputs between 19:41 and 00:19, and it becomes very difficult to specify a narrow latitude range for the search without following “hunches”.

  128. Richard Godfrey says:

    @Victor

    Using Barry Martin’s fuel model based on RR Trent 892 engines, the absolute maximum range, based on a FMT immediately after 18:40 UTC, ended at 38.5S.

  129. Victor Iannello says:

    @Richard Godfrey: If I recall, Barry primarily used the LRC tables to derive range. What PDA and temperature correction did Barry use?

  130. Richard Godfrey says:

    @Victor

    Both PDA and temperature correction are user parameters in Barry’s Excel Spreadsheet, which can be downloaded at http://www.aqqa.org , where I used his default settings.

  131. Victor Iannello says:

    @Richard Godfrey: I can’t find his default values in his spreadsheet. Without understanding the PDA and temperature correction he used, it is hard to understand how he arrived at 38.5S latitude, which I believe is unreachable by a fairly significant amount.

  132. Don Thompson says:

    @Peter Norton

    The Geoscience Australia website “The Data Behind The Search for MH370” shows overlays for the MBES seafloor survey and all side scan imagery that was acquired. The map and overlays are directly generated

    The deep tow survey lines did not all end at precisely the same latitude. I can correlate the website overlays to the data I have on GIS and Google Earth to confirm that the entire swath was comprehensively swept by the deep tow UVs to a line perpendicular to the 7th arc at S39.3º. In the extreme SW corner, S40º was reached.

    Beyond the position just quoted by Richard, southwards, 26,600km² has been surveyed.

  133. TBill says:

    @Victor @all
    Re: Fuel Consumption and Aircraft Performance

    Based on recent IG analysis, it would appear MH370 climbed to FL430+ and sped up after the U-turn at IGARI, seemingly defying the normal B777 aircraft operating envelope. This possibly implies unorthodox manipulation of the aircraft…techniques such as turning off bleed air and perhaps cutting off IDG’s are suggested.

    What I have been trying to understand is, what are all of the possible overhead panel electric configurations that could account for SDU (Left Bus) outage at IGARI? And what would the performance and fuel consumption of the aircraft look like for each config?

    1. Turn OFF L IDG and L TIE
    2. Turn OFF L GEN CNTRL and L TIE
    3. and 4. Options 1 and 2 and also L XFER OFF
    5. Turn OFF L and R IDG (keep APU Off in all cases)
    6. Turn OFF L and R GEN CNTRL…
    7. and 8. Options 5 and 6 asnd L XFER OFF
    9. Turn off Breakers in MEC Bay (however, out of scope of this list)

    My flight sim version PSS777 does not tell me when SDU is off, so I may need to graduate to PMDG777, if I thought it was accurate about giving the “SDU off” message in all of the possible elec config cases.

    Bottom line I am not too confident of fuel consumption limits.

    >>My issue with 39S is that the “perfect” BFO fit falls apart after about Arc5 suggesting MH370 could have changed heading to the east. If we have an active pilot making all kinds of maneuvers up to 18:40, why does it suddenly make sense that all maneuvers would cease? BFSkinner said so? The 39S story is the active pilot got lazy for 6 hours and did a straight flight southwest, woke up, and glided the aircraft 120nm off the prior search zone. I see no logic or unique fit to the data, unless we argue for a sleeping pilot.

    The whole problem with finding MH370 is the assumption that straight, passive flight is somehow the more likely, more rigorous base case, and active pilot is somehow a weak speculative case. But passive, straight flight is perhaps the greatest speculation of all.

  134. Victor Iannello says:

    @Peter Norton asked: What is the southernmost latitude on the 7th arc that…was searched

    On the 7th arc, the furthest point south was around 39.40S latitude. At that latitude, the search perpendicular to the arc and to the southeast extended about 50 NM to (-40.12,85.99).

  135. Don Thompson says:

    @Peter Norton,

    Yes, the criticism is fair. Lest I missed any important points I reviewed, again, Hardy’s three video clips.

    As I wrote, the seabed in the area that he first advocated was searched by Fugro. ATSB gave his work some consideration, I assume they bounded it with their own conclusion for a maximum distance, wide of the arc, to ocean impact. The area was searched out to +90km beyond the 7th arc. Hardy persists that his thesis is correct so moved his ‘X’ further along his ‘constant speed line’, i.e., at main engine fuel exhaustion the aircraft altitude was near cruise level, therefore, it ‘glided’ further. In the 60 Minutes MH370: The Situation Room programme Hardy states he does not believe the idea of an uncontrolled descent. He believes 9M-MRO was controlled to the end.

    When considering impact points the IG, and others, have re-evaluated what is most likely using the available information at a point in time. The progress of the search brings new information: areas where the aircraft has not been found.

    I’ll comment on what I consider to be the weaknesses of Hardy’s thesis:

    Aircraft speed: Mick Gilbert mentions above, Hardy’s method considers a constant ground speed. Aircraft do not fly at constant ground speed. Hardy himself showed a glimpse of flight planning material for a KUL-PER flight. That worksheet included wind fields, yet he chose to ignore it. Considering winds for airspeed would revise the arc crossing point north-east wards.

    Tools: Hardy describes using Google Earth as a tool to determine his paths. Google Earth calculates paths between two points as great circles, so are only accurate if the start and end point are valid. An autopilot only navigates great circle paths when it’s receiving LNAV guidance from the FMS to a waypoint. This approach is inaccurate.

    Aircraft knowledge: Hardy appears to assume that a Boeing 777 will fly serenely onwards without engine power, holding the heading at the time of fuel exhaustion, descending without without stall or phugoids. I do know that Hardy described testing a dual engine failure scenario in his airline’s Level D simulator. His experience is at odds with the exercises undertaken by Boeing, @airlandseaman, and ignores that R-R Trent engines attempt to auto-relight (Hardy flies the GE powered 777-3F2). Those tests undertaken by Boeing and ALSM show that a 777 enters a spiralling descent after main engine fuel exhaustion. The Trent engine attempts, automatically, to relight if the fuel cutoff switch remains in the RUN position. Relights attempts may well cause attitude upset.

    Evidence: most significantly, he ignores the BFO metadata recorded during the 00:19 Log On. He accepts the BFO analysis that determined the aircraft flew south, but not that the records at 00:19 indicate descent >13,800fpm (per Holland).

    So, feelings or facts?

  136. Don Thompson says:

    Clarifying one of my paragraphs above…

    Aircraft knowledge: During 2017 it was described to me that Hardy tested a dual engine failure scenario in his airline’s Level D simulator finding that the aircraft, hands off, continued to hold heading with wings level and it began a stable descent without upset. That conversation was notable as it described a scenario contrary to the Boeing simulator trials and the exercise conducted by ALSM. However, during the 60 Minutes MH370: The Situation Room programme (May 2018) Hardy demonstrated a scenario in which the aircraft had entered a steep descent from 35000ft and, at reaching -9000fpm, he then recovered to controlled flight at approx 29000ft (PFD read off screen). No explanation was given for the conditions in which the descent had initiated, nor why he delayed his intervention. It would be useful for Hardy to explain his findings and present conclusion for the 777’s flight characteristics after fuel exhaustion. At this time, I don’t find his scenario consistent or persuasive.

  137. Tim says:

    @All,
    Since the release of the SIR report it is clear that the autopilot is OFF after IGARI(aircraft is not flying a constant track and is probably flying phugoids.

    To make the assumption that the autopilot is re-engaged at some time is not necessarily logical. This is a huge factor in where and how far it flew….. time to reconsider this as a flight with no nefarious Imputs to the end.

  138. Tim says:

    @Don,
    Hardy must have had the APU running before he failed the engines. Without the APU already running the aircraft would definitely quickly enter a steep spiral once the second eng fails.

  139. Richard Godfrey says:

    @Victor

    I have re-downloaded Barry Martin’s model V7-9-4 from http://www.aqqa.org as I had played around with the settings so much, I wanted to check that I had a clean version.

    On the tab “MAIN”, you can set up the start time, start position, altitude, etc. I left the default settings.

    The default setting for engine type is for the Rolls Royce Trent 892 engines. I selected the speed mode to LRC and the Path Navigation to True Track. In columns L and M, you can set up the FMT and True Azimuth. I just set a 180° True Track from 18:25.

    I found the PDA in column FK and row 16.

    The Temp Deviation is shown in Column AE and the data is in the tab “FIELDS” from row 135 onwards.

    There is an initial ∂Ts in column FM row 13.

    The remaining fuel is shown in column FV and the distance travelled is shown in column FN.

    For a PDA of 2%, fuel exhaustion is at 00:03 UTC at 38.493°S.

    For a PDA of 3%, fuel exhaustion is at 23:59 UTC at 37.994°S.

    It is a massively complex Excel and maybe I have made a mistake somewhere.

  140. Victor Iannello says:

    @Richard Godfrey: First, assuming no descent, the turn south had to occur between 18:28 and 18:40, not at 18:25, which is why it appears that 38.5S was reached. But looking at the fuel endurance results, and taking the results at face value (always dangerous without knowing the details), each 1% of PDA reduces the endurance by 4 mins. For a PDA of 1.5%, there would be 2 min more of endurance than 2%, i.e., fuel exhaustion of 00:05. That’s about 10 minutes less than observed.

  141. Don Thompson says:

    @Peter Norton

    Concerning your 4) above, review earlier post

    @Tim and “Hardy must have had the APU running“. He ought to know better.

  142. Peter Norton says:

    > Tim:
    > Since the release of the SIR report it is clear that the autopilot is
    > OFF after IGARI. […] To make the assumption that the autopilot is
    > re-engaged at some time is not necessarily logical.

    It’s very logical if the intention was a fast getaway requiring A/P+autothrust OFF to bypass overspeed protections. Once out of danger and radar range, A/P could be used again.

    > TBill:
    > The 39S story is the active pilot got lazy for 6 hours and did a
    > straight flight southwest, woke up, and glided the aircraft 120nm
    > off the prior search zone. I see no logic or unique fit to the data,
    > unless we argue for a sleeping pilot.

    Sleeping pilot is the least likely possibility.
    There are plenty of more likely ones, maximizing range being one of them.

  143. Don Thompson says:

    @Peter Norton and “maximizing range

    If one had reached the middle of nowhere at fuel exhaustion, why go any further?

    The Geelvinck Fracture zone reaches up to S36.5º, E89.4º, there is little to differentiate further south. The depths across the surrounding area in the Australian-Antarctic Basin don’t make for an easier seabed search (if there was any prior consideration that localisation and a search would be feasible).

  144. Peter Norton says:

    A scenario I would like to run by you:

    • Maybe the pilot didn’t know about the Inmarsat data at all.
    • So he just shut down ACARS (before 17:37) + disabled XPNDR.
    • SDU was not shut down intentionally but as an unknown byproduct of shutting down other systems to maximize performance for a fast getaway
    • once out of danger and radar range at 18:25, the extra speed was not needed anymore
    • therefore systems required for A/P were enabled again

    • explains loss of SDU (due to unorthodox max performance config)
    • explains manual flight after IGARI (to bypass overspeed protections)
    • explains SDU re-logon at 18:25 (systems required for A/P were enabled)
    • explains absence of ACARS messages after 18:25 (ACARS disabled intentionally before 17:37)

    thoughts ?

  145. Peter Norton says:

    @Victor Iannello, @Richard Godfrey, @Don Thompson:
    Many thanks for the fuel range and search area limits. That was very helpful.

    @Don Thompson:
    idem for EoF article reminder. In the first Google Earth picture, I see 2 out 9 (can’t find the cyan path) paths touching down on the inside and 7/9 outside. So I guess the problem is that although crashes outside the arc are more likely, crashes inside the arc cannot be excluded and unfortunately can occur at the same distance from the arc, so both sides must be searched to equal width. Is that the correct interpretation ?

  146. DrB says:

    For those interested in lessons learned in the ARA San Juan search, here is an updated map showing the locations of interest:

    https://drive.google.com/file/d/1aVPAVl2CNCiVc35uhR2oa6qTP4uHdy8D/view?usp=sharing

    1. The best predictor of the actual debris field location is the planned course at the most recently observed average speed until about an hour before the implosion event occurred, which was detected by two CTBTO stations and two land-based seismographs in Argentina.
    2. The actual debris field is outside the CTBTO error ellipse by a factor of 1.8X.
    3. The debris field is only about three miles off the planned course in the direction expected from the local wind/waves/current based on GDAS for that time and place.

  147. Andrew says:

    @Mick Gilbert

    RE: Your post I’m still trying to hunt down exactly what CFFORCE_PITCHCWS FOREIGN and LOCAL actually represent and what the unit of measurement is. What I’ve inferred is that 100 = Herculean, 0 = easy peasy and that FOREIGN and LOCAL probably refer to pilot applied forces and control column back pressure although that’s all conjecture, right down to which is which.”

    Bjorn Fehrm’s analysis of the Preliminary Report discusses the stick force data and describes the difference between the two traces:
    Indonesian authorities release preliminary Lion Air crash report

  148. Mick Gilbert says:

    @Andrew

    Outstanding! Thanks for that.

  149. Peter Norton says:

    re: Hardy’s area

    Since it was mentioned a couple of times that Hardy’s area is beyond fuel range, in order to avoid misunderstandings let me say just in case (although you probably know) that Hardy’s area is designed to be reached in a glide after fuel exhaustion.

    Crossover of the 7th arc was not at S39 but circa S38.6 judging from the screenshot which is pretty well aligned with Barry Martin’s maximum range of S38.5.

    Besides, would the area beyond S38.5 (to S39.4 – S40) have been searched if experts had deemed it absolutely certainly unreachable ?

    BTW, I don’t know if Cpt. Hardy’s numbers are precise calculations or only GE approximations, so the true path might be slightly off.

    Victor Iannello: “It is not trivial to determine the farthest point south that is reachable from fuel considerations because it depends on how the plane was flown between 18:22 and 19:41. Because a path at LRC speed does not have sufficient fuel endurance, significant portions of the flight have to be at lower speed, which may include some combination of a descent and slow down to holding speed before 19:41, or slower than LRC speeds after 19:41. It becomes more difficult to find an acceptable combination of speed/altitude changes for latitudes farther south than around 35S because the speed after 19:41 needs to be greater than LRC to satisfy the BTOs, and there is not enough time for a descent and/or slow down before 19:41 to balance the fuel required after 19:41.”

    If I understand the video demonstration [part 1/3] correctly, FMT occured at 18:36 at or circa ANOKO.

  150. Peter Norton says:

    (continued)
    According to Cpt. Hardy:
    FMT occured at 18:36 at or circa ANOKO and then the plane continued at 488kn on a straight path T188° until reaching the area shown in the screenshot above.

  151. Victor Iannello says:

    @Peter Norton said: Crossover of the 7th arc was not at S39 but circa S38.6 judging from the screenshot which is pretty well aligned with Barry Martin’s maximum range of S38.5.

    I really don’t think Barry M’s model predicts that, when considering fuel endurance, timing of the FMT, and adhering to BTO constraints. If I understand things correctly, Richard G describes a path that turns to 180T at 18:25 with the fuel running dry around 00:05. That path won’t satisfy the aforementioned criteria.

    I have never understood the fascination with Hardy’s analysis. It is an oversimplification that ignores most of the BTO data, meteorological conditions, and fuel endurance.

  152. Victor Iannello says:

    @Peter Norton said: There are plenty of more likely ones, maximizing range being one of them.

    Please explain why the pilot did not cut across Sumatra and then turn towards the SIO, which would have put him much further south at fuel exhaustion. The turn to the northwest over the Malacca Strait before the turn to the SIO wasted a lot of range to the south.

  153. Don Thompson says:

    @PN

    One item, “explains SDU re-logon at 18:25 (systems required for A/P were enabled)

    That the L (and possibly R) Main AC Bus were isolated would not affect operation of the AFDS (autopilot flight director system).

  154. Peter Norton says:

    > Victor Iannello:
    > I have always been fond of 34.3S because of the simplicity of the
    > BEDAX-SouthPole great circle path and because the BTO and BFO fits
    > are excellent (which is why Inmarsat chose a very similar path).

    I agree that this scenario is a good fit.

    > However, if there were pilot inputs after 00:19, there is no reason
    > to assume there were not pilot inputs between 19:41 and 00:19

    after 00:19 = pilot inputs required for stable flight
    19:41 – 00:19 = no pilot inputs required for stable flight

    This would be one reason, no ?

  155. Peter Norton says:

    > Victor Iannello: “Please explain why the pilot did not cut across Sumatra”

    Sorry, I don’t have better explanations than those already discussed:

    1. risk of interception over Sumatra
    2. disguise of the target region (crossing Sumatra shows SIO as target, whereas laying a false trail up the Malacca Strait within radar range at 18:22 points to Europe/Asia as approx. target).

    Many explanations were discussed, maybe there were other good ones.

    It sure would be interesting to know why Sumatra was avoided, but the fact is that this is what happens, so we have to work with this fact anyway.

    There could have been 3 distinct phases of the flight, each with different priorities:

    phase 1 [KLIA-IGARI]: priority = making sure everything looks normal until deviation
    phase 2 [IGARI-18:25]: priority = fast getaway + reducing risk of interception + laying false trail
    phase 3 [18:25-EoF]: priority = maximizing range

    If it were not for the unknown Inmarsat data, the flight through the Malacca Strait would have worked perfectly as a red herring.

    Maximizing range could have been an overall priority but had to take a back seat in phase 2 where other things were more important.

  156. Viking says:

    @airlandseaman

    Concerning the weather impact of satellite communication, I do not claim that there should be transmission interruption or measurably increased bit error rate. All I say is that there may very likely be a power attenuation around 2dB near a tropical thunder cloud due to the plasma effect of upward going lightning.

  157. Ventus45 says:

    @Peter Norton

    Re Victor’s “Please explain why the pilot did not cut across Sumatra” and you reply “reducing risk of interception”.
    (a) Please explain “from whom” (RMAF or TNI-AU) or both, and
    (b) Your logic for either – or both, and detail or describe the relative risks you perceive.

  158. Peter Norton says:

    Don Thompson: “That the L (and possibly R) Main AC Bus were isolated would not affect operation of the AFDS (autopilot flight director system).”

    Thank you. These are separate points in my scenario:
    • Autopilot/Autothrust would have been OFF to bypass overspeed protections.
    • SDU would have been an unknown side-effect of shutting down other systems to maximize performance (imagined or real), as discussed above by TBill for example.

  159. Peter Norton says:

    @ventus45:
    1. re interception: The question is not what is safer but what the pilot, who wasn’t a military expert, considered safer.

    2. Flying over water along an international airway can arguably be considered to pose less risk of interception than invading sovereign airspace.

    3. If you think that avoiding Sumatra for fear of interception is not a good argument, then we still have the rationale of laying a false trail as a red herring. This is reason enough.

    4. It doesn’t really matter. We know where MH370 was at 18:22. It didn’t cross Sumatra but flew around it. This is a fact.

  160. Viking says:

    @Richard Godfrey

    Thanks for including hurricane Gillian in the simulation. The effect goes in the correct direction, but is roughly an order of magnitude too small to explain the deviation.

    However, when simulating the impact of a hurricane I presume the effect is critically dependent on the exact position of the debris when it hits, and on how it is floating in the water. If it is floating very high in the water (as one may expect early after the crash where the water has not yet penetrated the finer pores of the composite material and no biofouling is present), the hurricane will have much higher impact. In extreme cases it may even jump across the tops of the waves in the same way as when one throws a flat stone in the water at the correct angle.

  161. Viking says:

    @Richard Godfrey

    Thanks for informing me on your reference point for the 26.1Hz BFO deviation. When I do a crude estimate with the same reference I get slightly over 24Hz so we are very close. I am sure your estimate is most precise, since I used an approximation which will lead to this level of deviation.

  162. mash says:

    re: Making Sense of No Sense: the “Dive-glide-dive” Paradox

    (Paradox) Reasoning:
    By assuming the dive-glide-dive scenario is true, almost everything is not explainable/reasonable; therefore, the dive-glide-dive scenario fits well with the other on-going ‘self-contradictory’ patterns – and so it should be seriously considered as a possible/probable scenario. [[think imaginary (vs real) number]]

    [The problem is, perhaps, no one is willing to consider/believe a ‘third’ (but unknown) alternative; therefore, so-called ‘simple’ solution’s explanations (of the observed events) must be very ‘complicated’.] [[think heliocentric (vs geocentric) system]]

    [Contradictory Patterns (end to begin – roughly – just to bring home the idea)
    dive – why glide?
    glide – why dive?
    glide – why no IFE (except APU down ‘coincidence’)?
    ditch planned – why fuel exhaustion?
    EOF pilot active – why (likely) ‘passive’ flight after FMT?
    Satcom on again – why phone calls not answered?
    Satcom off – why on again?
    SDU reactivated – why deactivate IFE (if no ‘coincidence’)?
    Negotiation – why no ‘direct’ communication?
    Depressurisation – why mobile phone (re-)connection detected?
    Negotiation – why depressurisation?
    Etc.]

    [Based on above ‘analysis’, should dive-glide-dive be considered seriously? Perhaps so: at least it is ‘consistent’; and as long as one is still ‘ignorant’ (of the “centre of origin of the system”).]

    [[Actually (imagining above ‘logic’ acceptable), as a high risk venture, now is also the best time to reconsider a particular (optimal) ‘point’/’hotspot’ (centre of a [larger] circle) again – still assuming a pre-programmed flight but a “contradictory” dive-glide-dive scenario.]]

  163. Peter Norton says:

    ———
    Victor Iannello: First, assuming no descent, the turn south had to occur between 18:28 and 18:40, not at 18:25, which is why it appears that 38.5S was reached. But looking at the fuel endurance results, and taking the results at face value (always dangerous without knowing the details), each 1% of PDA reduces the endurance by 4 mins. For a PDA of 1.5%, there would be 2 min more of endurance than 2%, i.e., fuel exhaustion of 00:05. That’s about 10 minutes less than observed.

    Peter Norton: Crossover of the 7th arc was not at S39 but circa S38.6 judging from the screenshot which is pretty well aligned with Barry Martin’s maximum range of S38.5.

    Victor Iannello: I really don’t think Barry M’s model predicts that, when considering fuel endurance, timing of the FMT, and adhering to BTO constraints. If I understand things correctly, Richard G describes a path that turns to 180T at 18:25 with the fuel running dry around 00:05. That path won’t satisfy the aforementioned criteria.
    ———-

    ok, my counter-arguments:

    • The model doesn’t include all the aforementioned range-extending parameters. I mean, it doesn’t deal with bleed air for example, right ?
    • I assume the model is geared towards accuracy and thus doesn’t follow a conservative approach (safety margins).
    • As far as I can see, Cpt. Hardy’s model doesn’t consider wind. This could move the real path further north.

  164. Ventus45 says:

    @Peter Norton.

    Then we will simply have to agree to disagree, on the 18:22 radar hit, and all that it entails.

    For Z to elect to fly what are essentially two sides of a triangle, when he could simply, and in half the time, fly the third side direct, (if getting as far south as possible was the goal), defies any sensible logic, unless he thought the risk of interception by TNI-AU was unacceptable.

    Even then, quite frankly, given the location of their radar sites, and their fighter bases, and even if the TNI-AU were “on alert”, crossing Sumatra quickly and getting “out of range”, in a tail-chase as it were, is far preferable, to going up around Ache, and then coming back down the west coast of Sumatra, which would give the TNI-AU time to scramble, and position the fighters off the west coast, indeed, they could be there waiting for him, by the time he got there. Going around Ache was definitely not a smart move. Going straight across Sumatra, like a bat out of hell, was the smartest option, by far.

  165. airlandseaman says:

    Viking: I assume you are referring to variations due to ionospheric scintillation, not thunderstorms. The rain attenuation due to a strong thunderstorm is <0.1 dB. In any event, it did not happen. Here are the measured values:

    http://bit.ly/2AQYqJ7

    This graph demonstrates that the inbound rcv pwr at Perth was essentially constant for ARC1 -ARC7 (within normal variations due to antenna pointing, etc.)

  166. Victor Iannello says:

    @Peter Norton: Some comments:

    1) I think that your “false trail” up the Malacca Strait is possible, especially if the turn to the south occurred after 18:40 and beyond the (perceived) range of Indonesian radar in Sumatra.

    2) To defeat the envelope protection, it is not enough to disengage the A/P and A/T. Rather, the flight control mode has to transition out of NORMAL. Two ways for this are to remove power from the transfer busses (by isolating the AC busses and turning off the IDGs and backup generators, for instance), or to disconnect the PFCS, either of which would degrade the flight control mode to SECONDARY.

    3) You can’t first suggest that air packs were turned off to extend range, and then claim the plane was actively piloted after fuel exhaustion.

    4) When the effect of temperature corrections and engine PDA are properly incorporated in the fuel model, I do not believe there is the range and/or endurance to reach 38.5S latitude for a path that satisfies the BTO data.

    5) As mentioned previously, Hardy’s model does not properly account for meteorological conditions, autothrottle modes, the entire set of BTO data, and fuel consumption. It is overly simplistic and inaccurate.

  167. Victor Iannello says:

    @airlandseaman: Correct me if I am wrong, but there are other effects beyond antenna pointing that cause the signal strength to vary: 1) At the log-on at 18:25 and 00:19, there is an adjustment of power between the log-on request and acknowledgement based on information from the GES. 2) The gain of the satellite’s “bent antenna” varies over time as a function of the total traffic, i.e., higher total traffic results in less power per channel.

  168. Victor Iannello says:

    @mash: Unfortunately, your comment is both absurd and perversely sensible. Said another way, it is very difficult to come up with ANY scenario that checks all the boxes. However, I don’t think we should give up trying.

  169. Victor Iannello says:

    @Viking: Is your stance that a path with 26 Hz BFO error is nonetheless acceptable?

  170. David says:

    @Mick Gilbert. The NTSC lists the Tianjin Binhai to Manado flight as being on 26th. If as per your list that was on 27th and supposing also that their 27th listing of two others has them reversed then with the Denpasar-Lombok return flights added to that day that increases the consecutive flights with the Speed and Altitude flag evident to at least 5, not the NTSC’s 3.

    Also there would have been 4 consecutive flights with the Speed Trim Fail light illuminated vice 2.

    That said, the ‘Speed Trim Fail’ defect terminology might be misleading to ground and aircrew as to the immediacy and thoroughness of the attention it should get having subsumed MCAS failure effects. In other words a pre-MCAS ‘Speed Trim Fail’ might have been more benign. The MCAS stabiliser nose down trim rate can be 2.5 deg in a 10 sec burst. That rate first caused dives before trim was selected off (28th flight) and flaps were lowered (29th). Such a dive could be from a very low altitude I take it.

    @Andrew. Thank you for that reference. Bjorn Behm comments that the pilot of the flight to Denpasar did not report the trim isolation necessary because “…this is no classical trim runaway. He consequently didn’t mention “Trim Runaway” in his post-flight reports”. I would have thought that use of the trim switch to overcome excessive stick forces, which had led to a dive, might be worth mentioning whatever it was called.

    The NTSC said he should not have flown on with stick-shaker active, the aircraft being unairworthy. He made no mention of that either, written or oral.

  171. Peter Norton says:

    > Don Thompson:
    > re: /maximizing range/
    > If one had reached the middle of nowhere at fuel exhaustion, why go any further?

    Why not ?
    If the intent was to maximize range, why not maximize range ?
    I don’t see your point.

    > The Geelvinck Fracture zone reaches up to S36.5º,
    > E89.4º, there is little to differentiate further south.

    I can only speculate. Either the seafloor topography/geology was not considered at all and is just a coincidence.

    Otherwise it might have been just an ancillary bonus somewhere along those lines:
    Pilot wanted to maximize range after 18:22 (as explained above), made similarly calculations on Google Earth as Captain Hardy and found the same target area. And either the fracture zone was visible on GE, or he checked the coordinates in a seafloor map. In this case, it may not have been the primary goal (which was maximizing range) but within that target area the fracture zone was chosen as an ancillary/secondary goal, given that it might complicate undersea detection. (Based on the sonar images I think this is true.)

    re: criticism of Hardy’s method

    > Don Thompson:
    > the seabed in the area that he first advocated was searched
    > by Fugro. ATSB gave his work some consideration, I assume they bounded
    > it with their own conclusion for a maximum distance, wide of the arc, to
    > ocean impact. The area was searched out to +90km beyond the 7th arc.
    > Hardy persists that his thesis is correct so moved his ‘X’ further along
    > his ‘constant speed line’, i.e., at main engine fuel exhaustion the
    > aircraft altitude was near cruise level, therefore, it ‘glided’ further.

    Sorry, I have no idea what you are referring to. I don’t know of any goalposts being moved, and even if he did, it wouldn’t bother me. Everyone refines their models. Cpt Hardy wouldn’t have the right to revise his target by a few nm, but the IG can revise by 100 times as much? Sorry, I fail to see your point. And I really don’t want to focus on that. I only know the 3 part video series. So I want to focus on that. (No goalposts are moved there either.)

    And no, Don, the area shown in the video cannot have been exhaustively searched, because it extends up to 120nm from the 7th arc, whereas you say that the search area only covered +90km beyond the arc. So based on that alone, there are 30km missing (regardless of the arc segment, which might not match either).

    > Hardy’s method considers a constant ground speed.

    I don’t know why. I assume for the sake of simplicity with the youtube video just being a proof of concept?

    I agree that wind has to be considered. But I think this just moves the path a bit NE (rather than completely invalidating his method).

    > Considering winds for airspeed would revise
    > the arc crossing point north-east wards.

    Well, all the better, no? This makes the Hardy area even more likely in terms of fuel range.

    > Tools: Hardy describes using Google Earth as a tool to determine his
    > paths. Google Earth calculates paths between two points as great
    > circles, so are only accurate if the start and end point are valid. An
    > autopilot only navigates great circle paths when it’s receiving LNAV
    > guidance from the FMS to a waypoint. This approach is inaccurate.

    Same answer: I think there are a couple of ways where his method is just coarse, I assume for demonstration purposes on youtube. Another case in point: Hardy clicks ~approximately where the 3 lines intersect and then lines up a straight path at a certain angle up to ~ANOKO. It’s approximate geometry as a proof of concept, not precise calculation. Rather than rejecting his model, it should be refined.

    > Aircraft knowledge: Hardy appears to assume that a Boeing 777 will fly
    > serenely onwards without engine power […]

    To my knowledge he never once in his 3-part video states whether the flight is actively piloted or not.

    > I do know that Hardy described testing a dual engine failure scenario in his
    > airline’s Level D simulator.

    I can’t speak to that. I don’t know the parameters of his scenario. Given your experience, I can imagine that your scenarios model the reality better than his.

    > Hardy tested a dual engine failure scenario in his airline’s Level D
    > simulator finding that the aircraft, hands off, continued to hold
    > heading with wings level and it began a stable descent without upset.
    > That conversation was notable as it described a scenario contrary to
    > the Boeing simulator trials and the exercise conducted by ALSM.
    > However, during the 60 Minutes MH370: The Situation Room programme
    > (May 2018) Hardy demonstrated a scenario in which the aircraft had
    > entered a steep descent from 35000ft and, at reaching -9000fpm, he
    > then recovered to controlled flight at approx 29000ft (PFD read off
    > screen). No explanation was given for the conditions in which the
    > descent had initiated, nor why he delayed his intervention. It would
    > be useful for Hardy to explain his findings and present conclusion
    > for the 777’s flight characteristics after fuel exhaustion.

    Unfortunately I don’t know more about the 60min scenario than you. I agree it would be useful for Hardy to explain the different simulations if he has not done so.

    > Evidence: most significantly, he ignores the BFO metadata recorded
    > during the 00:19 Log On. He accepts the BFO analysis that determined the
    > aircraft flew south, but not that the records at 00:19 indicate descent
    > 13,800fpm (per Holland).

    Yes, although all the explanations/theories that legitimate a dive-glide scenario at any other point at the 7th arc would obviously also be valid in this case. I mean, if one is willing to search wider, for example on DrB’s path or Victor Iannello’s or Inmarsat’s southpole path, Hardy’s area would be equally legitimate from the point of view of a glide scenario.

    The other possibility is that the BFO value was produced by another phenomenon than a dive. Don’t ask me why. I have no idea. I’m not saying this is likely. I just mention that it was discussed.

  172. DennisW says:

    @Peter

    I mean, if one is willing to search wider, for example on DrB’s path or Victor Iannello’s or Inmarsat’s southpole path, Hardy’s area would be equally legitimate from the point of view of a glide scenario.

    No one is going to search anywhere if they listen to your ramblings.

  173. Peter Norton says:

    Victor Iannello: “I have never understood the fascination with Hardy’s analysis. It is an oversimplification that ignores most of the BTO data, meteorological conditions, and fuel endurance.”

    Ok, first regarding the attraction:

    In his model
    (1) his maximum fuel range circle from ANOKO at 18:36
    (2) the 7th arc
    (3) a straight track at T°188 flown at a constant 488kn
    all intersect at the exact same spot.

    My question to all mathematicians here is:
    Is line (3) exceptional ?
    Or would any path, even random actively piloted curved paths, create BTOs that allow for a (wrong) straight line interpretation ?
    If no, how likely is such a random coincidence ?
    Is it possible to assign a probability from 0% to 100% ?

    JW once asked something along those lines: If you take a seven hour walk in a city, randomly turning right and left at crossroads, what are the odds, that your locations at full hours all lie on a straight line by mere coincidence ?

    I don’t know if he has a point there, but this – combined with the 3 lines intersection – is why I find Hardy’s method attractive.

    As for your 3 points of criticism:
    • BTO data: To my knowledge all BTO values are accounted for. Or do you mean BFO?
    • fuel endurance: as discussed above, I think this might just be the flight path geared towards maximum range
    • wind: Yes, but this only moves the path a bit more to the NE and makes the scenario even more likely in terms of fuel range.

  174. airlandseaman says:

    Victor: Yes, there are several factors affecting the receive power. I gave the AES HGA gain vs. pointing angle as one example. The AES HPA is power agile, so the L band uplink EIRP can be (and is) adjusted occasionally to maintain the nominal rcv pwr at the S/C L band rcv antenna. It is changed by the SDU depending on the data rate to maintain a nominal Eb/N0 at the GES demod.

    The I3F1 satellite uses fixed gain transponders. There is no AGC, but the transponder gain can be adjusted by ground command as needed. Thus, the instantaneous traffic loading does not affect the transponder gain on a minute by minute basis.

    I have been over the unredacted Inmarsat log with a fine tooth comb and I see no evidence of signal drop out at any time attributable to weather. There is evidence of low C/N0 in a couple of cases, but it was due to co-channel interference, not low signal strength. That happened at 18:25:27 for example.

  175. Perfect Storm says:

    @airlandseaman: your inbound Rx power graph is interesting. Can anything be inferred from it concerning MH370’s flight path?

  176. DrB says:

    @Victor Iannello.

    You said (to Peter Norton): “You can’t first suggest that air packs were turned off to extend range, and then claim the plane was actively piloted after fuel exhaustion.

    Maybe in fact one could say that. I seem to recall that the supplemental O2 system was good for quite a few hours for one pilot, although I don’t know how long the pilot could keep up when breathing in a low ambient pressure. In addition, one could shut off bleed air for one or several hours (say from 17:23-19:00), and then repressurize the aircraft. That would save a bit of fuel, although I don’t think squeezing out a few more miles was part of the plan.

    On the subject of why fly out the Malacca Strait, my own view is that this was planned as a false trail. The subsequent turn to the south occurred far enough to the west that it was not captured by anybody’s radar. This combination of maneuvers seems to me to be an excellent plan to disappear the aircraft. If so, achieving maximum range from Malaysia is not the consideration. It is instead a plan to achieve an undetected and major course change and then to go as far as fuel allows into the middle of the SIO. I doubt he navigated to any underwater geological feature or far-south waypoint (that might be guessed as a destination). Except for his ignorance of the presence of the BTOs, it was a masterful plan executed with precision.

  177. Peter Norton says:

    Ventus45: “Going straight across Sumatra, like a bat out of hell, was the smartest option, by far.”

    Then everyone would have known MH370 is in the SIO.

  178. airlandseaman says:

    Perfect Storm: Nothing about the path can be inferred from the RCV PWR, or the C/N0 or BER. What all these parameters show is normal operation all the way to the very last record at 00:19:37 when the BFO indicated a descent rate of 15,000 ft/min. Note that the next to last R1200 record at 00:10, prior to FE, was within 0.5dB of the final R1200 record at 00:19:37.

  179. David says:

    @Peter Norton. Range to f.e.
    Another contributor, energy height.
    If that will power a glide 120nm then theoretically a powered glide would use the energy in the same way even if not gaining all that 120nm.
    At lowish level a bunt down after that extended f.e., directing the velocity vector down, would satisfy BFOs.

  180. Perfect Storm says:

    airlandseaman: “the next to last R1200 record at 00:10, prior to FE, was within 0.5dB of the final R1200 record at 00:19:37.”

    Ok, what does that tell us ?

  181. airlandseaman says:

    David: What is “energy height”? Actually, I can’t understand anything you wrote above. Please try again.

  182. airlandseaman says:

    Perfect Storm: It suggests no radical attitude (like inverted) at 00:19:37. Could be in a high bank angle though.

  183. Peter Norton says:

    DrB: “On the subject of why fly out the Malacca Strait, my own view is that this was planned as a false trail. The subsequent turn to the south occurred far enough to the west that it was not captured by anybody’s radar. This combination of maneuvers seems to me to be an excellent plan to disappear the aircraft. If so, achieving maximum range from Malaysia is not the consideration. It is instead a plan to achieve an undetected and major course change and then to go as far as fuel allows into the middle of the SIO. I doubt he navigated to any underwater geological feature or far-south waypoint (that might be guessed as a destination). Except for his ignorance of the presence of the BTOs, it was a masterful plan executed with precision.”

    This is exactly what I tried to convey in my comment above. You presented this idea way more eloquently. I agree this is one of the more plausible scenarios and explains quite a few observations.

    Yes, as you say waypoints add an unnecessary risk. The only point I would differ is the underwater topography. I don’t know how much data was available to him, but if the goal was to make the plane disappear by flying “as far as fuel allows into the middle of the SIO”, then why not choose difficult terrain? It would be a bonus and contribute to the same goal.

  184. Peter Norton says:

    @Victor Iannello: In reply to your comment above:

    (1) re: Malacca Strait. Yes, good point. I always thought that the LKP@18:22 and the SDU-relogon@18:25 was too closely timed (particularly when you deduct the start-up time) to be a coincidence. But maybe it really was just a coincidence and the FMT only occurred after leaving the presumed radar range at 18:40.

    2) re: defeating the envelope protection. “remove power from the transfer busses (by isolating the AC busses and turning off the IDGs and backup generators, for instance)” Wouldn’t this also knock-out the SDU as a side-effect? (This would be in line with my scenario above explaining why the SDU lost/regained power).

    3) I am unfamiliar with that topic. I’ve just seen lots of times here on your blog reduction or disabling of bleed air being mentioned as an idea to increase range/performance. Using oxygen might be a possibility as DrB points out.

    4) Then why was the search area extended to S39.3+ ? I concede that many here have invested considerable efforts into the fuel consumption models and I don’t intend to disrespect them. I just argue that the maximum range comes relatively close to Hardy’s path and when all the aforementioned range-extending parameters are considered OR when wind is accounted for (see above), then I think the path can work.

    5) I have addressed these points in my replies to you and Don Thompson.

  185. TBill says:

    @Peter
    “Is line (3) exceptional ?”

    I think it just means if you assume a final turn FMT south at ANOKO at 18:40, and a passive flight on a straight-line course, you get Hardy’s end point.

    The easterly extreme is Victor and Richard’s NZPG course, they assume later turn and VOCX for FMT, and you get “straight” line course to approx. 26.9 South@Arc7, which would be where the fuel exhaustion circle would also intersect Arc7.

    Those two cases map out the west/east boundaries of the passive flight paths suggested to date. I believe Rob would say the ultimate clincher for Hardy’s path is that sun rises at that time, which is valid hypothesis, but not convincing for many of us (that an active pilot would have crash exactly at sunrise as the main strategy for the flight).

  186. David says:

    @ALSM. “What is “energy height”? Actually, I can’t understand anything you wrote above. Please try again.”

    Otherwise known as potential energy. As a crude example, if at 40,000 ft when almost out of fuel, nigh-on at the 7th arc the pilot retarded the engine(s) to idle and used his height to maintain ground speed, fuel exhaustion following later when low down, he would have gone further than staying under power a few moments more at altitude.
    If you say that would carry his 7th arc BTO further than it was, well not necessarily. If he had turned right at the 7th arc (remember this is a crude example) then travelled along the 7th arc to fuel exhaustion, that would have extended his crash latitude south.

    In a more refined example that principle would be integrated with speeds, courses, BTO’s and BFO’s in the same way as step climbs would need to be.

    I support those looking for alternatives to the offerings north or wider offerings not because there must be something better but because there might be, still. Getting out of the unpiloted straight jacket may be as yet underdeveloped.

  187. David says:

    farther vice further!

  188. Warren Platts says:

    @DennisW

    1. Our posts crossed. More to say later.

    2. My guess is that you are living with your parents. Am I right?

    Dude. Rilly? FYI since you are so curious about my personal life, both of my parents are recently deceased.

  189. Peter Norton says:

    TBill: re: Is line (3) exceptional?: “if you assume a final turn FMT south at ANOKO at 18:40, and a passive flight on a straight-line course, you get Hardy’s end point. The easterly extreme is Victor and Richard’s NZPG course, they assume later turn and VOCX for FMT, and you get “straight” line course to approx. 26.9 South@Arc7, which would be where the fuel exhaustion circle would also intersect Arc7. Those 2 cases map out the west/east boundaries of the passive flight paths suggested to date.”

    Yes, that’s what I meant. How likely is it that an actively piloted path with random turns has straight path “cousins” (Hardy or NZPG)? Is this highly likely or highly unlikely?

    TBill: “I believe Rob would say the ultimate clincher for Hardy’s path is that sun rises at that time, which is valid hypothesis, but not convincing for many of us (that an active pilot would have crash exactly at sunrise as the main strategy for the flight).”

    When I first came up with the sunrise theory in 2015 my idea was that the pilot could have aimed for a spot that is (a) as remote as possible and (b) in daylight for a controlled ditching. This was at a time when everyone still wondered where the heck MH370’s debris is. 3 months later the flaperon was found. Given the debris we have today, if a ditching was attempted, it seems to have gone wrong.

  190. Mick Gilbert says:

    @David

    Re: ‘The NTSC lists the Tianjin Binhai to Manado flight as being on 26th.

    Date/time entries in the AFML are UTC, the dates that I have provided are local. JT2748 Tianjin – Manado operates 2355 CST – 0550 (+1) WITA, both being UTC + 8.

  191. David says:

    @Mick Gilbert. Thanks. So I take it that confirms the defects were carried for more flights than the NTSC has stated.
    Even so I doubt that will be raised in the final report, not being directly related to the accident.

  192. Peter Norton says:

    comment from AvHerald re LIN610: “ATC: As a crew, I would have told them to shut up. I said pan pan, you move the other traffic away and stop bothering me. Reading the communication record appalls me.”

  193. Richard Godfrey says:

    @Victor, @Peter Norton

    I am not an expert using Barry Martin’s Path Model and was only trying to help answer the question, what is the theoretical maximum latitude that could be reached by MH370, following an early FMT.

    The answer to that question is around 38.5°S, making assumptions that the altitude was constant at 35,000 feet, LRC speed mode was selected and the PDA was 1.5%. However, as Victor points out the endurance does not fit, with fuel exhaustion occurring at 00:05:30 UTC, about 10 minutes too early.

    To be clear, my personal view is that there was no early FMT. I think there was a late FMT following a holding pattern.

    I also think Simon Hardy’s path model is grossly over simplified, erroneous and misleading.

  194. Richard Godfrey says:

    @Viking

    Regarding your comments on my path model, a BFO Error of 26 Hz is unacceptable.

    Regarding your comments on my drift model, please note that Tropical Cyclone Gillian passed 80 NM in front of the simulated debris path from your end point.

    At 05:30 Local Time on 23rd March 2014, the centre of Tropical Cyclone Gillian was at 11.8°S 104.25°E. The simulated debris was at 11.7°S 105.6°E at this time.

    I used the actual GDAS wind strength and direction data at the time and location to calculate the simulated debris track. The sustained wind strength was only 32 knots at a distance of 80 NM from the Tropical Cyclone Gillian centre. This aligns to the fact that the 64 knot wind radius over open water, was 35-40 NM according to the NOAA Storm Centre 12-hourly reports at the time.

  195. Mick Gilbert says:

    @Peter Norton

    Re:’LIN610: “ATC: As a crew, I would have told them to shut up. I said pan pan, …’

    JT610 didn’t make a PAN PAN call, while they did advise ATC of a flight control problem they didn’t declare an emergency of any sort.

  196. Viking says:

    @airlandseaman

    It is my impression that the plasma contribution from the atmosphere is partly due to solar induced effects and partly due to upwards lightning. I think the details are still being researched and not fully understood.

    The link to Perth will not be affected by thunderstorms near Indonesia.

  197. Viking says:

    @Victor, @airlandseaman

    Concerning the antenna contribution and its dependence on traffic, I would presume traffic goes up in the morning hours in Asia (more than it goes down near Africa at the same time). This would promote a more rapid drop in power towards the 7th arc. I am not sure how big this effect is, but at least it strengthens my point on the power curve.

  198. Viking says:

    @Victor, @Richard

    I agree that 26Hz deviation is not acceptable. However, if one assumes that there was a pressure drop in the cabin some time between 20:41 and 21:41 this gives a contribution with the opposite sign and of similar size.

  199. Victor Iannello says:

    @Peter Norton: If you go back to a previous article, I examined paths constrained by the BTO and BFO data that were great circles starting with 19:41. Hardy’s path falls in that category, except many effects that he ignored are included in the analysis. Included in that post was a CSV file that summarized the results assuming LRC and constant Mach autothrottle modes. To reach 39.5S latitude requires a speed of M0.84 at FL350. That’s faster than LRC. Despite what you keep saying, by a large margin, there was not enough fuel to reach 39.5S latitude.

  200. Viking says:

    @Richard

    Concerning the critical sensitivity of the debris drift to distance from the hurricane, I think your numbers confirm this. If the debris was just 40nm further west it would have experienced double wind speed. Slightly further west the push from the hurricane would have reached maximum.

  201. Don Thompson says:

    @ALSM

    I’ve added an annotation to your plot, with explanation, here.

    Cloud cover data from Japanese MTSAT2 is archived for 2018-03-07/08. Some cloud was evident along the west Sumatra coast, however, Christmas Island was clear. I don’t consider that the cloud cover had any effect on the signalling, contributing this info for information only.

  202. Warren Platts says:

    @Victor Ianello: Please explain why the pilot did not cut across Sumatra and then turn towards the SIO, which would have put him much further south at fuel exhaustion.

    @Tbill:>If we have an active pilot making all kinds of maneuvers up to 18:40, why does it suddenly make sense that all maneuvers would cease? BFSkinner said so? The 39S story is the active pilot got lazy for 6 hours and did a straight flight southwest, woke up, and glided the aircraft 120nm off the prior search zone. I see no logic or unique fit to the data, unless we argue for a sleeping pilot.

    The whole problem with finding MH370 is the assumption that straight, passive flight is somehow the more likely, more rigorous base case, and active pilot is somehow a weak speculative case. But passive, straight flight is perhaps the greatest speculation of all.

    @Ventus > tail-chase far preferable position the fighters not smart move bat out of hell smartest option

    @Don Thompson > If one had reached the middle of nowhere at fuel exhaustion, why go any further?

    People: this is exactly the sort of speculation we should avoid. It is a rabbit hole that leads nowhere. Instead, I would rather recommend a Dan Dennett-style intentional stance/Skinnerian behaviorist analysis that does not attempt any hypotheses regarding the mental states of the pilot. Indeed, I find it helpful to imagine that the aircraft was taken over by a gray alien from Zeta Reticuli or the Evil Tiger Spirit ala Stephen King’s Carrie. Why? Because it is impossible to try to get inside the mind of an alien or evil ghost with any hope of accuracy: hence there is no point in trying.

    What we do have is the behavior of the aircraft prior to the FMC. And it exhibited the following characteristics for the most part:

    1. High altitude
    2. Fast cruising speed
    3. The aircraft was in “LNAV” mode in the weak sense that it was apparently navigating using established waypoints
    4. There was some indication it was being hand flown
    5. There was a tendency to avoid directly crossing sovereign airspace

    Thus to answer Victor’s question, the turn up the Malacca Strait was simply a continuation of the behavior exhibited in the prior leg of the flight. Period. In the first leg, the aircraft straddled the border between Thailand and Malaysian airspace; in the Molacca Strait leg, the aircraft straddled Thai and Indonesian airspace. (Which is why I am skeptical of a late FMT all the way up by ANOKO or more, as that would require flirting with Indian airspace; a better continuation of the same behavior imho would an FMT consisting of two turns, one at NILAM or SANOB, followed by another at POVUS–this would also considerably relieve maximum range models.) That is all there is to it. There is no need to speculate about scrambling fighter jets–although it is certainly the case that airliners straying into the wrong airspace have been shot down before: cf. MH17, USS Vincennes, KAL700.

    As for “LNAV” mode, I am not saying the autopilot itself was necessarily engaged in LNAV mode, although it could have been. Rather, the aircraft could have been hand flown. The pilot in that case, however, was still navigating by aiming for waypoints: ABTOK at Khota Baru, ENDOR then OPOVI at Penang, followed by VAMPI, MEKAR and probably NILAM or maybe SANOB.

    I have Mr. Exner’s Khota Baru PSR paper, that if mostly correct, would provide evidence that not only was the aircraft being hand flown at that point, but also that the pilot was pushing the envelope of what is possible in a B777. Then there is that one radar point just after Pulau Perak that sticks out like a sore thumb. Now, either that is some sort of meaningless error outlier, or the pilot was hand flying because it would have required a 75+ degree bank that is way past the overbank protection–again indicating a pilot pushing the performance envelope of his aircraft. The significance of that is we should expect a continuation of the same behavior at the end of the flight. That is, we should expect there to be a glide at EOF merely because that would continue the earlier behavior of “pushing the envelope”.

    What about after the FMT? Well, to continue the same behavioral pattern of high, fast, waypoints, combined with the BTO lines of position, you are pretty much forced down one corridor: POVUS, ISBIX, MUTMI, RUNUT that are all on a 189T azimuth. If you follow that course to the 7th arc, it is virtually identical to Capt. Hardy’s 188T path, yet derived by entirely independent means. So really, there is a 4-way consilience here: 7th arc, maximum range line, straight-mostly-equal-speed line, plus a waypoint path that is consistent with earlier patterns of navigation.

    Yes, this pushes the limit in terms of what is possible fuel-wise, but as Peter Norton points out, that is feature, not a bug. Note also when I said 39S, I was just rounding up. Anybody who thinks they can predict the location to better than a degree of resolution is just fooling themselves tbqh. At any rate, 188T comes out at about 38.6S; the 189T course crosses the 7th arc at 38.7S. So if you tell me the absolute limit is 38.5S, you can’t tell me that 38.6S or 38.7S is beyond the error bar of that estimate and thus theoretically impossible. 39.0S would only require flying an extra 38 nm further than 38.5S a difference of 1.4% from the FMT. Judging from the previous search pattern, the searchers apparently considered a path seemingly aimed at 69S,69E to be their westward limit. That path crossed 7th arc at 39.1S. So no one here is proposing something that is completely off the wall, like crossing the line at RERAB (~43S–curiously, if you run a max range calculation on a course straight from IGARI itself, it will get you to RERAB just barely!) or even 39.5S–not sure where you got that idea Victor.

    As for Victor’s point that a crash at the far SW zone should have produced debris that washed ashore in east Africa sooner than was detected, we can know for sure that is in fact what happened! Any flotsam produced by the crash would have collected a lot of barnacles, and since all the pieces collected in east Africa were clean, there is no telling how long they were sitting on the beach. In fact, out of all 30+ pieces, only two can be relatively precisely timed: the flaperon, and the engine cowling bit. The latter is proof of what I’m talking about: a person took a picture of it when it was still covered in sea life, but did not pick it up; some months later someone else found it, and it was bone clean by then. Therefore, no precise dates can be given for the time of arrivals for any of the other debris items.

    As someone pointed out, the Hardy zone has been searched to a distance of 90 km, which is 50 nm. That would take into account a 60 nm glide on the same 188/189 course. Thus a further extension of 50 nm should catch a glide up to 120 nm. If it was extended from say 39S up to 34S (where Victor said was his favorite path from BEDAX to NZSP that also happens to correspond to the ATSB crash site proposed in 2017) that would be a search area about the same size as the 25S to 20S search area that is also being proposed. Big, but not unmanageably so.

    At any rate, I hope this little essay wasn’t tltr. If you made it this far, thanks for paying attention! 🙂

  203. Wall says:

    @all
    Between which latitudes have the authorities been looking for MH370?
    @Victor Ianello
    My first question was about your assumption/statement that the fourth possibility (south of 39.6S) can be ruled out. And I have to admit that the fuel consumption of the airplane is an extremely important part in finding out where to look. Moreover, drift models showed that there’s a high possibility that a significant amount of debris should have washed ashore on Australian beaches, which is not the case. According to your analysis, it was not possible for the plane to reach the far south areas. But if the lowest part of the search took place in the 39 area (39.6), then why did that happen. What was the reason for them to search that far for MH370?

  204. Wall says:

    @warren platts
    Such dedication!

  205. airlandseaman says:

    Don: Thanks Don for highlighting the reason for the logon sequence eirp changes.

    Viking: It seems you are conflating the transient luminous event (TLE) phenomenon (discovered here in Colorado and sometimes referred to as sprites and jets) or ionospheric scintillation with neutral atmosphere rain attenuation. Whatever it is you are talking about, it is not applicable to the signals transmitted by 9M-MRO on 2014-03-07 and received by the Perth GES. The log proves that. So why keep harping on this? Also, as I noted above, the traffic does not have any affect on the transponder gain. There is no AGC on the I3F1 transponders.

    Other readers: Don’t be distracted by this rabbit hole. Every singe record in the Inmarsat log has been analysed and rechecked and cross checked. We understand it in depth at this point. There is ZERO possibility that any BFO or BTO or RCV PWR or BER observations were materially affected by the neutral atmosphere weather, sprites, jets, or ionospheric scintillation.

  206. Victor Iannello says:

    @Warren Platts: Whether you understand it or not, you have speculated many times in your comment. By the way, I allow speculation as long as it is identified as such.

    You have my comment about drift exactly backwards. Further south on the arc would have taken longer to drift to East Africa than observed.

    As for fuel calculations, reaching 38.5S is not possible by a wide margin. Do the calculation for yourself if you don’t trust my numbers. Or, if you provide me with a path of Mach number, altitude, and temperature offset versus time, I’ll do the calculation for you.

    @Wall asked: But if the lowest part of the search took place in the 39 area (39.6), then why did that happen. What was the reason for them to search that far for MH370?

    At that point in time, the CSIRO drift model was still being refined, and the DSTG Bayesian analysis incorporated a very crude fuel model that was essentially useless in constraining southern paths. There was also a lot of pressure to search Hardy’s spot.

  207. Richard Godfrey says:

    @Warren

    You stated “The whole problem with finding MH370 is the assumption that straight, passive flight is somehow the more likely, more rigorous base case, and active pilot is somehow a weak speculative case. But passive, straight flight is perhaps the greatest speculation of all.”

    Complete and utter nonsense!

    The plot of the “handshake” Burst Frequency Offsets (BFOs) is quite smooth and nearly linear in the timeframe 19:41 UTC and 00:11 UTC. The plot of the “handshake” Burst Timing Offsets (BTOs) is quite smooth and fits a cubic polynomial in the timeframe 19:41 UTC and 00:11 UTC.

  208. Warren Platts says:

    @ Richard Godfrey:

    No no. That was Tbill who said that. My approach was to enter every waypoint in the Indian Ocean I could find into my google earth, and then find a path that met the three constraints of following waypoints, flying at cruising speeds, and being consistent with the BTOs. It just worked out that the path happened to be straight. That the path had to be straight was never a working assumption the way I did it.

    @Victor: yes of course everything I say is speculation. That goes without saying, but so is everything else. Take the drift models: these are based mainly on buoy data that are designed to measure surface ocean currents, not pieces of jetsam with a density not much more than styrofoam. You’ve got strong anti-cyclones blowing through with strong SE winds. It could be that in storms small pieces blow through the air from one whitecap to the next [speculation] and thus travel much more quickly than a standard drift model might predict. We don’t know that doesn’t happen. Therefore all else is speculation.

    As for your fuel exhaustion model, what is the error bar on it? And how many sigmas are we talking about. And what is the error estimate of the error estimate? (Yes I would like to take a look at it. Perhaps there is a dropbox for it?)

    Also, Barry Martin’s model said that 38.5S was doable, so apparently there is room for reasonable–and knowledgeable–people to have somewhat different, honest answers. Yes, fuel exhaustion happened 10 minutes too soon, according to Richard above. But 10 minutes on a 7.5 hour flight is like a 2% error. Are you asking me to believe your all’s fuel exhaustion models are accurate to 2%? What is the percentage difference between your model and Martin’s model? That would at least give a ballpark figure of the proper error estimates.

    Was even the initial fuel load accurate to within 2%? What about the PDA? What if it was in reality only 0.5% instead of 1.5% or 2%? Then there’s the winds. Did NOAA have a bunch of weather balloons at FL350 in the SIO that night? Nope. So all we have are wind models. Are those accurate to less than 2%. I suspect not. If not, then there is no way any fuel exhaustion model can be accurate to less than 2%. All these errors and assumptions add up.

  209. Warren Platts says:

    WARNING: the following post on the philosophy of science of MH370 contains some speculation

    Re: the alleged oversimplicity of the Hardy model

    If I have one overall constructive criticism of the IG’s general approach, it is that you all are all too prone to fall for the “illusion of technique”. Do not take this the wrong way. You all are brilliant, and everyone appreciates the time and effort you have put into this endeavor. This is merely my opinion: it may or may not be helpful.

    The phrase was coined by the philosopher William Barrett who was thinking of Whitehead and Russell’s Principia Mathematica, a gigantic tome that attempted to formalize all human knowledge using symbolic logic. Obviously, their project failed: all the logic in the world cannot capture the world.

    We have the same problem with MH370. There is an idea that if only we calculate enough, we can deduce the crash site. So the IG has produced these incredibly sophisticated models. All that is highly commendable.

    What what are they really? They are not just models: they are models built of models that are based on estimates are in turn built on models. Every step of the way introduces a source of error, no matter how small. And when you have layer upon layer of assumptions, small errors add up to large error bars. Then there are the error bars on the error bars.

    The other problem is that all this leads to a certain arrogance. Again, I don’t mean this as an insult in any way. It is just that when you spend a lot of time building a model, it is easy to gain an overconfidence that may not be warranted. Economists and climatologists also build sophisticated models, but how accurate are those?

    If MH370 was found today, the IG could take their newfound spare time and join forces to start a new hedge fund. They could take their diverse skills and complicated spreadsheets and attempt to make millions. But in reality, the odds are they would not be able to consistently beat the S&P500 index.

    Such overconfidence can then breed a superbad case of confirmation bias.

    I know this guy on twitter: Dwarakanathan. He thinks MH370 crashed into a “cavity” on the island of Sumatra. The man is mentally ill. So to cure him of his illness, I egged him on to going to the spot and looking for himself. And by Krishna he did it. And I feel really bad about that because he had to travel from India, drive for hundreds of miles down the spine of the island and then double back up the southwest coast, then hire a river boat to take him up to this remote village, and then hire more guys to guide him several miles on foot up a mountain in a tiger infested jungle. He disappeared of the internet for several days. So I was very relieved when he made it out alive.

    Sure enough, he found his cavity. A giant mud hole actually: either an abandoned mine or a weird geological phenomenon. Did it cure him of his delusion? Nope. The plane must be buried in the mud! Of course he didn’t find any debris, but he chalks that up to the fact that it was getting late and the guides were getting antsy because they didn’t want to get caught in the dark and eaten by tigers.

    That is an extreme case. But the thing is, we all got a little Dwarakanathan in us. This is what I am afraid of: that the search has devolved into a sort of knee-jerk, brute force attack without even realizing it. The strategy simply seems to be if you don’t find the wreck at point A, then go to point B to the north, rinse, repeat. And of course, as each new search area to the north gets picked, the IG is there to turn out new sophisticated spreadsheets and models to justify the new search area. Which just goes to show a sophisticated model can be concocted to predict just about anything on an ad hoc basis. Heck, Jeff Wise has a sophisticated spoofing model that predicts the aircraft is in Kazakhstan!

    So when I am told that a simple, common sense model that is admittedly near the margin of what is sanely possible is in fact strictly impossible because a thousand line spreadsheet says so, I tend to take that with a grain of salt. To put it bluntly, if the IG was so smart, the mystery of MH370 would have been solved a long time ago. You guys had your chance. People are right to skeptical.

    So the thing to do right now that there’s a break in the action, is to take stock. Maybe the 25S to 20S is indeed the best course of action at this point. It is certainly no skin off my nose. However, I will not be surprised in the slightest if the results turn up negative. AGAIN. So what I want to know is: What is the Plan B?? Because I don’t see one.

    What I want to know is if nothing is found from 25S to 20S, do we return to the SW and push the search out another 50 nm beyond the 7th arc? Or would it be better to root around in Kazakhstan? Or something else?

    This is my bottom line advice to the IG as philosopher of science (and I do happen to have a ABT M.A. in it): figure out exactly what the Plan B is. And then think very carefully about whether the Plan B should really be the Plan A, and if not, why not. Then you will at least have a principled reason to go with your Plan A, whatever that is.

    In particular, you all should ask yourself how happy you are in your marriage to that one BFO value at 24:11. Because everything hinges on that one interepretation, doesn’t it? If another fruitless $70 million spending spree would be enough to make you consider a divorce, then maybe you should consider a divorce now. Or is the choice going to be till death to us part, leaving only the option of revisiting the same old places, like my poor friend Dwarakanathan?

    Again, the above is offered in good faith as hopefully constructive criticism. Make of it what you will.

  210. Victor Iannello says:

    @Warren. You have said so many things that are incorrect. I’ll address some.

    1) My fuel model is based on values from the Boeing LRC tables and holding tables, with physically models based on lift, drag, and engine performance data for in-between values of speed. Those tables are very accurate, as the underlying models are the basis for fuel planning and the Boeing value proposition. Meanwhile, the PDA number is calculated accurately from historical data for 9M-MRO. The error should be less than 1%. Meanwhile, at LRC speed, PDA=1.5%,and ISA+10K temperature correction of 3.4%, the fuel quantity is short by 3.1%. That’s quite a lot. At speeds higher than LRC (which the path ending at 38.5S requires), the shortage is even greater. I offered to do the fuel calculation for a path of your choosing. I’ll make the results available so that others like Bobby can compare to their models. You haven’t yet given me anything. Why not?

    2) Barry Martin’s model does NOT say that 38.5S is reachable. The results that Richard presented were for a turn due south at 18:25 at LRC speed. The endurance fell short by about 10 minutes. But, we know from the BTO and BFO sequence that the plane did not turn south until after 18:28, and to match the BTO values, an early turn had to be to a track angle greater than 180T and at speed greater than LRC. In fact, since Barry Martin derived the fuel flows from the same LRC tables as I used, so for the same conditions, Barry’s model is almost identical to mine. You are latching onto Barry’s results because you think it agrees with your hunch without understanding the underlying models and assumptions.

    3) You seem fascinated with Hardy’s model, despite the fact that our best estimates of fuel consumption and drift say the impact was further up the 7th arc. So you would prefer we throw out the models which you believe are inaccurate and search along Hardy’s path. You provide no basis for doing this other than you don’t believe the models which point elsewhere. If you believe the models are inaccurate, then you should explain the inaccuracies and demonstrate that Hardy’s path is acceptable within the error bounds you establish. That would be the scientific way to refute claims that his path is not acceptable. Instead you bring up the silliness of Dwarakanathan, which has no place on this blog.

    4) You agree that searching to the north on the 7th arc possibly a good way forward. Yet, you are worried about what to do next if that search fails. In the post above, I set out three possibilities, including searching wider. I don’t know what the probability is to find the debris field searching north, and I certainly won’t say it is very likely, but I’d say there is a reasonable chance of success. If we eliminate one possibility, then we consider the two remaining based on all the facts at hand. I don’t need a degree in philosophy to know this, nor is the IG soliciting your advice.

    5) Dennis is correct that it is unlikely that the next search will be anywhere other than further north. The other options require larger search areas and in more challenging seas. So it’s likely a choice between further north or no search.

    6) I have no idea what you are talking about when you say we are in a “marriage to that one BFO value at 24:11”. Perhaps you are referring to the TWO BFO values at 24:19 that show a rapidly increasing descent rate.

  211. Don Thompson says:

    @Warren Platts

    I’ve experienced worse. You will understand pedantism, the elusive secret is held by two BFO records: those recorded at 0019.29Z and 0019.37Z. It is worth paying close attention.

    Plan A isn’t fully closed out, quite a number of questions remain unanswered. I do believe answers to those questions exist.

    My Plan B, should it be necessary, is to wait out the next step change in efficency for seafloor search. One was achieved last year with Ocean Infinity operating eight AUVs simultaneously. A small step has recently been achieved with battery technology adding 50% to an AUV’s mission endurance. The Shell Ocean Discovery XPrize competition is presently underway with eight teams competing to perform a deep sea survey off Karamata, Greece, with no human intervention from shore to depth and back to shore again. Perhaps that competition brings a breakthrough.

    1969 the US Navy spent 4 months searching out 144mi² to find the USS Scorpion wreck, today one Hugin AUV can survey that area in half a day.

    Perhaps you would pose a similar question to Elon Musk? I find shooting spors cars off into space rather pointless, there’s a lot of earth that remains to be explored.

  212. David says:

    @Victor. You have indicated that at least 3.1% more fuel is needed to get to 38.5 deg S.

    I have raised the availability of the aircraft’s energy height as a source of virtual fuel. At fuelless weight at 40,000 ft and 480 knots ground speed that would be equivalent to 500 lb of fuel, ie it would take 500lb at 100% thermal efficiency to lift get that weight to that height.

    Add that to the delta in kinetic energy in slowing from that speed to say 200 knots, another 100 lb of virtual fuel, and the energy in the aircraft going a-wasting at fuel exhaustion is 600 lb.

    3.1% of 49000 lb take off fuel is 1519 lb, so that virtual energy without supplement from step climbing etc is well short. As it stands therefore the aircraft’s energy would be insufficient so I now put that thought aside.

  213. Peter Norton says:

    @Mick Gilbert / re: PAN PAN – you are right, I should have verified the comment before copying it. The main thrust is true though. When an aircraft says it has flight control issues and begins flying erratically, ATC should have handled the flight as an emergency even if not declared as such by the crew.

  214. Mick Gilbert says:

    @Peter Norton

    ATC should have handled the flight as an emergency even if not declared as such by the crew.

    I’m sure that Andrew will have a view on that.

    The two air traffic controllers involved were quite experienced and relatively fresh (only 60 and 90 minutes into their respective shifts). They were not, however, mind readers. Around that time of the morning they would have been managing somewhere in the vicinity of 30-odd departures and half a dozen arrivals.

    While the ATC directions would have been distracting, JT610 had a few options available; simply asking ATC to stand by or actually declaring an emergency were amongst those.

  215. Don Thompson says:

    @Andrew, Mick

    NTSB NTSB/AAR-97/05 (Airborne Express DC-8-63 accident report)

    “Following a December 20, 1995, fatal accident involving an American Airlines (AAL) B-757 near Buga, Colombia, the Safety Board recommended that the FAA:

    A-96-94
    Require that all transport-category aircraft present pilots with angle of attack information in a visual format, and that all air carriers train their pilots to use the information to obtain maximum possible climb performance.”

    “In a December 31, 1996, response to the Safety Board’s letter, the FAA said that it had “initiated an evaluation to assess the operational requirements for an angle of attack indicator to obtain maximum airplane climb performance. This evaluation will include implementation and training requirements, the complexity and cost of the system, and other functions as well as indicating the angle of attack for maximum rate of climb. If it is determined that angle of attack indicators are warranted, the FAA will take appropriate regulatory action.”

    Pending the Safety Board’s evaluation of the FAA’s completed action, on April 11, 1997, the Safety Board classified Safety Recommendation A-96-94 “Open—Acceptable Response.”

    The NTSB’s classification after the FAA’s final response was: ‘Closed – Unacceptable Action

    Regardless of the FAA’s decision, AAL and DAL made AoA visual indication a requirement for their 737NGs & MAX (other Boeing models in AAL/DAL fleets to be confirmed).

  216. Mick Gilbert says:

    G’day Don,

    Thank you for that background information. Boeing published an interesting piece on angle-of-attack in their Aero magazine No 12 back in October 2000.

    I think that an issue here is that with the MAX for the first time on a B737 AOA is being used directly by a flight control system to command control surface movements. Previously AOA had only been used to inform stick shaker activation, the Pitch Limit Indicator and the speed tape display.

    When customers are optioning their aircraft they need to be able to make informed decisions. As a manufacturer, when you elect to offer an AOA Indicator and an AOA DISAGREE alert as an option, I think that it is incumbent on you to ensure that your customers are fully informed as to how AOA data is used. When you make a fundamental change to the way that AOA data is used and fail to inform your customers of that, my view is that you have failed to meet your basic obligation to your customers. I have no doubt that legal teams on either side of the Pacific are formulating their own views on that matter also.

  217. Ben S says:

    Regarding the pilot-suicide or other pilot involved theories:

    It is assumed by many that the SDU was turned off and on as a consequence of human action, but that is not proven.

    The SDU is an avionics device, and not some ordinary PC. Laptops and desktop computers will automatically shut down if they detect certain anomalies, like overheating. But an SDU might not shut itself down due to overheating or other interruptions. And if it wasn’t destroyed by those interruptions, it could well reboot spontaneously.

    In fact I have actually witnessed a small, battery-powered computer (pre-21st century) spontaneously boot (turn on, literally) because of static electricity between my finger and a close-by piece of metal.

    Of course, we know that the aircraft was almost certainly being actively piloted at the point when the SDU rebooted. Thus, we assume it to be a consequence (direct or indirect) of human action, but it is plausible there is a legitimate, non-human explanation.

    I say this because many people believe the pilot to be involved in spite of his relatively un-suspicious background. Also, while the plane was probably being piloted when the SDU rebooted, it was likely not being piloted at the time of its crash. To me, this is indicative of a slow-developing catastrophe, and one example of this could be a fire that moves around the plane and ultimately suffocates everyone. This might also offer an explanation for how the SDU could spontaneously reboot.

    Finally, it’s certainly possible that the SDU was turned off, directly or indirectly by human action, but that doesn’t necessarily mean such action was nefarious.

  218. Andrew says:

    @Don Thompson
    @Mick Gilbert

    An interesting study into the utility of AOA gauges in a few different scenarios:
    Angle-of-attack display in modern commercial aircraft Do we need that?
    The study concluded: ‘Probably not’.

    My view is that the PLI on the various Boeing types is more useful than an AOA gauge in situations where the pilots need maximum climb performance. However, on some types the PLI is only available when the flaps are down.

  219. airlandseaman says:

    Warren Platts: Your assumption that there is no plan B is 100% wrong (like other statements you have made). In fact, there are (at least) 3 possible approaches to continuing the search. Plan A (S25 to S20 +/- 22nm) may be no more likely than B or C, but it is safer, faster and cheaper to to try NEXT. That is the logic. That is why OI will try this approach next.

  220. Richard Godfrey says:

    @Warren Platts

    If you look at what you actually wrote, the previous paragraph was attributed to @TBill. The paragraph I quoted from “you” was not attributed to anyone else.

    It appears that almost everything you write is misleading.

  221. flatpack says:

    @Victor

    I have been busy for a while and just now quickly read-in, apologies if have missed someting relevant.

    You posted back on October 29, 2018 at 7:43 pm

    “An important question to answer is why the simulated flight was created, and why the flight files were saved in the manner they were saved. I don’t think the purpose was to plan and practice the diversion, as there are better tools for doing this.”

    Yes, yes there are.

    Better tools that is, and ZS would have had access to them. He may just have been prudently avoiding using them. However that does not address the manner of file saving.

    You are right, your question is an important one.

    Then on October 30, 2018 at 6:27 am you asked:-

    “However, in this case, the user moved the plane along a path, changing position, fuel level, and altitude. What planning or practice do you think he obtained when he conducted the simulation?”

    My view is that fuel dumping is part of the solution.

    One advantage of the sim over ‘better tools’ is that it allows visualisation of a flight from the perspective of others (or indeed by others). Remember that ZS was apparently unsatisfied with the realism of it’s portrayal in sim.

    So, my answer is that he obtained some idea of how visible fuel dumping would be at different points in the flight. Whether visibility reduction or enhancement (such as simulating a flight in trouble) was his goal, is another matter.

    Visibility would also include consideration of any slick deposited.

    The manner of file saving seems reasonable to me for such a comparative exercise that may need revisiting.

    Then on October 30, 2018 at 7:08 am you stated:-

    “@TBill said: It is frustrating when you suggest there is no planning value of the Flight Sim. That leaves joyride for the experience of pre-living a suicide mission as the only apparent purpose.

    I have long suspected there was another purpose.”

    Well, your suspicion intrigues me!

    My suspicion is that your suspicion does not involve fuel dumping or general contrail management.

    The question is whether there is anything in the file fragments that can falsify my scenario or yours.

  222. TBill says:

    @ALSM
    “That is why OI will try this approach next.”
    Great news!!??

    @flatpack
    Your post gets into the controversial topic of trying to predict what actually happened to MH370, versus a more generic approach to just search Arc7@20-25 South +/-22 nm.

    But I have tried to consolidate my personal MH370 articles/thoughts on Twitter, and I put my fuel dumping graphic over there for you. I am open to fuel dumping but my current hypothesis is the pilot was short on fuel. Keep in mind, once fuel gets to a certain low level, it cannot be dumped, which adds a possible reason for a hypothetical glide with APU or residual fuel on the left wing- to empty the tanks.

    https://twitter.com/HDTBill/status/1070717041856913408

  223. Warren Platts says:

    Re: right mistakes versus wrong mistakes.

    In my opinion [speculation], the MH370 problem is rather analogous to drilling a horizontal well in the Appalachian mountains. One is forced to proceed forward on a multi-million dollar undertaking under conditions of profound uncertainty. In my career geosteering such wells, I must say that my studies in the philosophy of science proved more useful than all my geology training put together.

    The problem in drilling in the mountains is that the formations are unpredictable. You know they are there; you just don’t know exactly where they are. So you make educated guesses, form working hypotheses, and learn as you go.

    An inherent part of this game is that mistakes will be made. Knowing that, the key is to try to make the right mistakes, and avoid making the wrong mistakes.

    To a normal person, “right mistake” sounds like a contradiction in terms and “wrong mistake” sounds like a redundant tautology. But that is not the case. A right mistake minimizes the collateral damage, whereas a wrong mistake results in hugely expensive consequences.

    It is the same with MH370. We know mistakes will be made because mistakes have already been made. The heady confidence with which the initial predictions were made proved to be nothing but hubris. However, my goal here is not to point fingers and assign blame, but to assist in the process of stock-taking.

    In the case of MH370, a right mistake would be if one’s best-guessed that the aircraft crossed 7th Arc at 38.5S when the reality is that it crossed at 35.5S–or vice versa. The truth is not far from the wrongness. The resources required to rectify the mistake are relatively minimal.

    An example of a wrong mistake would be to correctly assess that the aircraft crossed at around 35.5S, but then to barely miss finding the wreckage because of a mistaken theory about how far to the south the wreck could be. That mistake then ramifies. One falsely concludes that the aircraft did not cross at 35.5S; the entire search is then led on a hundred thousand square kilometer wild goose chase costing multiple hundreds of millions of dollars.

    Here is another example of a wrong mistake. If it is the case that drift models are the decisive piece of evidence for finding MH370, then the wrong mistake was not to realize that fact from the very the beginning. That is, the plan should have been to work out detailed drift models, and then just WAIT, keep our powder dry, and conserve our resources until the results came in. And just maybe even, god forbid, actually send out formal search parties to look for debris items. With the drift models in hand ahead of time, we then could make a reasoned deduction that the aircaft must have crossed at 25S to 20S–or wherever–once the debris was found. Thus the wrong mistake in that case was to plunge ahead on the 100,000 square kilometer wild goose chase to the SW.

    Either way, the search for MH370 thus far has been a colossal wrong mistake. Given the resources expended, the wreckage should have been found. The question is whether the next search–if there is one–will also turn out to be a wrong mistake. After all, given the dismal results so far, why would anyone in their right mind plunk down another 100 or 50 million dollars?

    This is my fear: the original search to the SW was based on presumably good reasons. Whether it was Captain Hardy, the ATSB, or the IG, though they differed in precise predictions, they all pretty much suggested the same general region to the SW. Those good reasons, many of them based on independent methods, did not necessarily go away just because the wreckage was not found relatively close to 7th arc.

    Meanwhile, I am afraid that the uncertainties associated with the drift models are being underestimated. We must beware of the “shiny” factor. After all, we only have a sample size of about 30 some pieces. That in itself is piece of evidence, if previous debris studies are any guide. That is, the sample size suggests a relatively small initial number of a few thousand debris items rather than a debris field consisting of a few 10s of thousands of confetti. The ratio of interior to exterior items is also suggestive imo.

    –My two cents

  224. Victor Iannello says:

    @flatpack: Why do you believe the captain was interested in dumping fuel? The fuel endurance indicates that little or no fuel was dumped.

  225. TBill says:

    @Warren Platts
    Are you in Pa. drilling, my old stomping grounds? I don’t think Va. and Md. are doing much. I think that leaves Pa. and less possibly WV. And here I had you as a Hardy supporter from Australia.

  226. haxi says:

    @ALSM,

    What do you mean by saying “That is why OI will try this approach next”? Any insider news?

  227. Victor Iannello says:

    And here’s a new article from The Australian by Robyn Ironside. (David Gallo seems resentful that his offers to help the investigation were ignored.)

    MH370 debris find ‘not enough’ to reopen probe

    Five new pieces of debris almost certainly from missing Malaysia Airlines flight MH370 are unlikely to be enough to reopen the investigation.

    Handed to Malaysia’s Transport Minister in Kuala Lumpur last week, the fragments collected in Madagascar in the southwest Indian Ocean include a piece of flooring from a Boeing 777. Certainty on the piece’s origin stems from a serial number matching a Boeing Material Specification.

    Despite representing an important part of the puzzle, Uni­versity of Western Australia oceanographer Charitha Pattiaratchi said the debris did not shed further light on the final resting place of the Boeing 777, which went missing in March 2014.

    “They were found in areas where our models predicted they would end up,” Professor Pattiaratchi told The Australian, after inspecting the debris at the request of American adventurer Blaine Gibson.

    “But along with the other items found, (the debris) represents something like 0.01 per cent of the whole plane. If you tried to put the 32 pieces we’ve found into a plane, they would cover only a very small area.”
    Even Mr Gibson acknowledged the bittersweet reality of the latest discoveries.

    “They are small pieces of a very large puzzle,” he said. “It’s useful information but it’s not enough. It tells us something about the where, it tells us something about the what. It doesn’t tell us the who, it doesn’t tell us the why.”

    As a tireless supporter of MH370 families who have mounted their own campaign to solve the near five-year mystery, Mr Gibson desperately wants Malaysia’s government to reopen the investigation, which was put on ice in the absence of “new credible evidence”.

    Transport Minister Loke Siew Fook repeated that condition of resuming a search last week after taking delivery of the debris.

    Hopes that private companies might step in came to fruition earlier this year when Ocean Infinity undertook a 90-day search on a “no find, no fee” basis. But at this stage it would appear the company has no intention of another search despite its recent success in locating the missing Argentine submarine ARA San Juan.

    Woods Hole Oceanographic Institute also ruled out launching its own search, after showing interest in the first year after MH370 vanished. Director of special projects David Gallo said he was not convinced about the seventh arc hypothesis but his approaches to Malaysia and Boeing were ignored. “No emails or phone calls were returned,” he told The Australian.

    And while the latest pieces may not pinpoint the plane’s whereabouts, Mr Gibson said they strongly suggested a violent end for the 777-200ER.

    “The piece of flooring is just one more piece of evidence added to the others, that proves that the main cabin tragically shattered on impact,” he said.

  228. mash says:

    @Victor Iannello

    That’s exactly the point. Why is it “both absurd and perversely sensible” (at the same time)?

    Perhaps that indicates that one is (clearly) not looking at the problem at the right perspective.

    Maybe that also suggests that a more ‘complicated’ solution could explain the “contradictory events” more ‘simply’.

    Any suggestions?

    [Maybe there are really two opposite/opposing forces at work during the flight …]

  229. Viking says:

    @Victor

    Thanks for informing me on the Australian news article.

  230. Victor Iannello says:

    @Viking: You already know how many of us feel about your neglecting the second BTO, the large BFO error, and your observations about the received signal strength. I haven’t commented before on your suggestion that the captain jumped from the plane. At a speed of 220 KIAS, which would be the approximate minimum speed on autopilot, it would be suicide. Also, the only door whose location could possibly accommodate a jumper would the bulk cargo door, and it is only about 3’x4′. Surely the captain would be aware of these challenges and wouldn’t have tried it if his intention was to live.

  231. Warren Platts says:

    >“But along with the other items found, (the debris) represents something like 0.01 per cent of the whole plane. If you tried to put the 32 pieces we’ve found into a plane, they would cover only a very small area.”

    Hmmm… 32 pieces / 0.01% = 320,000 pieces

    Now that would be confetti. If there were really 320,000 bits, thousands or at least hundreds of debris items would have been found by now.

  232. jinott says:

    Bout time Beach searches took shape at potentially more useful locations to enable the drift studies to gain accuracy.

  233. Warren Platts says:

    >”“The piece of flooring is just one more piece of evidence added to the others, that proves that the main cabin tragically shattered on impact,” he said.”

    Not exactly sure what “main cabin shattered” means. If that means the entire main cabin was turned into confetti, one single piece of flooring surely does not “prove” that. If there was a Comoros Island-style impact, the fuselage could very well be split into two or three pieces.

    In that case, certainly pieces of flooring and other interior materials would be released, but the majority would go down with the rest of the wreckage. Meanwhile, the exterior of the aircraft would have to withstand the brunt of the impact, entailing that more exterior pieces would be released than interior pieces.

    On the other hand, if the entire aircraft was turned into “confetti”, the energy is distributed more or less equally. The entire aircraft is shredded, and there would be no preferential sorting of exterior versus interior pieces. If anything, there would be more interior pieces since they would not have to be made of structurally strong metals that would sink.

    Thus, what do we have: 32 pieces, of which perhaps 3 were from the interior. That is an exterior/interior ratio of 10 to 1. How does that “suggest a high-speed impact”?

  234. Don Thompson says:

    @Warren Platts, asked “what do we have: 32 pieces, of which perhaps 3 were from the interior.

    What do we have: 236 comments, of which perhaps 21 demonstrate an incomplete grasp of facts (e.g. debris count from cabin interior: 6, not 3).

    @all

    Seems, like me, others have noticed that certain sections of the press maintain watch on arxiv publications, and that arxiv applies only moderation, no peer review criteria, in its process.

  235. airlandseaman says:

    Regarding my comment at December 6, 2018 at 7:27 am: “… there are (at least) 3 possible approaches to continuing the search. Plan A (S25 to S20 +/- 22nm) may be no more likely than B or C, but it is safer, faster and cheaper to to try NEXT. That is the logic. That is why OI will try this approach next.”

    Some have asked me to clarify the last sentence. No, as far as I know, OI has NOT made a decision to return to the search as of today. My point was that if they do return, there are multiple choices already under consideration, contrary to Warren’s assumption, but Plan A (S25 to S20 +/- 22nm) is the preferred place to restart, for the reasons given, if they do restart.

  236. Wall says:

    @haxi

    Please don’t believe this kind of crap. It’s rubbish. It contradicts everything we know about this plane. Don’t read information about this airplane if it was published by: Daily Star, The Sun, Daily Mail, Sputnik International, the Star Online and Express.

  237. Warren Platts says:

    @airlandseaman:

    Fair enough. But what exactly is the Plan B and Plan C? Cheaper, faster, safer: that used to be a NASA slogan..

    @Don Thompson:

    Thank you for clarifying the exact exterior/interior ratio. It is only 5:1, rather than 10:1.

  238. Mick Gilbert says:

    @Warren Platts

    Regarding the ratio of interior to exterior floating debris, what were you expecting?

  239. Warren Platts says:

    RE: Martin Kristensen https://arxiv.org/ftp/arxiv/papers/1811/1811.09315.pdf

    “Therefore, the only plausible explanations are that they wanted to land in Bandar Aceh or abort the flight by parachute. Since the airplane did not land, the only option is parachuting. In order to do this they had to fly low and slow (as observed by Kate Tee) to open a hatch and get out. They programmed a return to normal flying-height into the autopilot before jumping. Therefore the plane returned to 11 km height after Bandar Aceh without a pressurized cabin (due to the leak through the open hatch) causing death for everybody on board who might still have been alive.”

    This is what I mean about the “Illusion of Technique”. It can lead to absurd conclusions, but because it has all the mathematical bells and whistles, he is 90% sure he is correct.

  240. Wall says:

    @Warren Platts

    He is most likely completely wrong. The crash site he is talking about doesn’t even lie on the 7th-arc. Don’t believe everything you read.

  241. Wall says:

    @Warren Platts,

    I meant to say, it does lie on the 7th arc but is not consistent with drift models. It just doesn’t make any sense. Besides, there is a consensus about the intention of ZS. There’s just too many coincidences.

  242. Warren Platts says:

    @Mick Gilbert: >Regarding the ratio of interior to exterior floating debris, what were you expecting?

    [speculation]Well, it is a question worth asking, isn’t it? Just looking at some of the pictures coming from the latest Lion Air crash, it seems like there is a lot of interior debris being recovered.

    Think about what goes into an airliner. You have your exterior parts. Some of these are made of honeycomb material that will float, but much of it is metal because of structural requirements.

    Then there are the interior parts. Things like floor panels, side panels, seat cushions. Here the design emphasis is going to be on light weight, rather than structural strength. Area-wise, the amount of material on the inside is not going to be much less than the outside.

    Now if you look at the other Lion Air crash in Bali in 2013, it was a water “landing” at low speed. Yet the fuselage is split in the middle. So obviously it does not take a lot of energy to split an airliner in two. In such circumstances, some interior debris will be produced, but not as much as much as in a high energy impact that reduces the entire aircraft to “confetti”.

    Do you see? The exterior/interior debris ratio is a possible means to determine the truth of two or three rival hypotheses: low energy impact versus medium-energy impact versus high energy impact.

    In a low energy impact, it would be like the Hudson River: the trailing edges of the wings were shredded pretty good, but hardly any interior debris was released. Therefore, since we have some interior debris, we can exclude a Hudson River-style low energy impact.

    Now we are stuck with an apparent exterior/interior ratio of 5:1. What are we supposed to make of that apparent fact?

    We have analogues. There was the Comoros event where the attempted ditching split the aircraft fuselage in to two or three big pieces, then there is the recent Lion Air crash where the entire aircraft was disintegrated.

    Clearly, in an ET961-style crash, some interior debris would be released. But should we not expect the exterior/interior debris ratio to be lower for a JT610-style crash? Shouldn’t that ratio be more like 1, or even less than 1, rather than 5?

    I don’t know the answer to that question, but it is worth asking, I think.[/speculation]

  243. Peter Norton says:

    @David
    Thank you for pinging me with your thoughts. I have not seen this issue raised before. If I understand you correctly, what you are saying is that:

    using the last fuel reserves by gliding with idle thrust = longer distance than
    flying at cruise speed until fuel exhaustion + glide without power

    correct ?

    > David:
    > If you say that would carry his 7th arc BTO further than it was, well not necessarily.
    > If he had turned right at the 7th arc (remember this is a crude example) then travelled
    > along the 7th arc to fuel exhaustion, that would have extended his crash latitude
    > south. In a more refined example that principle would be integrated with speeds,
    > courses, BTO’s and BFO’s in the same way as step climbs would need to be.

    You mean instead of flying along the 2 catheti (with ~90° turn right at ARC7) the pilot could have anticipated the final end point by flying along the hypotenuse, correct ?

  244. Richard Godfrey says:

    @Warren

    You stated, with respect to @Viking “he is 90% sure he is correct.”

    How do you conclude he is not 100% sure?

  245. Warren Platts says:

    >@ TBill: Are you in Pa. drilling, my old stomping grounds? I don’t think Va. and Md.

    Pa. yes. I am out of the business these days, semi-retired. But I still run the informal “Platts Frack Tank Farm Index”. In my town there is a parking lot where they keep frack tanks. After the 2015 bust it was almost full. Then it mostly emptied out a year or so ago. However, the other day, I saw that it is starting to fill back up, so things are definitely slowing down again.

  246. Warren Platts says:

    @Richard Godfrey: You said with respect to @Viking “he is 90% sure he is correct.” How do you conclude he is not 100% sure?

    It is in the Arxiv paper:

    “We propose instead a new, focused search zone of 3500 km2 centred at (13.279˚ South, 106.964˚ East) with slightly elliptical shape along the 7th arc and a total length of 140 km and width of 30 km. The probability of finding the plane there is above 90%.”

    But that is not the first time someone has published similar estimates based on sophisticated mathematical models. There was supposed to be an 80% chance that the Ocean Infinity search would be successful. Now there is an 80% chance the aircraft is between 25S and 20S.

    This is what I mean about excessive reliance on mathematical models, spreadsheets, and numerical simulations: it can lead to overconfidence. The constructors of such models are too close to them, and thus underestimate the inherent uncertainties involved. I am of course as guilty of this as anyone, given my marriage to the waypoint model.

    Did you see the latest Science article on the AlphaZero program that has defeated StockFish at chess, despite the fact that the number of moves it evaluates is 1,000 to 10,000 times less than Stockfish? More than that, it has developed an entirely new, agressive style of play where it willingly sacrifices lots of pieces in order to swarm the King.

    It is a tabula rasa design. That is, they just give it the rules of chess, and it figures out its own strategies. Whereas Stockfish has all kinds of hand-tailored heuristics inputted by humans.

    Do you see? Stockfish contains human biases, whereas AlphaZero does not. Something like AlphaZero would be very nice to have to help solve the MH370 problem: a tabula rasa A.I. where we could feed everything we think we know into the computer, and it would print out a solution that would contain no human bias. 😉

    http://science.sciencemag.org/content/362/6419/1140/tab-pdf

  247. Warren Platts says:

    @Victor: > I offered to do the fuel calculation for a path of your choosing. I’ll make the results available so that others like Bobby can compare to their models. You haven’t yet given me anything. Why not?

    I would like to redo my model from scratch. Do you or anyone else have a kmz file or a list of distances to the satellite that would describe the latest, best ping rings. I have 3 or 4 sets, but I’m not sure what is the best to use.

    IIRC the very first set of ping rings, a constant velocity model worked pretty well. However, with the revised arcs, it seemed like the final leg required some slowing down of ground speed. I guess this could happen if as the plane’s fuel load declined, it might gradually gain altitude, and since the air is thinner, at 0.84 mach the speed and the drag would be lower, and that might relieve your fuel model a little bit.

    At any rate, the path itself is easy enough to describe: MEKAR NILAM POVUS ISBIX MUTMI RUNUT, followed by 189T. Thus from POVUS to 7th Arc it is a perfectly straight (rhumb) line. It pops out just a few miles from Hardy’s path. If MEKAR NILAM POVUS is cheating, you could start it at about 7N 95E, or ANOKO. I would be very interested and grateful if you ran this through your fuel model.

    The other thing is this: I first discussed this model at airliners.net, and almost to person, the aviation wonks told me that “Real pilots fly magnetic courses, not true courses, except in polar regions.” Thus unless another waypoint was manually entered, the aircraft would fly the same magnetic heading it was on at the last waypoint. At RUNUT, that is about 192M. If it followed 192M after that, it would curve to the left and wind up at 7th Arc considerably to the NE of 38.7S. In that case, there would be too much fuel and the aircraft would fly past the 7th Arc unless other weird stuff was going on.

  248. Richard Godfrey says:

    @Warren

    My question was rhetorical! 😂

    That means you do not need to feel obliged to write another comment even longer than my arm!

  249. airlandseaman says:

    Regarding Plan B and Plan C: We have been discussing them here for some time. Perhaps it would be more accurate to say we have been discussing options, not specific plans per se. But let’s not split hairs.

    Plan A = Look NE of areas already searched (S25-S20)
    Plan B = Recheck all questionable PoIs in area already searched (remembering the SJ experience)
    Plan C = Search slightly wider than +/-22nm…expanding to maybe +/-30 nm?

    I don’t think there is any merit in searching south of the area already searched (~S39.5).

    Don’t include me in the gaggle that thinks there is now an 80% chance Plan A is where the plane is. A, B and C are all about equal chances IMO, but A is far faster, easier and safer to eliminate first.

  250. Wall says:

    Hey folks,

    I was wondering, does anyone know where these pieces of debris have been found. Every item that has been found so far is listed in the last report. But the only thing I could find about the items is that they were all found in Madagascar, which is a big country.

  251. Victor Iannello says:

    @Warren Platts: I would like to redo my model from scratch.

    Why don’t you do that. When you something definitive that you want to check for fuel consumption, I’ll run it through my model. I need Mach number, altitude, and temperature offset from ISA versus time.

  252. Wall says:

    Never mind, I already see there’s a list!

  253. Victor Iannello says:

    @Wall: If you’re talking about the new items, Blaine provided a description that includes the location of the find. I included that description in a zipped file along with low resolutions photos of the debris.

  254. lkr says:

    @Wall: I don’t think the localities have been stated, but IMO he’s going back to resort areas on the north and central east coast, and that at least some of these pieces have been picked up over time by local contacts from previous searches.

    It hardly matters. Madagascar gets a big chunk of drift from any reasonable POI so even on first pass, these pieces don’t help determine where it lies on the 7th arc.

  255. TBill says:

    @Warren
    You might have a point that the official MH370 Arcs/kmz’s are hard to find in the literature (slight gap in the online knowledge base). I am thinking they were orignally on Duncan Steel’s site but since taken down.

    If you want me to, I can email you what I use which I believe ALSM kindly provided me a year ago or so. I also use DrB’s readily available reference CBTO CBFO calc worksheet, if I want to do it by hand at different altitudes.

  256. airlandseaman says:

    Re 5 new pieces of debris: These new pieces do not tell us anything new about the 370 POI. What they tell us is that:

    There are likely many pieces of small debris still floating around
    Debris arrival times cannot be trusted to pin point a starting point
    The flaperon and flap segment remain the only large pieces found
    The floor debris is more evidence pointing to a high energy impact

    These are all clues, but they do not help ID the POI.

  257. airlandseaman says:

    Here are the ARCs again: http://bit.ly/2Mg5E1O

    The 7th arc was recomputed (refined by ~2nm) for the OI search, available here: https://goo.gl/cHPuUU

  258. airlandseaman says:

    This link has the revised 7th arc and the corresponding +/-22nm search limits used by OI:
    http://bit.ly/2RFGybn

  259. Warren Platts says:

    @Victor Ianello: > I need Mach number, altitude, and temperature offset from ISA versus time.

    Might as well keep it simple: Mach number = 0.84; altitude = FL350; temperature = ISA – 20C constant over time.

  260. Victor Iannello says:

    @Warren Platts: ISA-20K offset is not close to being realistic. A more reasonable number is an average of ISA+9K for time period of 18:40 until fuel exhaustion for a path that ends near 38.5S.

  261. Warren Platts says:

    @airlandseaman: The floor debris is more evidence pointing to a high energy impact.

    Then how do you explain the 5:1 exterior/interior debris ratio, as well as the overall small number of pieces found? Previous debris studies indicate that the recovery rate should be on the order of a percent.

  262. airlandseaman says:

    Warren: Honeycomb construction materials make up a large majority of the materials that remain afloat for a long time. I do not know the ratio of internal vs. external honeycomb materials, but I do know that much of the interior does not consist of honeycomb materials. For example, all the seats are fabricated from metal, plastic and fabrics. The seat cushions are designed to float for a few days or weeks, but not years.

  263. Warren Platts says:

    Guys: thanks for the arc information. The .kmz files provided by Mike loaded up beautifully.

  264. airlandseaman says:

    Regarding the ARC accuracy: Note that all the arcs from 2015 were calculated with a slightly different bias calibration. They are off by ~2 nm vis a vis the bias used for the 2018 7th arc calculations. I did not bother updating ARCs prior to 7 since their accuracy does not materially affect the search area.

  265. Victor Iannello says:

    @Warren Platts: At an average temperature ISA temperature offset of +9K after 18:39, an average PDA of 1.5%, and speed of M.84 at FL350, the fuel tanks run dry at 23:46, or about 29 minutes sooner than we estimate based on the time of the final log-on. That’s not even close to being an acceptable path.

  266. Warren Platts says:

    @airlandseaman: Do you know the 2-sigma expected error bar for the latest arcs? That is, how many nm away from the arc could be considered reasonable?

    @Victor Ianello: Why is the temperature offset +9K instead of -20K? Would that make any difference?

  267. airlandseaman says:

    ARC location uncertainty: ±5.3 nm (99%)
    Altitude uncertainty at 00:19:37: 1.5 nm (95%)
    Post FE path uncertainty: ±10 nm (~80%?), ±20 nm (95%?)

    Obviously, the post FE uncertainty is dominant.

  268. Victor Iannello says:

    @Warren Platts: I am using GDAS temperatures from the day of the accident. ISA+9K means that the temperature at FL350 is (on average) 9K warmer than the standard temperature at that altitude. It was a bit warmer at northern latitudes, and a bit colder as the plane flew south. The temperature has a significant effect on fuel flow rate, and therefore the fuel flow rate has to be corrected for temperature deviation from ISA conditions. In any event, your M0.84 path ending near 38.5S will not have enough fuel, which is really no surprise.

  269. David says:

    @Peter Norton. Thank you for your interest.
    In reverse order, “You mean instead of flying along the 2 catheti (with ~90° turn right at ARC7) the pilot could have anticipated the final end point by flying along the hypotenuse, correct ?” Yes, bearing in mind that his speed would need to be higher to match the 7th arc log-on timing having covered a greater distance.

    “If I understand you correctly, what you are saying is that using the last fuel reserves by gliding with idle thrust = longer distance than flying at cruise speed until fuel exhaustion + glide without power”

    No, I think fuel exhaustion range plus a glide would yield a greater total flight distance. However current fuel consumption estimates the 7th arc log-on will be short of 38.5 deg S. In concept by delaying the log-on to the end of the glide means the aircraft will have travelled farther before the log-on ie that will be farther SW along the arc. Thus a powered glide or engine restart (supposing still that APU restart on fuel exhaustion is what led to the log-on) will delay the log-on, not in time of course but, with higher speed, at a greater distance.

    As an example of a more sophisticated model he could begin a very gradual descent from say the 5th arc, largely maintaining power, gradually using that energy and saving fuel, that contributing to a speed increase needed for the increased distance between that and the 6th, then 6th to 7th.

    Were this all on the same course he had been flying since FMT naturally the whole track south would have been kicked to the west slightly, requiring other integration measures.

    However a problem with this example is that as his altitude decreased, drag would increase consuming some if not all of the 600 lbs of virtual fuel.

    Thus there is an optimum point at which he starts a descent and the trade-off between descent rate and engine throttle back. To me, most likely he would aim for a low down fuel exhaustion and after a short delay following fuel exhaustion, that would be followed by a very steep nose down for a last few thousand feet, prompting the BFOs and a high ROD at impact at the 7th arc. Maybe he simulated this at home.

    Another part of the equation is the BFO at the 6th arc. From memory that would accommodate a slow ROD at that point but if that proved a stopper, descending sharply after the 6th at low power while increasing ground speed slightly, while flying that hypotenuse, would be an (inelegant) get-around.

    This is all rather abstract and ‘in principle’. I must admit to not having a clue as to how much further SW on the arc he could get but based on what @Victor said was the extra fuel he would need to get to 38.5 deg S otherwise, at best it would be well short of that so would need supplement by step climbs and maybe some bleed air reduction.

    I did say that the above depended on the APU start as the log-on power source. However having a pilot allows alternatives such as him selecting power which happened to reboot the SDU, as most probably he did at 18:25, then selecting it off. Related particularly to more northern latitudes this could have been well before fuel exhaustion. Bear in mind that with the left tie breaker isolated, an APU autostart subsequently at actual fuel exhaustion would not repower the SDU for another log-on.

    >Finally, and for @TBill on the lateral thinking side, as to bleed air reduction, to achieve that without risk of decompression sickness a pilot could simply manually shut the outflow valves.

  270. Ben S says:

    Has it ever been ruled out by a preponderance of evidence that the flight could have ended somewhere on the 7th arc north of the 22nd parallel and south of Christmas Island?

    To be clear, I’m not promoting the “analysis” of Martin Kristensen, but I have previously wondered about his principle insinuation, that the plane is in this area. In light of the barnacle fouling researching, it seems reasonable to consider areas north of the old search area.

    Is there some kind of probability-based cutoff for how far north from the ATSB/Ocean Infinity search area the plane could possibly be?

    If you believe the recent debris analysis, the plane was not piloted at the end. But it seems pretty reasonable to think that the plane was piloted around the time of its turn and/or satcom reboot. Given that we can’t say with any certainty whether it was piloted after 19:41, are there any probabilistic routes or areas above the 22.0 latitude?

    Finally, on a separate note, the idea of MH370 being pursued by a fighter jet is certainly interesting. For example, suppose the edited logs contained details of an intercept, and/or MH370 was briefly intercepted but the military was unable to assist them. Malaysia’s failure to provide assistance or do anything after intercepting it would be hugely embarrassing because of all the people who died, and since they don’t actually know the final resting place anyway, they could have decided to cover it up altogether. Of course, if all of this had any truth, it would also mean that Malaysia has a much better idea of what was happening as it flew over the Malacca strait.

  271. Warren Platts says:

    @Victor Ianello: > “The temperature has a significant effect on fuel flow rate, and therefore the fuel flow rate has to be corrected for temperature deviation from ISA conditions. In any event, your M0.84 path ending near 38.5S will not have enough fuel, which is really no surprise.”

    This is an interesting bit of epistemology. We have two models, a weather model of a remote area based on few observations, and a flight path model, linked by a fuel efficiency model.

    The question is: Why should the weather model constrain the flight path model, rather than vice versa? What we want to know is whether the flight path model is within the realm of possibility.

    The temperatures predicted by the weather model are uncertain: there must be an error bar to the temperature estimates of at least a few degrees K.

    Then there is the sensitivity of the fuel efficiency model. How many degrees K deviation would be required to get the fuel efficiency model to close with the flight path model?

  272. Peter Norton says:

    ———-
    Richard Godfrey: “assumptions that the altitude was constant at 35,000 feet […] Simon Hardy’s path model is grossly over simplified, erroneous and misleading.”

    Victor Iannello: “To reach 39.5S latitude requires a speed of M0.84 at FL350. […] Despite what you keep saying, by a large margin, there was not enough fuel to reach 39.5S latitude.”
    ———-

    Victor, you keep misquoting the latitude and other things.
    I have stated multiple times already that …

    • the path to Hardy’s area, as presented in the 3-part video series, does not cross ARC7 at S39.5 but S38.6

    • Hardy’s proof-of-concept video assumes wind=0, probably for simplicity. If you account for wind, his path crosses ARC7 farther NE ! Don Thompson also mentioned that already.

    • I also stated that all range-extending conditions should be taken into account. For example, I don’t assume FL350 but specifically mentioned a gradual, progressive climb to the more fuel-efficient maximum altitude as much as progressive weight loss permits.

    • I also mentioned other possible range-extending techniques.

    Not that I am suggesting these points are misquoted deliberately, but I note that I have had to correct them several times already, which I find a bit curious and which makes me feel like Hardy’s model isn’t treated correctly, and, by extension a southern end-point, generally speaking.

    The points above may or may not be enough to reach Hardy’s exact area, but that’s hardly the point, or aren’t we interested in potentially finding MH370 in its vicinity? While I want to stress how much I value the scientific accomplishments everyone here has contributed in all these years, I’m disappointed about the wholesale dismissal of Hardy’s method let alone the general possibility of a southern end-point.

    Does it matter whether MH370 is in the specific area shown by Hardy in his youtube video or a bit outside in its vicinity? Imagine MH370 lies slightly outside Hardy’s area and we don’t find it, because you dismiss Hardy’s model on technicality. Can we really dismiss all the parameters magically coming together in his model as pure “coincidences”?

    If there is more to it than just coincidence, i.e. the basic concept is correct and only some aspects of it need refinement and precision, then MH370 will be found in this general area and not in S25-S20. If your criticism of Hardy’s model amounts to nothing more than the neglect of wind and that it’s at or slightly beyond the edge of your fuel calculations, then MH370 may very well lie just a bit further NE of Hardy’s area.

    I see this caricature in my mind where MH370 is found in 10 years at S37.9 (or wherever you think the max fuel range is) farther beyond ARC7, with Victor saying “that’s not Hardy’s area!”, Don saying “that doesn’t count, Hardy forgot the wind”, and Richard saying “that was misleading”.

    I hope you don’t get upset about this little attempt of humour. I value of all of you. I didn’t feel strongly about the south before, but the more it gets dismissed summarily, the more I feel someone has to step up to make sure MH370 was not narrowly missed down there.

    To use a software development analogy, I think we should not adopt a proprietary approach to path models, but an open source approach where existing path models are not treated as “misleading” and are only destructively criticized but constructively enhanced together. I think we all want to find MH370, so if you see areas for improvement in Hardy’s model, why don’t you want to improve it rather than dismiss it in its entirety?

  273. Victor Iannello says:

    @Warren Platts: I give up. When a model doesn’t predict what you want it to predict, you keep questioning the model, even when your path requires the results to be well outside reasonable error bounds. Your wishing to impose a temperature offset of ISA-20K demonstrates you are outside of the realm of reality.

    Perhaps @DrB will be willing to work with you. His fuel model will have similar results to mine. I won’t continue to try to persuade somebody that really doesn’t want to be persuaded.

  274. Victor Iannello says:

    @Peter Norton: At @Warren Platt’s request, I presented results for a flight at FL350 that ends at 38.5S latitude. That’s very close to the latitude of interest to you (38.6S). By my estimate, the fuel endurance falls short by 29 minutes.

    Yes, climbing to a higher altitude lengthens fuel endurance for two reasons: The fuel flow is reduced for the same temperature offset from ISA, but also the ISA offset is reduced. However, it takes additional fuel to climb to the higher altitude. Now during a normal flight, that extra fuel is “recovered” due to the longer idle descent from the higher altitude. In our case, the benefit of reduced flow at the higher altitude has to be balanced against additional fuel consumed to get to the higher altitude.

    @DrB has studied fuel constraints versus crossing latitude on the 7th arc much more closely than me, including a very systematic investigation across the parameter space. Perhaps he’ll weigh in.

  275. Peter Norton says:

    Victor Iannello: “To reach 39.5S latitude requires a speed of M0.84 at FL350. That’s faster than LRC.”

    What speed do you calculate for 39.5S 38.6S with FMT at 18:36 at ANOKO ?

    Given that Cpt. Simon Hardy is a very experienced British Airways B777 pilot and instructor, I’m wondering why he obtains to such different results.

    Quoting from his videos:
    ———-
    Cpt. Hardy says: “488 knots is the cruising speed of a triple seven”

    Cpt. Hardy further says that he “used an airline planning system” to derive the fuel range of 2760 nm at which point he shows a route chart sheet with wind vectors (so apparently he has factored meteorological conditions or at least the wind to calculate the 2760 nm).
    ———-

    Cpt. Hardy’s data, as far as I was able to gather from his video:


    ARC7 crossing: at ~S38.6 (my measurement from his GE screenshot)
    FMT: at 18:36 at ANOKO
    heading: 188° true track
    speed: 488 kn
    fuel range from ANOKO: 2760 nm
    distance from ANOKO to ARC6 crossing: 2722 nm
    distance from ANOKO to ARC7 crossing = fuel range + 2 min (1min APU + 1min SDU) = 2760 + 16 nm

  276. Peter Norton says:

    Victor Iannello: “@Peter Norton: At @Warren Platt’s request, I presented results for a flight at FL350 that ends at 38.5S latitude.”

    yes, sorry I have 2 days worth of comments to catch up, including your entire conversation with Warren Platts.

    Victor Iannello: “@Warren Platts: I give up. When a model doesn’t predict what you want it to predict, you keep questioning the model, […] you are outside of the realm of reality. […] I won’t continue to try to persuade somebody that really doesn’t want to be persuaded.”

    oh … seems like that’s very bad timing to come with this (probably, will have to read) similar issue :/

  277. Victor Iannello says:

    @Peter Norton: I don’t know what Hardy is doing. When he makes claims like “488 knots is the cruising speed of a triple seven”, it doesn’t give me a warm feeling. I don’t have the time or patience to once again sit through his videos and watch somebody draw lines on a flattened Google Earth image and measure distances with a ruler and expect any reasonable level of accuracy.

    Also, don’t forget that many of our first estimates for the crossing was around 37.5S latitude. That area was searched to +/- 50 NM width and deemed much less likely after drift and fuel models suggested a more northern terminus, in addition to the final BFOs which suggests the plane was in a steep descent at that time.

  278. Peter Norton says:

    Victor Iannello: “Yes, climbing to a higher altitude lengthens fuel endurance for 2 reasons: The fuel flow is reduced for the same temperature offset from ISA, but also the ISA offset is reduced. However, it takes additional fuel to climb to the higher altitude. Now during a normal flight, that extra fuel is “recovered” due to the longer idle descent from the higher altitude. In our case, the benefit of reduced flow at the higher altitude has to be balanced against additional fuel consumed to get to the higher altitude.”

    I’m aware of that in principle, but I wasn’t sure if the net effect was positive or negative and I just wanted to put it out there in case this actually lengthens the maximum distance to ARC7.

    Victor Iannello: “@DrB has studied fuel constraints versus crossing latitude on the 7th arc much more closely than me, including a very systematic investigation across the parameter space. Perhaps he’ll weigh in.”

    I’d be very interested.

    What factors (other than a gradual climb to higher altitude >FL350, if applicable, as discussed above) have been identified so far as range-extending ?

    TBill mentioned “2 ways to increase thrust, or increase range, are to disengage the Right and Left engine generators, and turning off of the bleed air into cabin” and other possibly range-extending electrical configurations.

    Are there other range-extending factors ?

    What about an FMT as early as 18:25 ?

    If anyone has ever done this calculation, what did you calculate as the southernmost latitude on ARC7 that could have been reached assuming the most favorable factors and conditions for maximum range:

    • gradual climb (if net effect is positive, as discussed above)
    • FMT at 18:36 (Hardy) or 18:25
    • MRC instead of LRC
    • bleed air off (has to breathe O2? probably unrealistic?)
    • turning off whatever could extend the range (TBill mentions engine generators, anything else?)
    • other range-extending factors ?
    • stay very conservative (i.e. include safety margins) for all values including fuel consumption, wind, etc.

  279. TBill says:

    @Ben S
    Alls we really have right now is Arc7 in the SIO and nothing is ruled out definitively. I have previously guessed 5% chance it is north of 20 South but I just made up the number.

    We have the persistent rumor that diversion to Xmas or Cocos might have been the flight plan, but nobody knows if that is fact or fiction or Plan A, and nobody knows why that diversion did not happen (Plan B or failed Plan A). MH370-Captio.net proposal is based on the Xmas concept, and that hijackers crashed short of Xmas island due to inexperience.

    I believe I am aligned with Victor that we need some extra new proof or verification that the plane might have gone north of 20 South, if someone expects us to give more serious consideration of it. This is because it is hard to believe that what looks like a straight flight path on the satellite data, was actually a highly complicated set of manual turns, altitude and speed changes that just happen to result in the perfect set of BTO/BFO that would look like straight flight.

    Thnak you for posting the hanger locations of the defunct MAS B777 aircraft on JW. That was interesting to me.

  280. Peter Norton says:

    Victor Iannello: “I don’t have the time or patience to once again sit through his videos”

    that’s why I sat through the videos and provided the data above

    > final BFOs which suggests the plane was in a steep descent

    Yes, although this is not specific to Hardy’s area but a factor any glide-scenario has to explain (in which case +50nm is not enough).

  281. @All

    We have extended the time period of the “collected” actual meteo data and produced a new, more complete debris drift analysis for flaperon type items (debris). The actual meteo data are in-situ measurements made on the very days during the drift.

    It demonstrates that location (12S;107E) on Arc-7 is definitely a possible origin of the debris drift.

    To our views, this shows that unfortunately because of all debris drift analyses possible found crash locations, all studies (incl. ours) do NOT bring sufficient discriminating evidences to make a decision on where to search for the wreck.

  282. Peter Norton says:

    Victor Iannello: “draw lines on a flattened Google Earth image and measure distances with a ruler”

    Agreed, the method would have to be applied to the figure of the Earth. But it’s possible Cpt. Hardy has done this. He says a couple of times in the video that he has done the calculations beforehand, so what he shows in the video is a proof of concept only. I think it’s just a visual presentation adapted for a wider audience. (Otherwise I have no idea of the magnitude of error this would introduce.)

    > expect any reasonable level of accuracy

    This leads directly back to my comment above.
    We have to decide if the entire method is wrong or just its accuracy.
    – If the method is wrong and can’t be improved, it must be thrown out the window.
    – If only the accuracy is in question, then the whole model cannot be thrown out the window on for this reason alone but instead the accuracy needs to be increased.

  283. Victor Iannello says:

    @Peter Norton: What is “Hardy’s model”? Matching some (but not all) of the ping arcs with a straight path and assumptions about speed? That’s been done many times by people here with much more precise analyses that include uncertainties in parameters. For instance, I showed in a previous post that there are many great circle paths that match the BTO and BFO data with reasonable error.

    Yes, MRC increases range versus LRC, but that occurs at a slower airspeed, meaning a shorter distance [since we are still constrained by the time of fuel exhaustion].

    Let’s assume there was enough fuel to reach as far south as 40S latitude. Let’s also assume that drift models are all wrong and debris could have traveled from an impact point that far south and reached the shores of East Africa at the time that debris was recovered. Let’s again assume a long glide occurred after pull out of a steep descent. (All three are very unlikely assumptions, in my opinion.) We still have to decide the latitude and range to search. How do we without bias choose that latitude range? Do we just search along Hardy’s path because you like his video, knowing that a path that crosses near 35S is a better fit to the BFO data and the drift data? Or knowing that a single pilot input to the autopilot’s directional settings will make his assumptions (or anybody else that assumes no pilot inputs until fuel exhaustion) invalid? It would be unlikely that anybody would pursue this search strategy because either you have to follow a “hunch” to limit the size of the search area, or you are left with an unmanageably large search area.

    There’s been a lot of talk about a glide followed by an attempted ditching to maximize the distance into the SIO and to minimize the debris. If there was a real attempt to successfully ditch, any pilot would want to do that with flaps down and with engine thrust available. On the other hand, the evidence we have is consistent with flaps up and fuel exhaustion, not to mention the steep descent at 00:19 which, is not consistent with maximizing the glide distance.

    In my opinion, our best bet is to search further north along the 7th arc. That said, it’s still a bet, and it may not be successful.

  284. TBill says:

    @Victor
    The big question is if OI would come back to search.

    Feels like, to me, that Malaysia needs to finish the proposed 20-25 South +/-22nm of Arc7 to call it “done” as far as best effort on their part. I say this even though I am not an optimist for Arc7 closeness, but I am willing to give the other investigator’s theories/pins a chance.

    Obviously it gets very contentious beyond that.

  285. Victor Iannello says:

    @TBill said: I am willing to give the other investigator’s theories/pins a chance.

    That’s not a decision you can make. Or any of us. The ball is in OI’s court to decide what they wish to do.

  286. DennisW says:

    @VictorI

    Given the success of the Argentine search, it seems unlikely that OI would attach significant marketing value to finding MH370. Their place in the world is secure for the moment. My guess is that we will not see an unfunded MH370 search undertaken or even a search on the same terms as the previous OI MH370 effort.

  287. Ed says:

    @ Peter Norton
    There is this brilliant guy GlobusMax (similar to the brilliant guys here) who has a blog and his endpoint is S40.1743E84.6945. He had formulated his waypoint hypothesis in 2014.His 23 rd Jan., 2018, post titled “Could MH370 have reached the Waypoint Hypothesis Location?” is interesting in that he has quoted Dr.Ulich’s fuel model, VI’s Great Circle analysis. An interesting quote from his above post -” there’s about a 5% chance MH370 makes it to the observed point with air conditioning packs off, and a 0.5% chance if the air packs are on. Statistically, it can work, time-wise, but chances are low. We are essentially 5.5 to 11.5 minutes short on a 7.5 hour flight, or 1-2%, but when error is accounted for, it could fit exactly.”

  288. Peter Norton says:

    @Ben S:
    re: spontaneous SDU reboot / non-human explanation

    I might miss something, but the SDU can only lose power
    (1) by accident (you mention fire)
    (2) by human intervention
    right ?

    Having depowered the SDU, a fire would not be able to repower it.
    And power removed by a human would need human intervention to restore power.

    So how do you see power being restored to the SDU without human intervention ?

  289. Peter Norton says:

    from the article above:

    « Director of special projects David Gallo said he was not convinced about the seventh arc hypothesis but his approaches to Malaysia and Boeing were ignored. “No emails or phone calls were returned,” he told The Australian. »

    Does anyone know what he doubts exactly and why ?

    I have come to like David Gallo over the years, he always made a very trustworthy and solid impression on me. That’s why I was a little surprised to see him voice those unspecified doubts about the Inmarsat data and apparently also the debris. I tried to find if he had elaborated on that somehwere, but all I could find was this twitter conversation between him, Victor Iannello and others, where he doesn’t really explain anything. Do you know more about this ?

  290. Peter Norton says:

    « his approaches to Malaysia and Boeing were ignored. “No emails or phone calls were returned” »

    This is also a little surprising. Given his high-profile role in the search for AF447, one would think that they would at least talk to him, if nothing else.

  291. Peter Norton says:

    Wall: “the search took place in the 39 area (39.6), why did that happen?”
    Victor Iannello: “[CSIRO drift model… crude fuel model…] There was also a lot of pressure to search Hardy’s spot.”

    Although it keeps being repeated, Hardy’s spot was not searched. We’ve been over that before, I have posted the screenshot above (green overlay area). Hardy’s area extends from ARC7+32nm to circa ARC7+85nm (or ARC7+90nm in case his margin of error is not included in the screenshot).

    Don Thompson: “1969 the US Navy spent 4 months searching out 144mi² to find the USS Scorpion wreck, today 1 Hugin AUV can survey that area in half a day. […] Ocean Infinity operating 8 AUVs simultaneously.

    Based on these numbers 8 AUVs would search Hardy’s 7000 km² area in 1 single day …

  292. Warren Platts says:

    @Victor Ianello: >”Yes, MRC increases range versus LRC, but that occurs at a slower airspeed, meaning a shorter distance.

    >”How do we without bias choose that latitude range? Do we just search along Hardy’s path because you like his video, knowing that a path that crosses near 35S is a better fit to the BFO data and the drift data?”

    Well, actually, with the arcs provided by @airlandseaman, on the 189T path I have charted, the distance from 4th Arc to 5th Arc is about 486 nm, that, for 1 hour, is about M0.84, that would be almost a perfect LRC (long range cruise) speed.

    But the distance from 5th Arc to 6th Arc is 678 nm. Since that is over a time of 1.5 hours, that works out to an apparent ground speed of 453 knots (~M0.78 @ FL350, assuming zero wind). Of course given a 25 nm positional error bar around the arcs, one can fudge these distances and speeds to get them more even, but at face value, the arcs indicate a slowing down consistent with MRC (maximum range cruise) mode towards the end of the flight on this alleged path. Combine that with some throttling back and gradual descent, it seems highly probable imho that a 38.5S to 38.7S, powered crossing of 7th Arc could be achieved.

    Why should we think this latitude is any more probable than say 35S? Well, the answer is that Hardy’s (an experienced 777 pilot) path is about the only LNAV flight possible at more or less standard cruising speeds. That is, it is about the only “normal” flight path consistent with the BTO arcs, if it makes any sense to say there is anything normal about MH370. You can call that a hunch if you want. However, such a flight path is consistent with the observed behavior of the aircraft in the observed, initial phase of the flight. That is an extra bit of information that maybe should not be ignored.

    The preceding is of course all speculation on my part.

  293. Warren Platts says:

    Here is an excellent, very understandable article from Boeing on cruise speed fuel conservation strategies in case anyone is curious:

    https://www.boeing.com/commercial/aeromagazine/articles/qtr_4_07/article_05_1.html

  294. Victor Iannello says:

    @Warren Platts said: I have charted, the distance from 4th Arc to 5th Arc is about 486 nm, that, for 1 hour, is about M0.84, that would be almost a perfect LRC (long range cruise) speed.

    LRC is not a constant M0.84. That’s the maximum speed. For LRC, the speed decreases as the weight decreases, just as for MRC.

    Well, the answer is that Hardy’s (an experienced 777 pilot) path is about the only LNAV flight possible at more or less standard cruising speeds.

    There are a LOT of paths consistent with LNAV mode (great circle paths), as I tried to demonstrate in this post.

  295. Peter Norton says:

    Warren Platts: “My approach was to enter every waypoint in the Indian Ocean I could find into my google earth”

    Isn’t there a .kmz file with all waypoints available ?

  296. Peter Norton says:

    ———-
    TBill: “The whole problem with finding MH370 is the assumption that straight, passive flight is somehow the more likely, more rigorous base case, and active pilot is somehow a weak speculative case. But passive, straight flight is perhaps the greatest speculation of all.”

    Richard Godfrey: “@Warren, you stated “The whole problem with finding MH370 is the assumption that straight, passive flight is somehow the more likely, more rigorous base case, and active pilot is somehow a weak speculative case. But passive, straight flight is perhaps the greatest speculation of all.” Complete and utter nonsense!

    Warren Platts: “No no. That was Tbill who said that.”

    Richard Godfrey: “The paragraph I quoted from “you” was not attributed to anyone else. It appears that almost everything you write is misleading.”
    ———-

    @Richard Godfrey: Warren is right, your quote is from TBill (see links above). And personal attacks are not helpful.

  297. Peter Norton says:

    Victor Iannello: “@Warren Platts: I give up. When a model doesn’t predict what you want it to predict, you keep questioning the model, even when your path requires the results to be well outside reasonable error bounds.”

    @Victor, I’ll let Warren speak for himself, but I don’t think his intent was to question the model per se, but to say that any model has a certain margin of error and to find out its order of magnitude (“temperatures predicted by the weather model are uncertain: there must be an error bar to the temperature estimates […] Then there is the sensitivity of the fuel efficiency model. How many degrees K deviation would be required to get the fuel efficiency model to close with the flight path model?”“).

    I think I understand his motivation. As I said before, I never felt strongly about the south, but this collective wholesale dismissal just creates this kind of reaction. I guess Warren, Wall and others (including myself) would feel less strongly about southern latitudes or Cpt. Hardy’s model if both were greeted with more respect and carefulness: e.g.
    – stop misquoting latitudes
    – stop saying Cpt. Hardy’s area was searched, when this is not true and has been pointed out several times
    – don’t reject the question of what you think is the southernmost latitude and what you think is the (overall) margin of error: S xx.x° ± x.x° .

    These questions have been sidestepped. However, your fuel calculations for Warren were certainly helpful. Can we please continue with a/this constructive approach?

  298. Peter Norton says:

    With regards to the second question: “What is the overall margin of error“:

    Is the correct method to use best estimates for all factors (wind, temperature, fuel consumption, etc.) and to just add a margin of error on the end result ?

    Or could these errors be cumulative in a worst case scenario, so that for each factor the worst (most range-friendly) case must be assumed to be on the safe side and to be certain that the plane could not have flown beyond ?

  299. Warren Platts says:

    > Isn’t there a .kmz file with all waypoints available ?

    Not that I’m aware of. I guess I could try to make one if anyone is interested.

  300. Don Thompson says:

    @Peter Norton,

    1/ Hardy is not a British Airways flight crew member, he flies for THY.

    2/ I do have it on authority that in Oct 2015 Hardy prescribed two impact locations, about 20km apart. These two locations were within the boundary of the ATSB’s contracted deep ocean search. If you wish to contend that, take it up with Hardy himself.

    3/ Apologies, I was much too imprecise about potential coverage of the OI Hugin fleet. The estimate for daily coverage during Ocean Infinity’s 2018 MH370 search was, per AUV, 140km² per 24hrs (launch, recovery, descent & ascent imply variable non-productive time per mission), so the USS Scorpion search might have been completed by two typical AUV missions of 48hrs.

    4/ Again, Hardy’s revised conclusion/location following Fugro’s coverage of the Oct 2015 positions is not openly documented. Best that he describes what assumptions are involved in that revision and what his error bounds might be.

  301. Victor Iannello says:

    @Peter Norton says: First, you better take a deep breath because you are throwing around statements that are not true, and I have limited patience for that.

    Although it keeps being repeated, Hardy’s spot was not searched. We’ve been over that before, I have posted the screenshot above (green overlay area). Hardy’s area extends from ARC7+32nm to circa ARC7+85nm (or ARC7+90nm in case his margin of error is not included in the screenshot)…Although it keeps being repeated, Hardy’s spot was not searched.

    If your criterion is that a path associated with a particular crossing of the 7th arc is “not searched” unless the search extends to +90 NM, then NO spot has been searched. A possible glide past the 7th arc is not particular to Hardy’s path. On the other hand, 38.5S latitude has been searched to 50 NM past the 7th arc, which is farther than any other part of the 7th arc.

    stop misquoting latitudes

    I can’t keep track of the latitudes that you and @Warren are proposing. In any event, the last fuel calculation I performed was for a 7th arc crossing of 38.5S latitude. I don’t believe that is “misquoted”.

    – don’t reject the question of what you think is the southernmost latitude and what you think is the (overall) margin of error: S xx.x° ± x.x° .

    I gave you some insight into the error when I provided the percentages for an LRC flight at FL350 and ISA+10K. Here it is again: “However, the average engine PDA is 1.5%, and the extra fuel flow due to the ISA+10K temperature is 3.4%, which means at LRC speed, the fuel quantity is short by about 3.4% + 1.5% – .8% = 3.1%.” That means the LRC fuel model has to be 3% wrong, which is unrealistically high.

    @Warren asked me about the required fuel for M.84, FL350, and ISA-20K. I explained that ISA-20K is not consistent with the conditions on the day of the flight. For the path crossing at 38.5S, the average temperature offset at FL350 is closer to +9K. As I said in an earlier comment, “At an average temperature ISA temperature offset of +9K after 18:39, an average PDA of 1.5%, and speed of M.84 at FL350, the fuel tanks run dry at 23:46, or about 29 minutes sooner than we estimate based on the time of the final log-on. That’s not even close to being an acceptable path.” Said another way, the fuel model would have to be wrong by 6.8% for this path to be acceptable.

    These questions have been sidestepped…As I said before, I never felt strongly about the south, but this collective wholesale dismissal just creates this kind of reaction.

    Sidestepped questions? Wholesale dismissal? Based on the comments both of you are making, I’ve given his path more serious attention than either you or @Warren. The reason that there is not a lot of support for that area is because it doesn’t agree with fuel models (by a large margin), by drift models (by a large margin), or by previous search results (out to 50 NM from the 7th arc). You might disagree with all those reasons, and if you do, you need to refute with facts, but to claim that his path has been dismissed out of hand is incorrect.

  302. Peter Norton says:

    @Don Thompson:

    ad 1: I don’t know. All news articles refer to him as “British Airways pilot”, “senior British Airways captain” or “former British Airways pilot”. That’s why I referred to him this way.

    ad 2: I have stated multiple times (including directly to you I believe) that I refer only to his 3-part video series.

    ad 3: re USS Scorpion: My intention was not to second-guess the numbers. All I wanted to say is that the 7000km² area is quite small and it wouldn’t take long to search.

    ad 4: agreed

  303. Peter Norton says:

    @Victor Iannello:
    I have no desire for an argument. All I said is true. I’ll let the facts speak for themselves:

    (1) Hardy’s area is the green area in the screenshot I posted above 5 days ago. It never was fully searched.

    (2) re: missstating Hardy’s latitude, just 2 examples (there were others):

    • Victor Iannello, 4 Dec 2018 at 9:22am: “there is not enough fuel to reach Hardy’s 39S

    • Peter Norton, 4 Dec 2018 at 7:01 pm: “Since it was mentioned a couple of times that Hardy’s area is beyond fuel range, in order to avoid misunderstandings […] Crossover of the 7th arc was not at S39 but circa S38.6

    • Victor Iannello, 5 Dec 2018 at 7:57 am: “Hardy’s path falls in that category […] To reach 39.5S latitude requires a speed of M0.84 at FL350. That’s faster than LRC. Despite what you keep saying, by a large margin, there was not enough fuel to reach 39.5S latitude.”

    (3) Still have not seen S xx.x° ± x.x°

    Can we please let the matter rest?
    Item 3 would be interesting though.

  304. Victor Iannello says:

    @Peter Norton: No, here are the facts. What you said is NOT true. Multiple times, Hardy’s path has been associated with 39S. Regarding 39.6S, you said that we should consider searching south of 39.6S. Here it is:

    Peter Norton says:
    December 4, 2018 at 3:00 am

    Warren Platts: re: “The fourth one is namely south of 39.6 S.” “Agree that remains a live possibility.”

    Victor Iannello: “I believe the prospects of MH370 crossing the 7th arc at 39S latitude and impacting at 50+ NM from the 7th arc as very unlikely, but not impossible.”

    ———-
    I agree with others here that this 4th scenario should be included as one that merits thought. At any rate by the standard that many here (I think including you, Victor) deem scenarios 2 and 3 highly unlikely.

    If there is confusion about what latitude you are referring to, that confusion was not created by me, despite your false accusation. I’ve done my best to understand what you are talking about. I revised my calculations to 38.5S to try to accommodate your proposal.

    And by your criterion of searching out to 90 NM past the 7th arc, NO area was searched. I’ve explained that already. So stop complaining that Hardy’s spot was not searched.

    Work with somebody else on the fuel calculations. When you’re done, I’ll critique the results. I’m reluctant to do more calculations for somebody that whines about the help I am giving.

  305. airlandseaman says:

    Peter Norton: As I understand it, you were once a strong proponent of an Inmarsat/US Gov’t conspiracy theory involving fabrication of the Inmarsat log and the murder of Inmarsat employee James Stuart Fairbairn, all designed to mislead searchers. Is that correct? If so, do you still hold to that theory? (apologies if this was a different Peter Norton)

  306. Warren Platts says:

    @Victor: “At an average temperature ISA temperature offset of +9K after 18:39, an average PDA of 1.5%, and speed of M.84 at FL350, the fuel tanks run dry at 23:46, or about 29 minutes sooner than we estimate based on the time of the final log-on. That’s not even close to being an acceptable path.” Said another way, the fuel model would have to be wrong by 6.8% for this path to be acceptable.

    So your fuel calculation was for constant M0.84, rather than for LRC or MRC?

    With the revised the arcs, Hardy’s constant speed model breaks down: the ground speed from the 5th to the 6th arcs does appear to slow down significantly, so perhaps that could make a difference, especially if the aircraft switched to MRC sometime after 22:40?

    As for LNAV, I agree there are many great circle paths, but as far as I can see, there is only one path that uses established waypoints and goes at more or less cruising speeds. Note that this is also the only (established) waypoint path the ATSB considered (MUTMI RUNUT)for the same reasons I mentioned, that the aircraft was apparently following established waypoints during the initial phase. They thought that path did not correlate as well with BTO’s/BFO’s as some other paths, although I speculate that it is not entirely outside the range of possible errors.

    Also thanks for the link. Note that paths through this SW zone at about 38.5S degrees leads to a cluster of Antarctic stations: Mawson, Progress, Zhongshan, and Davis. (I read off the internet somewhere where someone had suggested the hand-entered waypoint 69,69, that is in the middle of this cluster in Antarctica, might be significant if the pilot was mad because his friend had been thrown in prison for sodomy [pure speculation]…)

    Consider the fact, from your previous article, that the pilot had deleted the simulated flight to 45S,104E, indicating an apprehension that his computer might be searched. These days, one can never be sure that a deleted file cannot be resurrected by clever forensics. Knowing that, one can speculate that might motivate different targets in the Antarctic or wherever.

  307. airlandseaman says:

    Warren Platts: Why would any pilot plot a course via MULTIPLE way points. If you are planning a route to the SP, for instance, over water obviously, then why add any complexity? One WP would do it.

  308. Wall says:

    @ all

    Does someone know how the 7th-arc can be drawn in google earth?

  309. Victor Iannello says:

    @Warren Platts: So your fuel calculation was for constant M0.84, rather than for LRC or MRC?

    Yes, that was your request. I asked for you conditions. You said: Might as well keep it simple: Mach number = 0.84; altitude = FL350; temperature = ISA – 20C constant over time. I explained to you that ISA-20K is very far from reality. I used ISA+9K, which is much closer to a variety of paths at FL350.

    As for LNAV, I agree there are many great circle paths, but as far as I can see, there is only one path that uses established waypoints and goes at more or less cruising speeds.

    BEDAX and the South Pole are “established” waypoints. That path crosses the 7th arc near 34.3S latitude. We also have Car Nicobar to McMurdo Station, as well as other LNAV paths.

  310. airlandseaman says:

    Wall: ARCs as reported above: http://bit.ly/2RFGybn

  311. TBill says:

    @Warren
    “someone had suggested the hand-entered waypoint 69,69”

    …that would be entered as “6969S” (69S 69E) and it seems to represent a great circle path that goes quite closely thru ISBIX MUTMI RUNUT and about 39.2S on Arc7. It is almost a mirror image – towards the west – of the simulator path DOTEN to NZPG (or “78S67”) – towards the east.

    Other oceanic waypoints have been suggested as the anchor for the 38S path, but I forget the exact coordinates.

  312. TBill says:

    @Warren

    P.S. Re: “It is almost a mirror image…”

    The center point of the mirror image is approx. MUTMI, whereas I am suggesting MH370 went MUTMI easterly towards BEBIM/NZPG (representing the actual simulator path) and you are suggesting it went MUTMI RUNUT westerly towards 6969S. The fact that two paths are near mirror image might be why BFO’s support either path. The easterly path requires a slower speed/active pilot I assume.

  313. Victor Iannello says:

    @Peter Norton asked: I tried to find if [Gallo] had elaborated on that somehwere, but all I could find was this twitter conversation between him, Victor Iannello and others, where he doesn’t really explain anything. Do you know more about this ?

    He has made several quizzical comments without further explanation expressing his doubts about the search. I’ve encouraged him to say more, but he does not. He also made the statement that he thinks a new team should be formed to study the matter. I encouraged him to organize the team himself, since a team with the required skill set is unlikely to self-assemble. (I would not request to be part of that effort.) Again, no response.

  314. DrB says:

    In response to requests, HERE is a link to my MH370 publications, including my fuel flow and endurance models, and my analyses of speed and altitude combinations which can reach the 7th Arc and run out of fuel at the correct time. Just click on any link in the publication list to see and download that item.

    FUEL AND ENDURANCE MODELS (#28)

    Publication #28 is the fuel flow model and the endurance model (V5.6 from November 2017). The assumptions for these models are listed in the Notes in the first worksheet called Fuel Flow Model. There are 7 parameters that must be set to determine the fuel flow. The second worksheet is the Endurance Model, and it predicts the time for MEFE and the air miles traveled.

    PREDICTED MEFE TIMES (#29)

    It is not necessary to run the fuel and endurance models since I have already done that for the environmental conditions appropriate for MH370. The predicted MEFE times are presented in Publication #29. This contains a plot of predicted MEFE times for various speed control modes as a function of flight level.

    From Publication #29, the acceptable endurance solutions are as follows:
    1. MRC (or possibly ECON with a low value of the Cost Index << 52) at FL350-400
    2. HOLDING or 250 KCAS near FL400
    3. HOLDING or 250 KCAS near FL75
    4. 1-Hour HOLD at FL210, followed by LRC at FL300-400, both with Air Packs OFF.

    Solutions #2-3 involve HOLDING speeds at very high and very low altitudes, and these do not seem capable of matching the BTOs/BFOs.

    Solution #4 is the LRC case. With LRC, several fuel-savings measures are needed to extend the endurance until 00:17:30. First, a 1-hour HOLD at much lower altitude can save some fuel (but not quite enough to match exactly the observed MEFE), and the air packs also need to be off for both the HOLD and the subsequent LRC dash in order to equal or exceed the observed endurance.

    The most likely solution for the endurance, in my opinion, is #1 – MRC at typical cruising altitudes. This setting also provides the maximum range after the FMT.

    RANGE

    Publication #32 presents the predicted range and endurance of various speed control modes as a function of flight level. Figure 1 shows the predicted ranges as well as the calculated ranges to various latitudes on the 7th Arc. Thus, one can interpolate this figure to find the most southern point of the 7th Arc reachable with each speed/altitude combination.

    Here are my findings from this work:

    1. HOLDING INOP (Flaps Up) and LRC INOP have excess endurance at all altitudes.
    2. LRC has inadequate endurance at all altitudes.
    3. A 1-hour Hold at FL210 followed by LRC has marginally acceptable endurance from FL200-400 with Air Packs ON and acceptable endurance with Air Packs OFF.
    4. A 1-hour HOLD at FL210 followed by CI = 52 has acceptable endurance from FL300-400.
    5. MRC has acceptable endurance from FL220-270 and from FL350-400.
    6. CI=52 has acceptable endurance from FL320-400.
    7. HOLDING has acceptable endurance from FL60-90 and near FL400.

    To summarize, MRC at cruise altitudes fits the endurance and can also fit the arcs well. LRC (and M0.84) burns too much fuel, A very efficiently flown 1-hour HOLD near FL210, followed by LRC, is needed to reach marginal acceptability in terms of endurance. Turning the Air Packs OFF in addition to the HOLD is needed to achieve a comfortable LRC match to the endurance.

  315. Victor Iannello says:

    @DrB: So there is no confusion, what do you think is the likelihood that the aircraft crossed the 7th arc near 38.6S latitude, which is where Hardy predicts the crossing to be? (If you have been following the thread, I don’t view this as probable.)

  316. Ventus45 says:

    @DrB:
    Link to document 42 does not work.

  317. DrB says:

    @Ventus45,

    Thanks. There is an extra space in the URL. I’ll fix it tomorrow.

    @Victor Iannello,

    I would say in my opinion it is extremely unlikely that MH370 reached the 7th Arc near 38.6S for the following reasons: (1) no auto-throttle speed control mode, taking into account the winds, provides an acceptable match to the BTOs, (2) the BFO errors exceed what I consider acceptable (7 Hz), and (3) the best speed fits do not match the known endurance. While it is true that a reasonably good fit to the BTOs is possible with a straight path, it requires a constant ground speed and is not flyable with the auto-pilot, plus there is insufficient fuel.

  318. Peter Norton says:

    @Ed: Thank your for the hint.
    For reference, here is the link:
    https://globusmax.wordpress.com/2018/01/23/could-mh370-have-reached-the-waypoint-hypothesis-location

    @Warren Platts:
    re: Isn’t there a .kmz file with all waypoints available ?
    “I guess I could try to make one if anyone is interested.”
    Thanks for the offer, but I would be surprised if there wasn’t a file/list with coordinates officially provided somewhere. I have searched for that 3 times in past years but came up empty. If someone knows where to find that, please let us know. Thank you.

  319. Peter Norton says:

    From airliners.net about 737 MAX / LNI610 crash:

    « Some very interesting points which might point towards deliberate underplaying of MCAS to avoid simulator training:

    http://www.wsj.com/articles/behind-boeings-decision-to-omit-details-on-safety-system-in-lion-air-crash-from-manual-1544025884

    “Engineering, training and other experts inside Boeing had differing views on the precise language to be used in manuals. People familiar with the process said there was a sharp focus on one point: avoiding added simulator training.

    The decision to omit the new control system from manuals has put a Boeing design principle at the center of a probe into a fatal airliner crash for the first time in more than two decades. It has sparked public scrutiny of a typically behind-the-scenes process and threatens to tarnish Boeing’s reputation for safety and its tradition of prioritizing pilot authority over automation.

    Former Boeing and current airline and government officials said there was a strong push to keep 737 MAX training to a minimum—a common goal for the introduction of new models. One former Boeing official recalls a colleague expressing concern about keeping their job if regulators rejected the company’s proposed guidelines. The program was eventually approved.” »

  320. Peter Norton says:

    @airlandseaman: re: Fairbairn – I think there is a fundamental misunderstanding if you believe that following leads and arguing a case is the same thing as being a proponent. As far as MH370 goes, I have never been a proponent of anything. I don’t have any pet theory if that is your question. I have seen several different, well-argued theories to all of which I am sympathetic, although they are obviously mutually exclusive. I try to keep an open mind. This is the approach many MH370 observers have adopted perforce, given the somewhat contradictive evidence. I think this is what many of us have struggled with when formulating theories: It’s very hard to account for all the evidence. It’s as if you can align some of it, but then there are always some items left that are hard to reconcile, as mash, Victor and many others have pointed pointed out before. As far as Mr. Fairbairn’s death is concerned, I don’t recall the details, as it occurred more than 4 years ago, but it connected to other strange dots, which, taken together, I thought deserved to be questioned. That’s what I tried (without much success).
    I hope I have answered your question. May I ask in return, why you are asking ?

  321. Peter Norton says:

    Don Thompson: “Best that he [Hardy] describes what assumptions are involved in that revision and what his error bounds might be.”

    If anyone has his contacts I would ask him for clarification.

  322. Peter Norton says:

    @DrB, @Victor Iannello, @Warren Platts, @Wall:

    With regards to the question of the southern latitude limit:

    I have tried to find conclusions on that subject matter. Here is what I found yesterday:

    2017-03-31 Victor Iannello: “A case could be made that after the flight to the south began, MRC (ECON with CI=0) is the most logical speed for flying MH370 into the SIO. The fact that the plane was flown until fuel exhaustion means the intention was to fly as far as possible, and MRC satisfies that. Earlier parts of the flight, before the turn to the south, might have had different constraints.” Incidentally, I reached the same conclusion regarding the 3 different phases of the flight (see above).

    • 2017-05-15 DrB: “At MRC from 18:36 onward at FL350, fuel exhaustion is 5 minutes early.”

    2017-10-11 DrB: “MRC can fly until 00:18 at FL370, and it can go a few minutes longer at slightly lower altitudes. The southern latitude limit is 38.2S assuming a FMT at 18:29, 1.5% PDA, and a great circle southward. This is further south than my previous prediction, made with somewhat different assumptions and a less accurate latitude limit method, and it is close to the Boeing prediction (with as yet undisclosed assumptions). […] There may still be (hopefully small) errors.”

    • 2018-05-25 DrB: predicted locations at 7th arc”

    S limit:
    range limit (aircraft performance): ATSB/Boeing: S38, Ulich: S39
    best fit to BTO and BFO: Godfrey S39.5

  323. Peter Norton says:

    @DrB:
    I was about to say that to my eyes this data seems to allow S38.6 as latitude limit of the ARC7 crossing, but your comment now seems to suggest otherwise.

    What do you deem the southern limit and what do you estimate is the margin of error ?

    If “MRC can fly until 00:18 at FL370 to a southern limit of 38.2S”
    – can’t it reach 38.6S at higher FL within the time constraints
    (given higher MRC speed at higher FL, if I understand correctly) ?
    – isn’t 38.6S within the margin of error anyway, even at FL370 ?

    And a question about the altitude:
    I see you and Victor use FL350 very often as reference value for your calculations. Why is that? Is there something particular about this flight level?

    And why don’t we have to assume that MH370 continued above FL430 as this was its altitude over Kota Bharu calculated by Michael Exner with a reliable method ?

  324. Peter Norton says:

    [18:36]

    Last question, if someone can clue me in:

    Cpt. Hardy uses 18:36 as timing for the FMT in his video.
    I also see DrB used 18:36 eleven times here.

    Is there anything special about this particular time ?

    In search for an answer, it appeared to me, that 18:36 seems to have been used by the DSTG as timing for the FMT. Is this probably the reason why Cpt. Hardy used 18:36 for the FMT ?

  325. Victor Iannello says:

    @Peter Norton: I believe that when @DrB proposed that 38.2S latitude is possible at MRC speeds, he did not fit the BTO data. I remember objecting to the conclusion at the time. I don’t see how it’s possible to reach that latitude at MRC speed AND hit the ping arcs.

  326. Peter Norton says:

    @Victor Iannello:
    First of all, thank you for sharing your thoughts about David Gallo.

    In reply to your comment:
    Victor, I would like to maintain a good relationship with you, so I would like you to know that although I quoted you as an example, my plea was intended to be general, not directly to you. I apologize if you felt personally attacked, that was not my intention at all.

    With regards to S39.6, I think there might have been a misunderstanding, because Wall, Warren Platts and myself were initially discussing the idea that MH370 might have been missed in the south, generally speaking. And we were subsequently also specifically talking about Cpt. Hardy’s method and area. With the latter being a special case of the former. Personally, I have never associated Hardy with S39.6. Maybe others did, I don’t know, and my comments would apply to them, too (as I said, my comments were directed to anyone). I only found 2 quotes by you situating Hardy’s area at S39 and S39.6, with the second time occurring after I already pointed out the error. This in combination with repeated claims by others that his area (marked in green in the screenshot) was searched, which is not true, got me frustrated, as you saw, and as is comprehensible I think. Maybe others didn’t refer to the video, although I made clear countless times that I did. So maybe this was a misunderstanding, too.

    Please, Victor, with your allowance I hope this can be my last comment on this matter. Can we please let it rest? I don’t think anything productive can come out of it. I also don’t understand why this is such a problem. Some numbers were wrong, this can happen. Though honestly I would have rather expected a “thanks Peter for pointing that out” instead of the reaction I got.

  327. Peter Norton says:

    Victor Iannello: “I encouraged him to organize the team himself, since a team with the required skill set is unlikely to self-assemble. (I would not request to be part of that effort.)”

    In case you want to share: Just out of curiosity (given your passion for MH370), why wouldn’t you want to be part of such an effort (not even as an occasional adviser if time does not permit a more active role) ?

  328. Peter Norton says:

    Victor Iannello: “I believe that when @DrB proposed that 38.2S latitude is possible at MRC speeds, he did not fit the BTO data. I remember objecting to the conclusion at the time. I don’t see how it’s possible to reach that latitude at MRC speed AND hit the ping arcs.”

    Because MRC is too slow for the ping arcs, right ?
    But wouldn’t MRC be faster at higher altitudes (which we would have to assume anyway given the calculated FL430+ at Kota Bharu) ?

  329. airlandseaman says:

    Peter: What about the Inmarsat Conspiracy theory? Do you still believe that?

  330. TBill says:

    @Peter
    Re: 18:36 is approximate latest time an FMT turn south started to register the observed BFO at the 18:40 sat phone call. A turn takes about 3 minutes to execute. Whereas of course the early FMT is required for the 38-39 South theory. The alternate idea is so-called loiter, meaning that MH370 did not immediately turn south at 18:36.

  331. Victor Iannello says:

    @Peter Norton: Gallo is proposing that a new team assemble. In as much as I am associated by some as being part of the old team, I don’t qualify. I’d rather critique their work product than be part of the team that generates it.

  332. airlandseaman says:

    Peter Norton : Disregard my previous question. I located your response above at 07:41. Thanks for answering. Explanation accepted.

    I asked because I do not think it is at all helpful to even suggest such an outrageous possibility. If you know anything at all about the history of Inmarsat and its people, as I do, then you know there is ZERO chance anything remotely like what you suggested in 2015 ever happened. Such conspiracy theories raise questions about the writers motive. That is why I asked.

  333. Tim says:

    @All,
    It’s time to admit it is impossible to fine tune the exact flight path. Since the release(why did it take 4yrs) of the SIR it shows that this is an autopilot OFF flight. No fixed hdgs. No fixed altitudes. No nefarious pilot intentions. The Australians must be livid, as they based their whole search effort on the assumption this was an efficiently flown autopilot ON flight.

    If the arcs are correct, then it must lie further north up the 7th.

  334. Warren Platts says:

    @airlandseaman: Why would any pilot plot a course via MULTIPLE way points. If you are planning a route to the SP, for instance, over water obviously, then why add any complexity? One WP would do it.

    This sort of thinking–making assumptions about plausible mental states of the pilot–is exactly what I am trying to avoid by suggesting a strict behaviorist account that focuses solely on the observed behavior of the aircraft. After IGARI the pilot could have simply programmed in 45S104E (the apparent destination of Zaharie’s simulated flight to SIO, that just happens to be due south of IGARI) or NZSP (also due south) and been done with it. Instead, there was a complicated flight that was observed apparently involving several waypoints (IGARI ABTOK ENDOR OPOVI VAMPI MEKAR NILAM). Therefore, it would not be surprising, and indeed probably should be expected, that the latter phase of the flight might also involve multiple waypoints (speculation).

    Note that the simulated flight arguably included multiple waypoints: after the turn at point #3, the path flew over KETIV, our old friend MUTMI, followed by BEBIM, followed by some other waypoint, possibly a base in Antarctica. (The point #3 itself seems rather problematic: it is not a round number; perhaps it is an established waypoint that I am not aware of.)

    If we want to speculate about mental states, however, it could be that flying established air routes with your secondary radar turned off would be less likely to arouse the curiosity of any naval radar operators that happened to be in the SIO. But who knows? There is no telling what other people are thinking if you cannot ask them.

  335. airlandseaman says:

    Warren: I was basing my question on 50 years of experience as a pilot. NOBODY plots a course over open water, far from land, using multiple way points. There is absolutely no reason to do that. It simply makes no sense to a pilot.

  336. DennisW says:

    @Tim,

    It’s time to admit it is impossible to fine tune the exact flight path.

    That has been my position from the get-go. Now, however, we have an excellent opportunity to find the aircraft since the search boundary conditions have been refined by previous search efforts. There should be no lingering confusion on the optimal course of action at this point. Discussion of motive, Captain Hardy, dive/glide/dive, Inmarsat conspiracies,… are pointless IMO.

  337. Warren Platts says:

    @TBill: that would be entered as “6969S” (69S 69E) and it seems to represent a great circle path that goes quite closely thru ISBIX MUTMI RUNUT and about 39.2S on Arc7. It is almost a mirror image – towards the west – of the simulator path DOTEN to NZPG (or “78S[1]67[E]”) – towards the east.

    That is an interesting pattern I had not noticed before. Coincidence? 😉

  338. TBill says:

    @Tim
    “It’s time to admit it is impossible to fine tune the exact flight path.”

    Well your comment is in direct disagreement with the recent media articles on MH370 (MIT/Popular Mechanics) saying, based on @Vikings paper, that more could be done with path modeling.

    I actually tend to feel more could be done to pin down flight path possibilites- eg, perhaps a less contrained Bayesian approach – but I do not see too much potential for a serious effort unless Malaysia turned it over to FBI/NTSB or something like that, and that isn’t going to happen.

  339. Peter Norton says:

    @Tbill: re: “18:36 is approximate latest time an FMT turn south started to register the observed BFO at the 18:40 sat phone call. A turn takes about 3 minutes to execute. Whereas of course the early FMT is required for the 38-39 South theory.”

    Thanks for the quick summary. That’s very helpful.
    It’s interesting then that Cpt. Hardy uses 18:36 for the FMT. Judging from your comment an earlier one might better fit his area.

    @Victor: re: Gallo. Thanks, I can understand that. It’s off course a trade-off: expertise vs starting from scratch. The former would call for at least 1 IG member to be part of such an initiative. But judging from your earlier comment, there is no such initiative at the moment anyway.

    @airlandseaman: Why do you bring this up from 4 years ago sans current topical relevance. Do you intend to discredit me? I am open-minded and deem taboos and thought barriers never helpful. Twisting this into ill will is a low blow.

  340. Peter Norton says:

    Tim: “No nefarious pilot intentions.”

    How do you know ?

  341. Warren Platts says:

    @airlandseaman: I was basing my question on 50 years of experience as a pilot. NOBODY plots a course over open water, far from land, using multiple way points. There is absolutely no reason to do that. It simply makes no sense to a pilot.

    Well, according to a website used for making flight plans for airliners, it recommends LOTS of overwater waypoints. For example, for a flight from Kuala Lumpur (WMKK) to Jeddah (OEJN), the first overwater leg goes PUGER ANSAX IGEBO NOPEK DUBTA ELSAR GIRNA. Then it cuts across India, and the next over water leg consists of nine more waypoints. The total trip consists of like 25 waypoints.

    If you want one that is even more over water, here is the link to a trip from Perth to Jeddah that cuts across the area we are interested in. There are too many waypoints to even count, and it is not at all a great circle path. (Because of weather?) At any rate, here it is. Make of it what you will:

    https://flightplandatabase.com/plan/1488036

  342. Don Thompson says:

    @Warren

    flightplandatabase.com is for use with VATSIM & PC based sims.

    Real world, refer to NATOTS, PACOTS, AUSOTS.

  343. airlandseaman says:

    Warren: Let me be more clear. When filing a flight plan, and flying IFR over water, WPs are used for air traffic control purposes. I was addressing the case for MH370, which clearly had no intention of following an IFR flight plan.

  344. TBill says:

    @Peter
    There is not too much time for an earlier FMT. We have the BFO burst of SDU logon data at 18:25 – 18:28 suggesting MH370 was on an offset to N571 path and heading. So that gives us approx. 18:29 as the earliest possible start of FMT turn and 18:36 as the latest time to start the FMT turn, to be consistent with the 38-39 South theory.

    @Warren is suggesting a such a path that would turn almost immediately after 18:28 heading to 69S 69E (“6969S” manual waypoint) in Antarctica and that’s a great circle path that happens to pass almost directly thru POVUS ISBIX MUTMI RUNUT and about 39.2S on Arc7. So getting as far as 39S+ would probably be aided by the absolute immediate FMT say 18:29.

  345. TBill says:

    @Peter
    PS- 18:36 is more consistent with heading a little further out to IGOGU /ANOKO for the FMT.

  346. airlandseaman says:

    TBill, Peter, Warren: Re FMT time: The 18:25-18:40 BFO data shows that if there was a FMT within this time frame, and not a descent at 18:40, then the FMT was in process at the time of the 18:40 phone call burst of BFO values, not earlier. OTOH, if there was a descent underway, not a turn at 18:40, it might be possible that there was an earlier turn, but after 18:28, and then a descent at 18:40 (on a different heading).

  347. airlandseaman says:

    TBill, Peter, Warren: Here is an early assessment of the FMT. We were not seriously considering a descent at 18:40 at the time of the document, but it does explain why we thought the turn was underway at 18:40 (assuming no descent). http://bit.ly/2MfoHVj

  348. Wall says:

    @All

    Does someone know the precise coordinate of the Inmarsat satellite hanging above the Indian ocean? And I also need the exact radius of the circle d(satellite,7th-arc).

    Wall

  349. Warren Platts says:

    “Don says:
    flightplandatabase.com is for use with VATSIM & PC based sims.

    Real world, refer to NATOTS, PACOTS, AUSOTS.

    OK, I see what you mean. However, flightplandatabase has the virtue of being free, and it probably gives approximately the same answer. In any case, I asked it to give me a flight plan from Banda Aceh (WMMT) to the main Argentinian Antarctic base (Marabio Airport SAWB). This would be sort of like the simulated flight that kind of looked like it was aimed at McMurdo, but in the mirror image noted by Tbill. And sure enough, it recommends the flight corridor ISBIX MUTMI RUNUT (however, it crosses 7th Arc at 40S).

    airlandseaman says:
    December 10, 2018 at 2:00 pm
    Warren: Let me be more clear. When filing a flight plan, and flying IFR over water, WPs are used for air traffic control purposes. I was addressing the case for MH370, which clearly had no intention of following an IFR flight plan.”

    OK, again, you are attempting to read the mind of a dead man. Obviously, no flight plan was filed with anyone official, and the ADS-B was turned off. Nevertheless, there was an apparent flight plan that utilized multiple waypoints.

    The question you should ask yourself is: WHY? And the answer is: IT DOES NOT MATTER! The fact is, the aircraft followed a waypoint path as if a genuine flight plan had been officially filed. That is all we need to know in order to make reasonable predictions about the future flight path after radar contact was lost.

    Now, you are going to say that it doesn’t make sense to you. That is fine and dandy. However, anyone can equally speculate in the opposite direction. Perhaps the intention was to hide in plain sight. By following established airways with the ADS-B turned off, it would probably look like a B-52 on a military mission or something to anybody looking on primary radar. Similarly, for an aircraft heading into the SIO, if it was observed on the ISBIX MUTMI RUNUT corridor, it might not raise any eyebrows. Or maybe old habits die hard. Who knows? Again, this is pure speculation on my part. However, it is equally pure speculation to assume that the pilot cared nothing for waypoints tbqh.

  350. Warren Platts says:

    @airlandseaman says: The 18:25-18:40 BFO data shows that if there was a FMT within this time frame, and not a descent at 18:40, then the FMT was in process at the time of the 18:40 phone call burst of BFO values, not earlier. OTOH, if there was a descent underway, not a turn at 18:40, it might be possible that there was an earlier turn, but after 18:28, and then a descent at 18:40 (on a different heading).”

    At 18:25 the aircraft would have been at NILAM. WHat if the aircraft turned at 18:25 at NILAM and proceeded down airway P627? Then, it would hit POVUS at about 18:40 almost perfectly, at which point it would be natural to turn to the south somewhere (perhaps ISBIX MUTMI RUNUT?).

    That is, do we know for sure that there had to be only a single FMT, or could it be possible there were basically two turns, say, one at NILAM and one a POVUS? On the idea the pilot was following established airways for whatever reason, that would be the most natural path.

  351. Victor Iannello says:

    @Warren Platts: A turn to the south at NILAM around 18:25 would not satisfy the BFO value at 18:28. It is very unlikely that occurred.

  352. airlandseaman says:

    Warren: A turn before 18:28 is not consistent with the BFO data at that time. A turn “IN PROCESS” (85% finished toward 186) is consistent with the BFO data at 18:39-18:40. It is uncertain what happened in between. And of course, there could have been some other maneuvers in between, as long suspected. We don’t know for sure. The point is that the 18:40 BFO data is not only consistent with a turn circa 18:40, but the turn actually appears to be “IN PROCESS” then.

  353. Warren Platts says:

    @Victor & ALSM: OK, then how about a MEKAR SANOB path with a turn at SANOB? That would work out to about 18:28. Then the turn at POVUS would have to start a couple of minutes before 18:40. Thus we could have two FMTs and that would considerably relieve endurance models to the SW.

  354. Victor Iannello says:

    @Warrent Platts: Independent of exactly when and where the FMT occurred, a great circle path that crosses the 7th arc at 38.4S latitude and has acceptable BTO error would require a speed of M0.833 at FL350. The predicted flame out time is 23:50, or about 25 minutes earlier than observed. The fuel model would have to be in error by 5.8% to achieve the required endurance. As you’ve heard now many times, it is very unlikely that this occurred.

  355. Mick Gilbert says:

    @Warren Platts

    The fact is, the aircraft followed a waypoint path as if a genuine flight plan had been officially filed.

    Fact?! Well, there’s your problem right there!

    After the turn back at IGARI the only ‘waypoint path’ that the aircraft was known with reasonable certainty to have followed was between VAMPI and MEKAR (and it might have previously tracked through ENDOR). Contrary to your contention, all the evidence suggests that apart from VAMPI – MEKAR (less than 70 nm out of nearly 490 nm from the turn back or less than 15% of the observed flight) the airplane was not navigating by waypoints.

  356. DennisW says:

    @Mick

    +1

  357. Victor Iannello says:

    @Mick Gilbert: I’d put the Lido Hotel radar data in the category of questionable. Which means, the recorded radar targets might not have been along any airways after the turnback at IGARI.

  358. Don Thompson says:

    @Warren Platts

    @Mick hit one nail square on the head. The only factual information concerning waypoints is the flight plan describing the route for the 2014-03-07 16:42Z MH370 service to ZBAA.

    One topic you may have missed concerns the FMC on the B777-2H6 and its storage capacity. The MAS B777 fleet, excepting 9M-MRP and -MRQ, had AIMS-1 implying a limited storage capacity for FMC data. The FMC nav databases were not global, they contained a limited selection of navigation data. The configured data has remained privileged information, but it wouldn’t be safe to assume Antarctic data to be included.

  359. DennisW says:

    @VictorI

    The Lido data agrees quite well with the 18:25 ISAT data (which all of you IG people reject as anomolous). I regrad it as valid. Don’t be insulted when I say that you and the rest of your IG buddies are questionable.

  360. Victor Iannello says:

    @DennisW: To match the Lido Hotel radar data to the entire sequence of satellite data for the log-on at 18:25 requires a maneuver between 18:22 and 18:25. That’s the genesis of the lateral offset of N571. If the requirement for the maneuver qualifies as “agrees well”, then I agree with your characterization.

  361. DennisW says:

    @Victor

    Oh please.

    I am not stupid. When you say:

    @DennisW: To match the Lido Hotel radar data to the entire sequence of satellite data for the log-on at 18:25 requires a maneuver between 18:22 and 18:25.

    No maneuver is required until after 18:25 to account for the data from point onward.

    18:25 itself is very clean.

  362. Victor Iannello says:

    @DennisW: You are not understanding. We’ve been through this before.

    To make is simple, let’s ignore the BFO data for now. Just look at the BTO data. You cannot start at the last radar point at 18:22:12 and continue along at 500 knots and a track of 296T and match all the BTO values at 18:25 – 18:28. Yes, it will look as though there is a match for the BTO value at 18:25:27, but only because the uncertainty for that value is the highest, being an R600 channel (SD around 90 us). If you look at the subsequent values of the BTO on an R1200 channel, you will see errors greater than 2 standard deviations (SD around 29 us) for that path. Basically, the BTO trend of the path doesn’t match the BTO trend of the data between 18:25 and 18:28. The only way to match the radar at 18:22:12 with the subsequent BTO data is with a maneuver before the first BTO value, such as the lateral offset.

  363. DennisW says:

    @Victor

    I agree. Adjustments need to be made after 18:25:27. And yes, we have been through this before.

  364. Ventus45 says:

    @Warren Platts

    Oceanic Flights these days are not restricted to the old “defined great circle routes”.
    Modern computerised flight planning systems definately take weather into account, and go searching for the best winds (tailwinds) or least adverse (headwinds).

    Particularly in the case of westbound flights across the Indian Ocean, where headwinds are often ferocious, the tracks actually flight planned, and actually flown, are often a long way off “great circle” routes.
    To illustrate, Qantas 63 has taken off from SYDNEY (YSSY) and has just passed OTKED and is heading over Tasmania on it’s way to to JOHANNESBURG (FAOR).
    http://www.airservicesaustralia.com/wp-content/uploads/Sydney-Gates-South.pdf
    See https://flightaware.com/live/flight/QFA63#
    The flight plan route is:
    DCT WOL H20 OTKED J22 LT/M084F320 DCT 4447S14343E 5029S13647E 5215S13434E/M082F300 5852S12551E/M083F320 6202S11744E 6326S11120E/M084F340 6500S09727E 6405S08012E/N0488F350 6326S07500E 6255S07142E 6012S06150E 5827S05809E 5515S05323E 5129S04938E/N0476F370 4652S04533E 4142S04105E 3811S03801E 3458S03515E/N0484F380 3152S03241E DCT GYV UZ29 STV
    (https://flightaware.com/live/flight/QFA63/history/20181211/0035Z/YSSY/FAOR/route)

    The flight progress so far is:
    https://flightaware.com/live/flight/QFA63/history/20181211/0035Z/YSSY/FAOR/tracklog

    As you can see, (from that flight plan for flight sims site above),
    SYDNEY (YSSY) to JOHANNESBURG (FAOR)
    https://flightplandatabase.com/search?fromICAO=YSSY&toICAO=FAOR&sort=popularity
    there are many examples of tracks both north of, and south of, the great circle route.

    Just as a side note, as Don mentioned above, we do not know what waypoints were loaded in 9M-MRO.

    If we assume there were no Antarctic waypoints, that leaves two possibilities.
    Either (A) Indian Ocean waypoints (for say routes between FAOR and YPPH and/or YSSY; OR (B) manual input.

    Some Southern Indian Ocean waypoints.

    OKTOK
    S 33°12.20′
    E 113°23.90′

    VISIT
    S 36°44.00′
    E 105°00.00′

    EKUTA
    S 39°50.60′
    E 095°00.00′

    IGPOL
    S 41°53.80′
    E 085°00.00′

    KASDA
    S 35°10.90′
    E 105°00.00′

    OLPUS
    S 37°06.90′
    E 095°00.00′

    SEBRO
    S 38°08.70′
    E 085°00.00

    SUNKI
    S 38°18.70′
    E 075°00.00′

    UPNEK
    S 37°37.15′
    E 065°00.00′

    RAXES
    S 36°26.00′
    E 057°00.00′

    GEVIS
    S 29°41.62′
    E 033°57.63′

  365. Tim says:

    @victor,
    And how much would a climb or descent change these values? Let’s say 3000ft/min.

    Thanks

  366. mash says:

    @Peter Norton says:
    “There could have been 3 distinct phases of the flight, each with different priorities:
    phase 1 [KLIA-IGARI]: priority = making sure everything looks normal until deviation
    phase 2 [IGARI-18:25]: priority = fast getaway + reducing risk of interception + laying false trail
    phase 3 [18:25-EoF]: priority = maximizing range”

    Perhaps the major constraint of phase 3 is that it must be a pre-programmed/passive flight. That’s why all the suspected maneuvering/loitering required at phase 2 – mainly to adjust fuel to the right level (and to start at the right starting point [or even time, considering the “daybreak” ending]). The (likely) end point must be very precise, in the sense that it is very close to Australia (error of margin small). This also fits well with the observed data, contradictory/deceptive behaviours and the on-going patterns.

    Thus, the 3 phases could be:
    1. normal
    2. ‘controlled’.
    3. ‘uncontrolled’.

  367. Ventus45 says:

    @Warren

    I agree with airlandseaman, only one end point was required, or used.

    I forgot to add in the (previous post awaiting moderation), that personally, Z’s final waypoint was likely one of three possibilities (all at 85 east).
    Either SEBRO, or IGPOL, or half way between them, at a manually input 40 south, 85 east.

  368. Victor Iannello says:

    @DennisW: No, adjustments to the path needs to start BEFORE 18:25:27. Again, you are only considering the first BTO point with the most error.

    One thing that became apparent to those of us that studied the statistics of the MH371 BTO data is the uncertainty associated with the first value at log-on (the log-on request) is very high–I estimate it to be around 90. However, the uncertainty associated with the second value (the log-on acknowledge) is 30, which is about the same as the other R1200 values. Those values typically occur close in time. In our case, the time separation is 7 seconds, and the BTO along the path changes at around 1.5/s, so expected changed in BTO from one point to the other is only 10.5.

    The first BTO value at 18:25:27 has an uncorrected value of 17120, which gets corrected to 12542 after subtracting 4578. (The value of 4578 was determined by analyzing the MH371 data, and differs slightly from the value of 4600 proposed by Inmarsat.) The next value at 18:25:34 (just 7 seconds later) has an uncorrected value of 51700, which gets corrected to 12615 after subtracting 23 and then 5*7812.5, again using a correction determined by analyzing the “anomalous” BTO data from MH371.

    The remaining 6 values between 18:27:04 and 18:28:15 need no correction.

    A path starting at the radar point at 18:22:12 and proceeding at 500 knots on a track of 296T will match the BTO value at 18:25:27 with a BTO error of only 18, but the BTO error at 18:25:34 is 96, or >3σ. There are similar errors for the other BTO values.

    Now, the corrected values for the first two BTOs (which are only 7 seconds apart) are 12542 and 12615, a difference of difference of 73, which seems high, but is not high when considering that the SD is 90 for the first value and 30 for the second. Statistically, they are equivalent. However, the second value has much less uncertainty. The two values can be combined by weighting their values by the squared inverse of their uncertainties, which yields a value of 12595 with SD=28.6. For the 500 knot, 296T path with no lateral offset, the error for these combined points is 2.8σ. This makes that path very unlikely.

    In short, if we accept the validity of the radar point at 18:22:12, the BTO values only match with a maneuver before the BTO value at 18:25:34.

    Now, if you only consider the first BTO value at 18:25:27, no maneuver is necessary, but only because the uncertainty of that value is very high (90).

  369. Ventus45 says:

    @Warren

    I might add that QF64, is in flight, eastbound from FAOR to YSSY, and is approaching Victoria, and is about 90 minutes from YSSY.

    It’s flight plan route was:
    DCT APDAK UQ48 AVAVA UZ8 RBV DCT 3013S03340E 3231S03614E 3610S04108E 3816S04441E/N0485F350 4112S05126E 4153S05334E/N0483F330 4259S05700E 4345S05944E/N0473F310 4500S06338E 4636S06942E/M082F330 4748S07500E 4806S07633E 4834S08004E/M083F350 4848S08612E 4734S09854E/M080F330 4627S10550E 4548S11012E/M083F370 4432S11926E 4320S12706E 4221S13111E/M083F390 4039S13641E/M082F370 DCT SPODD/M083F390 DCT ML H129 DOSEL Y59 RIVET
    (https://flightaware.com/live/flight/QFA64)

    Makes an interestin comparison with QF63 that is westbound. They just passed a little while ago.

  370. airlandseaman says:

    Any attempt to imply that multiply way points were needed or used beyond ~19:00 based on the fact that NORMAL flight plans do, is simply absurd. Conflating the two is nonsense.

  371. DennisW says:

    @Victor

    The unceratinty at 18:25:27 may have been high, but it was not needed. The BFO/BTO values were spot on according to my calculations given the location, speed, and ground track at that time. Of course, Hollande, Mike, et. al. argue that the data was flawed and not trustworthy. Could be.

  372. Brian Anderson says:

    Actually there is very little evidence of any waypoints being followed after the aircraft departed from its planned route, after IGARI. There is some evidence to suggest that after IGARI the aircraft was being flown manually. Z, an experienced pilot would have known the air routes around the Malaysian peninsula like the back of his hand. He would also be very familiar with the general geography, and the “landmarks” from the air, and even at night the light patterns would be very familiar.

    It is entirely possible that the flight track up the Mallaca Strait just happened to look as though way points were being followed when in fact the track could easily have been set on the heading indicator.

    And, after the FMT there is no evidence that any waypoints were being followed. It is a fruitless exercise trying to guess a track into the SIO using waypoints. None are necessary to get there.

  373. DrB says:

    @Peter Norton,

    The results of the fuel model predictions I posted yesterday were based only on predicted endurance and range. They did not consider fitting the satellite data. My Publication #39 from last March does that and considers all the above simultaneously.

    Look at Table 2 and at Figure 11 in paper #39. They show the routes (for various combinations of navigation method, altitude, and speed mode) which fit the satellite data well (including the Inmarsat BFO criterion) and also match the known endurance. You will see on the right side of Table 2 the 7th Arc crossing latitudes. They range from 30.3S to 35.8S, using the assumption that there were no maneuvers executed after 19:41. So, with that assumption, my results do not allow the possibility of more southerly latitudes than 35.8S, which disallows all of Hardy’s predictions (unless multiple maneuvers are added after 19:41 which are tailored to better match the satellite data).

    I was developing and refining my fuel model in 2017, and some changes made during that 11-month process had a small effect on predicted latitude limits. This is particularly true for MRC, for which there are no Boeing fuel flow tables. The latest version of my model in Publication #28 was done in November 2018, and I believe it is slightly more accurate than earlier versions.

    As TBill said, a turn to the SSW circa 18:36 can be completed prior to the 18:40 BFOs. So, if one interprets those BFOs as a new course, which I do for reasons which have already been discussed at great length here, a turn is needed between 18:28 and 18:40, and this turn can be started as late as 18:38:22 near IGOGU. See Figure 2 in Publication #43, which shows a turn at 18:38 to 196 degrees true from IGOGU to BEDAX, followed by a FMT to 180 degrees true at 19:21 from BEDAX to BEBIM. Table 2 gives the maneuver details.

    FL350 is the last flight level reported by ACARS, and it was also confirmed by the radio calls to ATC just prior to diversion. Normally an even flight level would be used after the turn to the west at 17:21, but this may not have been a “normal” course change. We simply don’t know much about the altitude after 18:22, and I include it as a variable when route fitting. Adjusting the altitude has an effect on the average airspeed and on the fuel flow, so the altitude adjustment helps to fit the satellite data and run out of fuel at the proper time. For instance, I found good fits to the BTOs in Table 2 in Publication #39 with MRC at FL336 and with ECON/CI52 at FL400.

  374. MH says:

    Given we are talking about what happened at IGARI. Was all the radar data available from the time of take off to IGARI for all air traffic accounted for? Was it possible some other aircraft followed MH370 and intercepted near IGARI? It seems the turn back was a capture of other kind of aircraft ie: military due to their flying characteristics not that one expect of a passenger jet.

  375. Wall says:

    @all

    Does someone know the exact coordinates of the Inmarsat satellite hanging above the Indian ocean and the radius of the circle? I do know approximately the distance, but I noticed something went wrong.

  376. Don Thompson says:

    @Wall

    Refer to Duncan Steel’s post.

  377. Victor Iannello says:

    @DennisW: There were 8 BTO values between 18:25:27 and 18:28:15. The first had an uncertainty of 90 μs the other 7 have an uncertainty of around 30 μs. You are using the first one and ignoring the others, including the second which occurred just 7 seconds after the first. By doing that, you allow a much larger range of permissible paths. The subsequent BTO values are much more restrictive, both individually and collectively, and are much better able to discriminate an acceptable path.

    This thread started based on a comment I made about questioning the validity of the Lido radar. Perhaps people should ask themselves why the BTO value at 18:25:27 fits so well with N571 and the last radar point at 18:22:12 when the uncertainty for that BTO point is the greatest. Perhaps it’s just a coincidence.

  378. DennisW says:

    @Victor

    You are using the first one and ignoring the others, including the second which occurred just 7 seconds after the first.

    I am just doing what Inmarsat suggests. (see page 7 of “The Search of MH370”)

    The subsequent messages during the logon sequence were found to have
    unreliable delay and are believed to be an artefact of the terminal switching channel and frequency during logon and so are not used in this analysis. This means that the BTO data for 18:25:34 and 00:19:37 should be ignored, but that corrected BTO values of 12520 and 18400 μs may be derived from the Logon Request messages at 18:25:27 and 00:19:29 UTC respectively.

  379. Victor Iannello says:

    @DennisW: We’ve learned more in the 4 years since Sept 2014. We’ve “decoded” the switching channel delay, and have shown that with the correction, the value is very reliable. And how about the other BTO values at 18:27 and 18:28 that require no correction? Why do you ignore them?

    If I have time later, I’ll put together a graph showing with high confidence the BTO data is not consistent with the radar data at 18:22:12 and remaining on a track of 296T at 500 knots.

  380. TBill says:

    @MH
    Re: other aircraft IGARI
    We are told of no such aircraft.
    Malaysia is holding much radar data as a state secret, including all military radar data. We do have some significant civil radar for MH370 only. Therefore it would be up to Malaysia to disclose any primary radar for “hidden” aircraft near IGARI. Flight Radar24 gives us the commercial flights.

    The radar data we do have shows MH370 switching back at IGARI, and the recepted cell phone registration at Penang confirms that, as well as the Inmarsat satellite data confirms that the aircraft seen flying off course was in fact MH370.

  381. TBill says:

    @Ventus45
    That is interesting commercial flight plan with all of the detailed lat/long waypoints ovr the ocean.

    I tend to question if MH370 was using distant waypoints. That could show incriminating intent to fly off to the middle of nowhere, if the aircraft was ever found. I also tend to question if MH370 did actually fly to the middle of nowhere, for the same reason. I tend to view a flight path that looks like an aborted approach to COCOS or something like that, or something on L894 to Perth.

  382. Victor Iannello says:

    @DennisW: First, a summary of the statistical analysis of the BTO log-on data prior to and including 16:00 (including flight MH371) is summarized in this comment.

    This figure illustrates the BTO trends for the log-on at 18:25. Shown is a plot (dotted black line) of BTO versus time assuming the 18:22:12 radar point is correct and the path remains at 296T at 500 knots, i.e., no maneuver. Also shown (black circles) are the (corrected) BTO values for the entire sequence, which can be seen to be three clusters in time. For each cluster, I’ve determined a “best estimate BTO” value (blue squares) based on weighting individual BTO values by the inverse of the square of their associated uncertainty. In addition, I show a path (solid black line ) that starts with the last radar value, but turns to 307T and decelerates to 389 knots. The error bars for the BTO data represent 2σ.

    Some observations:

    1) The “no maneuver” path satisfies the error criteria for the first BTO value at 18:25:27, but no other BTO data point is satisfied. Although not shown, it also matches the BFO data for the first and last points.

    2) The three “best estimate” BTO values fall on a straight line, within uncertainty limits. The SD for the three best estimates at 18:25, 18:27, and 18:28 are 29 μs, 17 μs, and 17 μs, respectively. They represent a very accurate way to discriminate acceptable paths.

    3) The “turn, decel” path satisfies the error criteria for all individual BTO data points, as well as for the “best estimate” BTO points. Although not shown, it also matches the BFO data for the first and last points.

    4) The “turn, decel” path is one of many maneuvers that will satisfy the BTO and BFO data. The lateral offset maneuver, essentially considering of a right turn-straight-left turn sequence, is another. However, a single turn or a single change in speed would not suffice.

    These observations mean that either:

    1) There was a maneuver that commenced before the log-on request at 18:25:27 that cause a deviation from tracking along airway N571 with no offset, or

    2) The radar data point at 18:22:12 is not correct.

  383. Niels says:

    @VictorI

    “Although not shown, it also matches the BFO data for the first and last points”
    Victor, is that under the assumption that at 18:28 the OXCO had been stabilized; and that there would not have been a substantial FFB shift after AES restart?

    In the broader context I’m interested to find out if the 18:28 BFO’s can give us an indication about a possible shift in FFB after AES restart; without a substantial shift “near straight” paths reaching S20 – S25 would be very unlikely.

  384. DennisW says:

    @Victor

    Thx. Your explanation makes sense.

    I am guilty of old thinking I suppose. I never could get anything to work out between 18:25 and 19:40, and simply drew a white circle around that area with question mark labels. I still feel that way.

    https://photos.app.goo.gl/x3ypv8KLJAMTPeqq7

    I did review your offset paths some time ago, but never really internalized them choosing instead to start path analytics well after the FMT at 19:41. Perhaps laziness. Perhaps a belief that knowing what happened between 18:25 and 19:41 was not worth the effort (laziness with a reason).

  385. Warren Platts says:

    @Victor Ianello: >There was a maneuver that commenced before the log-on request at 18:25:27 that cause a deviation from tracking along airway N571 with no offset

    So you are saying there very well could have been a turn before 18:25:27? That would be consistent with a left turn to get on airway P627 at NILAM. My understanding is that when making such a turn, the FMC will attempt to cut the corner. But to do that, the turn would have to be initiated before 18:25. The aircraft would then proceed down P627 until it hit POVUS close to 18:40, in which case there could be another turn to the south at that point. This diagram shows what I mean: https://i.imgur.com/ZEz06Ji.png

    @Victor >”a great circle path that crosses the 7th arc at 38.4S latitude and has acceptable BTO error would require a speed of M0.833 at FL350. The predicted flame out time is 23:50, or about 25 minutes earlier than observed. The fuel model would have to be in error by 5.8% to achieve the required endurance.”

    Sir, thank you for running that model for me, but that was something of a straw man. These fuel endurance models depend crucially on at least 4 or 5 variables: (1) temperature; (2) wind; (3) altitude; (4) mach number. We only have best guesses for all of these. Yes, wind does not directly factor, but it does affect ground speed. We know to get to the 5th arc, the ground speed has to be about 486 knots thereabouts. That translates to mach 0.83 at FL350, but if it was at say FL300, the mach number could be lower. Also if there is a tailwind component, that would also allow a lower mach number. Then there is temperature. YOu know how it is: the weather channel predicts 37 degrees for a high tomorrow, in the event, it is only 33. It is the same with predicting the weather across space to remote locations where there are not direct observations. It would not surprise me if the temperature models were a few degrees off in the deep SIO.

    @Dr. B: and thanks for your input, but you are incorrect that a 38.5S crossing would require constant speed. With the revised arcs, there would have to be a significant slow down sometime after the 5th arc. Thus, I am not yet convinced there is not some combination of parameters that are within acceptable error bars, combined with a slowdown in the latter part of the trip, that would allow such a far south crossing at fuel exhaustion.

    @Ventus: which leads to your very interesting idea, and that is certainly progress: namely, that aircraft navigating the SIO do not necessarily fly a great circle from point A to point B. Typically follow idiosyncratic waypoint paths, rather than great circles in order to take advantage of or avoid certain types of weather. That is something I do not believe has ever been done. That is, has anyone ever gone to one of those flight planning sites cited by Don above that take into account the weather, and then using weather data for March 8, 2014, see what courses it recommends for a flight from say Banda Aceh to points in the deep south? That would be a very interesting exercise imho. After all, why would the pilot not want to take into account the weather when “filing” the flight plan. If the goal was to maximize range, then it would make sense to seek tailwinds and avoid headwinds.

    Also thanks for the waypoints. I did not have all of them in my system. IGPOL (-41.896667°, 85.000000°) is especially intriguing as it is just on the other side of the crossing I am talking about (albeit a POVUS ISBIX MUTMI RUNUT IGPOL would cross at about 39.3 that would definitely be pushing the limits of what is possible fuel-wise).

    Which leads me to:

    @airlandseaman: “Any attempt to imply that multiply way points were needed or used beyond ~19:00 based on the fact that NORMAL flight plans do, is simply absurd.”

    Sir, there is no epistemological warrant for your assumption. You are attempting to read the mind of a deranged person. Of course that is impossible, so you are projecting what you would do. But there is no reason to think that the hijacker (if the plane was indeed hijacked) thinks the same way as you.

    As I said above, what we do have is a partial record of the flight from the turnback at IGARI to the end of the Malacca Strait. And based on the radar path charted, assuming we can accept it at face value, there was indeed a flight plan involving multiple waypoints: IGARI ABTOK LOSLO ENDOR OPOVI VAMPI MEKAR NILAM. Therefore, something like a POVUS ISBIX MUTMI RUNUT IGPOL path should not be surprising, according to my speculation. At any rate, assuming you are correct, the pilot could have simply entered POVUS IGPOL, and the path still goes pretty much directly over ISBIX MUTMI RUNUT anyways, so yes, it would be redundant to enter in all those if IGPOL was the chosen direction.

    WHich brings me to:

    Brian Anderson: >”Actually there is very little evidence of any waypoints being followed after the aircraft departed from its planned route, after IGARI … evidence that the aircraft was being flown manually … the track could easily have been set on the heading indicator.”

    Yes, all those things are possible. But in my post above I said the pilot was evidently in “LNAV” mode in scare quotes in order to emphasize that I merely meant that the aircraft was apparently following a waypoint path. Whether the aircraft was being handflown, or the pilot was twirling heading knobs, or the FMC was actually in LNAV mode with a bunch of waypoints programmed in, that does not matter imo. The fact remains that it was navigating using established waypoints and airways.

    Yes of course waypoints are not necessary to get to random points in the SIO. However, it does not logically follow that trivially true fact that rational reasons for following a waypoint path into the SIO cannot be imagined. The behavior exhibited seems to be following a strategy of hiding in plain sight. Therefore, a flight path following POVUS ISBIX MUTMI RUNUT IGPOL would make perfect sense: if the aircraft were picked up on a primary radar somewhere, it would simply look like a military transport heading to one of those several Antarctic bases (they all have their own airports or at least skiways) that lie in the same general direction.

    YEs, all of the above is pure speculation. However, I am just trying to inject a fresh take into the discussion in order to maybe move it forward a little bit. Thanks! 🙂

    PS I found a Russian website that seems to have every waypoint on Planet Earth listed:

    https://www.avsim.su/f/raznoe-24/skript-bukinga-54006/zip?i=175

  386. airlandseaman says:

    Warren: Re: “Sir, there is no epistemological warrant for your assumption.”

    You are not listening. It is not an assumption. It is a fact that there is no reason to use multiple WPs. A fact, not an assumption. What actually happened is unknown, but you can’t keep pretending that multiple WPs are likely for any reason whatsoever.

  387. Warren Platts says:

    “It is a fact that there is no reason to use multiple WPs.

    Respectfully, that is most emphatically not a necessary fact. Indeed I have provided FOUR such reasons:

    1. The turnback from IGARI obviously involved multiple WPs;
    2. Flight planning software for overwater use typically recommend multiple WPs;
    3. Flying along an established airway marked by multiple WPs is a good way to hide in plain sight;
    4. There could be reasons to deviate from a great circle paths (e.g., weather) that would necessarily require multiple WPs.

    Your argument seems to be that the above goes out to window because MH370 was obviously not a normal flight. I agree with your premise, but it does not follow that the aircraft might not operate in a more or less normal configuration other than the fact that it was hijacked and in the process of disappearing. Indeed, practically any model of MH370’s flight involves assumptions of normality: e.g., the compass is typically assumed to be set on NORMAL, rather than TRUE? Or fuel economy models make assumptions about the “cost index,” even though obviously cost was not a concern.

    At any rate, even if you were correct there are ZERO reasons for using multiple waypoints, a single, distant, established waypoint could be used. There are plenty that lie on the other side or near the 7th arc that offer potential aiming points: e.g., RERAB SEBRO IGPOL EKUTA OLPUS VISIT KASDA any number of waypoints up the west coast of Australia, not to mention WPs and airports in Antarctica. That is lot of options, to be sure, but not really when you consider the nearly infinite set of unconstrained paths. Of those I just listed, IGPOL fits BTOs rather well at cruising speeds–assuming of course that the fuel endurance can be made to work. But maybe EKUTA or OLPUS might work–I haven’t looked at those in detail yet. I will report my results when I get them.

  388. airlandseaman says:

    Warren: You really are not listening…or understanding.

    Re:
    1. The turnback from IGARI obviously involved multiple WPs;
    Wrong. There were no WPs used between 17:22 and 18:02. NONE.

    2. Flight planning software for overwater use typically recommend multiple WPs;
    Irrelevant. Not applicable in the case of MH370 as discussed in detail, but you don’t seem to get it

    3. Flying along an established airway marked by multiple WPs is a good way to hide in plain sight; So many assumptions burried in that statement I won’t even attempt to clean it up

    4. There could be reasons to deviate from a great circle paths (e.g., weather) that would necessarily require multiple WPs. Yes, of course, but that is PURE speculation.

    All we know for sure is that there is no reason to use multiple WPs to get to anywhere on the 7th arc. NONE.

  389. Warren Platts says:

    @airlandseaman: Sorry: I don’t mean to be annoying, but….

    1. It sure looks like several waypoints were used if we can regard the published radar tracks as true, namely ABTOK LOSLO ENDOR OPOVI. https://i.imgur.com/lq3Shzd.png

    2. If Zaharie was the pilot, how do we know he did not use any flight planning software? Presumably he had a laptop with him. He could have downloaded a flight plan before the flight, and then uploaded the flight plan into the FMC after the hijack. Do we know this happened? Of course not. However, if A is not known, that doesn’t entail that ~A must be true.

    3. By straddling the airspace between Malaysia, Thailand, Indonesia, and India, the aircraft was visible to radar–obviously–but no alarm bells went off. That is what I mean by hiding in plain sight: it is an attempt to not look too weird.

    4. It is pure speculation, I admit, but then again, so is every other justification for any other path. Aren’t you curious as to what the flight planning software would have said for March 8, 2014? I am. Why hasn’t anybody looked at that?

  390. Victor Iannello says:

    @Warren Platts: The effect of temperature and wind are included in the model. Uncertainty in parameters could not account for a 5.8% discrepancy in endurance.

    I ran the calculation again at FL400. To reach 38.4S requires a higher Mach number because the ISA temperature at FL400 is lower than at FL350, and the average offset from the ISA temperature (+3.2K) is also lower. The endurance discrepancy is 3.2%, and fuel exhaustion is reached about 14 minutes early. That represents an improvement compared to FL350, but is still not acceptable.

  391. TBill says:

    @Warren
    “2. If Zaharie was the pilot, how do we know he did not use any flight planning software? Presumably he had a laptop with him. He could have downloaded a flight plan before the flight, and then uploaded the flight plan into the FMC after the hijack. Do we know this happened? Of course not. However, if A is not known, that doesn’t entail that ~A must be true.”

    I don’t think the lower end IFE system on MH370 allowed Internet connection. I tend to like the idea of the pilot using Google Earth to plot the flight path, but if so he’d probably have to hand enter the lat/long coordinates into Google Earth, unless there is some way connect the lap top to the aircraft flight data system.

  392. airlandseaman says:

    Warren:

    Re: “It sure looks like several waypoints were used if we can regard the published radar tracks as true, namely ABTOK LOSLO ENDOR OPOVI. https://i.imgur.com/lq3Shzd.png

    Sorry, but that cartoon is not a real radar path. This was the real path according to the KB Civil Radar (PSR). http://bit.ly/2RXlmhu Clearly, no WPs. Hand flown.

  393. Don Thompson says:

    @Warren P,

    MH370 & waypoints is so 2014. Really.

    Flight planning is conducted by the dispatchers in the OCC, the captain merely signs-off his agreement the flight plan. For a N-S oceanic track, AUSOTS information would have been useless.

    My earlier mention of the oceanic track management organisations related to your Perth-Jeddah route – it’s unrealistic. Oceanic tracks consider a number of issues: comms & surveillance – regular reporting (ADS-C or HF); weather – navigation; safety – all aircraft operating on defined tracks, minimal confliciton risk out of SSR range.

  394. Warren Platts says:

    @ASLM: Thank you for the .kmz files showing the exact radar plots: they make my point for me even better: apparently, the aircraft near Khota Baru was aiming for NOLIC rather than ABTOK, albeit NOLIC is in the “cone of silence” zone. https://i.imgur.com/V4inUcY.png

    However, the turn around Penang is unmistakeble: it nails the ENDOR and OPOVI waypoints better than my cartoon that explicitly used them (granted my cartoon was based on the earlier radar reports, so it is not surprising.) https://i.imgur.com/wRam1kS.png

    @ Don: According to the ATSB, the most likely mode of navigation was LNAV, in which case waypoints would likely have been used.

  395. MH says:

    @TBill, there is many problems with the cell phone argument of it registering at one tower when many other cell phones were not …

  396. airlandseaman says:

    Warren Platts: The KB and BU radar data PROVES the 370 track was not a straight line segment anywhere between 17:30 and 18:02. It was wandering and slowly turning to the right the whole way from 17:30 to 17:52 (http://bit.ly/2M45B4d) then a right turn, followed by a correction to the left before finally taking up the post 18:02 course (~295). No autopilot works that way.

    BTW Where are ENDOR and OPOVI? They do not show up on any Malaysian data base I have checked. The only ENDOR WP I could find appears to be in OZ.
    Latitude 35° 17′ 12.00″ S
    Longitude 149° 20′ 48.00″ E

  397. Victor Iannello says:

    @Warren Platts: Yes, MH370 flew over the waypoints ENDOR and OPOVI when south of Penang Island. However, the aircraft was likely not in LNAV mode, as the radar targets are not along a great circle between the waypoints. Either the plane was manually flown between those waypoints, or in HDG/TRK SEL mode with the pilot dialing the direction. Either way, the pilot was likely using the navigational display (ND) for path guidance, as waypoints appear on the ND when in MAP mode, with WPT selected, and with a scale of 40 NM or less.

    The lightly banked turn as the aircraft rounded to the south of Penang Island and passed OPOVI is also not consistent with LNAV. That path appears to be likely manually flown, or less likely in HDG/TRK SEL mode with multiple pilot inputs.

    These observations combined with the “wavy” path after near Kota Bharu make it very unlikely that the autopilot was engaged with LNAV mode selected as MH370 passed back over the Malay peninsula.

  398. Brian Anderson says:

    @Warren,

    “The fact remains that it was navigating using established waypoints and airways.”

    Emphatically no ! As we keep saying, there is no evidence of that, and no evidence that LNAV was being used.

    Trying to guess a waypoint track following the FMT is an exercise in futility.

  399. Victor Iannello says:

    @Brian Anderson said: Trying to guess a waypoint track following the FMT is an exercise in futility.

    I think there is a fair possibility that LNAV mode was selected at some point before 19:41, and the plane was following a great circle path towards a waypoint until fuel exhaustion. However, as you say, it is futile to guess that waypoint, especially since it is trivial to enter a custom waypoint in the FMC. It would be unwise to organize a focused search effort based on a waypoint hunch.

  400. Victor Iannello says:

    @Warren Platts said: So you are saying there very well could have been a turn before 18:25:27? That would be consistent with a left turn to get on airway P627 at NILAM.

    No. If we accept the radar data point at 18:22:12, if there was a turn before 18:25:27, it was to the right, not the left, in order to satisfy the BTO and BFO value during the log-on sequence.

  401. Brian Anderson says:

    @Victor,
    “I think there is a fair possibility that LNAV mode was selected at some point before 19:41,…”
    I agree, and partly because the data suggests a smoothly varying track, without sudden course changes. Perhaps as early as 18:xx even.

  402. Ventus45 says:

    As I see it, Z was “in his own back yard” and could quite easily navigate visually, virtually as a VFR/VMC flight, and did, no autopilot or any FMC modes required, or desired, all the way to say, the island of Palu Perak, perhaps even further, until heading to VAMPI. Discussion of modes prior to then seems moot to me, though he would definitely need the FMC post the FMT if he had an end point in mind.

  403. airlandseaman says:

    Warren: I located the WPs you referred to. And, as Victor acknowledged, the radar track passes close to these WPs. However, the whole track around Penang is inconsistent with LNAV AP control. “S” shaped turns like we see here in the violet radar plot after OPOVI do not happen on AP. http://bit.ly/2BbuUy9 (Red is the smoothed, digitized SK999 track for reference) Looks much more like tweaking the heading select or manual flight.

  404. Don Thompson says:

    @Warren P

    ATSB wrote “Although waypoints and air routes were examined and compared to possible tracks derived from analysis of the SATCOM data, there was insufficient evidence to positively determine whether MH370 intersected any waypoints associated with published air routes in the Southern Indian Ocean.” Definition of Underwater Search Areas, 18 Aug 2014.

    Your comment is at best misinformed.

  405. airlandseaman says:

    I agree with Victor, Brian, Don and Ventus45 comments above.

    The evidece is most consistant with flying by hand from the IGARI turn until after OPOVI. After the “S turn”, the observed path could have been flown by hand or heading/track select for the next 20 minutes on the 294 leg. There were some maneuvers between 18:22 and 18:40. After the loiter and/or FMT(s), before 19:41, a final heading or track or LNAV WP may have been used, but there is not enough evidece to be sure if any of these happened after 19:41. Given all this, and the fact that the bto/bfo data is consitent with a straight shot somewhere, a final heading/track, or singe LNAV WP seems most likely.

  406. Niels says:

    Triggered by the discussion about the 18:25 – 18:28 BTO/BFO data I checked @DrB’s paper “MH370 SDU OCXO Transient Analysis”.

    https://drive.google.com/file/d/1VZe9r4QvxfD2kgyjh6F6WOQ0F-k1afVC/view

    If I read this correctly, this analysis suggests that at 18:28 the OCXO can be expected to be close to stable again, and that we could work with a BFO value of 143 Hz. This value is perfectly consistent with 500 knots at 296 degrees TT. It is a rather strong indicator that there was hardly any shift in the bias frequency after SDU power-up.
    In combination with the smooth path signature of both the BTO and BFO data after 19:41 it makes me doubt if S20 – 25 as the impact region is realistic in the light of the Inmarsat data. To reach to these latitudes with near straight paths you typically need to accept mean BFO errors of 10 Hz or more. Without significant bias frequency shift after the SDU power-up, it is questionable if such high values for mean BFO error can be explained well.

  407. Victor Iannello says:

    @Niels: Yes, the need for significant bias drift to match the BFO data is one of the reasons we initially limited the search to 25S latitude.

    Yet, the DSTG has presented past flight data showing that significant bias drift is possible, and recommends that path reconstructions be performed assuming a SD of 25 Hz for BFO error.

    And the Car Nicobar – Cocos Airport route I previously presented, which includes a deceleration and left turn at Cocos Island, has acceptable (< 7 Hz) BFO error for all BFO data except at 00:11, which might be due to OCXO frequency shift due to cabin de-pressurization. As I've said many times before, every remaining possibility for the location of the point of impact has associated evidence that casts some doubt. The underwater search experts tell us there is little possibility that the debris field was missed in the area previously searched, and for that reason are reluctant to search those areas again. With this in mind, where would you suggest we conduct the next search? How do you suggest that we proceed?

  408. Richard Godfrey says:

    @Victor

    “As I’ve said many times before”

    This appears to be the phrase of the month …

    Why do we have to repeatedly, repeat what we have already repeated?

  409. DennisW says:

    @Niels

    To reach to these latitudes with near straight paths you typically need to accept mean BFO errors of 10 Hz or more. Without significant bias frequency shift after the SDU power-up, it is questionable if such high values for mean BFO error can be explained well.

    You might want to take a look at Figure 5.4 on page 30 of the DSTG book. I would characterize the data in that figure as being within the bounds of normal.

  410. TBill says:

    @Warren
    What they are saying is nobody can figure out an Autopilot route around Penang, just like the way the pilot flew it. The way I do it in flight sim is I set the bank angle at a very gentle 5 degree max and fly thru the radar path points manually entered. But the actual flight path is not a 5 deg bank more like 7 degree, so my simulation is not perfect, but it gives me what I want (which is a simulation of what the pilot(s) could see out the windows).

  411. Victor Iannello says:

    @TBill said: @Warren, What they are saying is nobody can figure out an Autopilot route around Penang…

    No, what I actually said is that the radar data shows that it is unlikely that the autopilot was engaged in LNAV mode, which is Warren’s incorrect claim.

  412. TBill says:

    @Victor
    OK

  413. Niels says:

    @DennisW
    From the DSTG book and my (short) communication with Dr. Ian Holland it is clear that fig. 5.4 represents one of the flights for which there were BFO modelling problems (the alleged “geographic dependency”).

    @VictorI
    You asked: “where would you suggest we conduct the next search? How do you suggest that we proceed?”
    And: “The underwater search experts tell us there is little possibility that the debris field was missed in the area previously searched, and for that reason are reluctant to search those areas again.”

    The ATSB has been transparent, see for example the fig. that you posted on 18th Nov:
    http://mh370.radiantphysics.com/wp-content/uploads/2018/11/Sonar-Coverage.png

    Indeed the coverage shown in the shared figure is high (except perhaps the part where there is 7.8% lower confidence coverage).

    I think it is important and it would be helpful if a similar map could be prepared and shared by OI.

    I remain worried about the coverage especially in the S31 – S33 range, which includes the Diamantina Escarpment / southern border of Broken Ridge.

  414. Warren Platts says:

    @airlandseaman: “BTW Where are ENDOR and OPOVI? They do not show up on any Malaysian data base I have checked.”

    The coordinates I have for ENDOR are -35.286667 149.346945; the radar track goes about 150 yards to the south of it on my system (using the .kmz files you provided). The track also goes about about 300 yards to the north of OPOVI. Too close to be a coincidence imho. Even when flying in LNAV mode, according to my understanding, the FMC will direct the aircraft to cut the corner a little bit as it makes its turn, and thus one should not expect a perfect overflight of a waypoint even when in LNAV mode.

  415. Victor Iannello says:

    @Niels: It’s true that we know less about the search conducted by OI than the ATSB-led search. However, the same individuals that were responsible for assessing the efficiency of the ATSB’s search have similarly assessed the efficiency of OI’s search, and the verdict is that there is little chance a debris field was missed, whether or not the details are ever shared with us.

  416. Victor Iannello says:

    @Warren Platts: The radar plot does not show evidence of following a great circle path between OPOVI and ENDOR. In fact, the deviation of the actual path from the waypoint-to-waypoint path is on the wrong side of the “corner”.

    You are clearly arguing about something you don’t understand. You’ve been politely corrected by several people already. Please stop.

  417. Warren Platts says:

    @Don Thompson: “ATSB … insufficient evidence to positively determine whether MH370 intersected any waypoints”

    “Positively” is the operative word. The only way to positively determine whether MH370 overflew any waypoints in the SIO is to find the flight data recorder and see what it says. Therefore, the ATSB statement above is trivially true and basically says nothing. Here is part of the rest of the quote:

    “Air routes and waypoints were then examined to see if there was any correlation with the possible southern tracks for MH370 obtained from the analysis of the SATCOM data. Relevant southern air routes that MH370 may have intersected/traversed were N509, N640, L894 and M641. Waypoints associated with these air routes were also considered as possible points on the MH370 flight path…

    [Figure 34: Southern Indian Ocean air routes and selected waypoints]

    “The waypoints at MUTMI and RUNUT were also considered as possible points that MH370 may have crossed. However ground tracks through these points did not correlate well with the most favoured paths generated through the analysis of the BFO and BTO data.”

    So the MUTMI RUNUT IGPOL path did not correlate well with the most favored paths generated through analysis of BFO & BTO data. So what? You cannot tell me that a path crossing at 20S correlates better with the most favored paths of BFO/BTO analysis. It is premature to rule out possible waypoint paths imho.

  418. DrB says:

    @Victor Iannello,

    You said: “However, the same individuals that were responsible for assessing the efficiency of the ATSB’s search have similarly assessed the efficiency of OI’s search, and the verdict is that there is little chance a debris field was missed”

    Has OI revealed what percentage of their searched area was not not scanned at all? I seem to recall for ATSB it was 2-3%. That qualifies as “little” but it is not zero, and I can’t understand why anyone would object to filling in the small unscanned areas before proceeding to a new area. Re-scanning marginal areas is a more difficult decision.

  419. Victor Iannello says:

    @DrB said: Has OI revealed what percentage of their searched area was not not scanned at all? I seem to recall for ATSB it was 2-3%.

    Actually, for the ATSB-led search, the fraction of the area that was “black” was only 0.5%. In addition, about 2.1% of the area had a lower probability of detection of 70%. I am not aware if similar numbers have been compiled for the OI search.

  420. Warren Platts says:

    @Victor Ianello: “The radar plot does not show evidence of following a great circle path between OPOVI and ENDOR….

    You are clearly arguing about something you don’t understand. You’ve been politely corrected by several people already. Please stop.”

    Hold on a sec. I don’t know where you guys got the idea that I am insisting that the FMC was in LNAV mode from the turn at IGARI to past Pulau Perak or wherever. I am saying it does not matter whether the FMC was in LNAV mode or not. What I wrote at least twice is the aircraft was flying in “‘LNAV'” mode using “scare quotes” followed by an explicit explanation simply meaning that the aircraft was apparently navigating via waypoints.

    Let me try again. The behavior of the aircraft demonstrates an intention on the part of the pilot to use waypoints as targets to aim for as the aircraft was being navigated. There are multiple means to achieve that goal: programming waypoints into the FMC is one way; or one can twirl the heading knob; or one can handfly the aircraft using the navigation display as you mentioned above.

    Either way, for our practical purposes, it does not matter whether the aircraft was in LNAV mode or being handflown. The aircraft was following waypoints either way. The radar track proves that.

    Note that I had predicted years ahead of time before the radar tracks became available that the path went through ENDOR and OPOVI. My hypothesis was nicely confirmed by the .kmz files, thank you ALSM, and ought to encourage confidence in the general waypoint theory.

    If the aircraft was intercepting waypoints in the initial phase of the flight, there is no reason to think that the latter phase of the flight would not also intercept waypoints. That is a HUGE bit of information because it VASTLY reduces the search space of possible flight paths. How can that not be a good thing?

    As for the particular claim that the aircraft was being handflow post-IGARI, I agree! I never said it wasn’t being handflown, and yes, if you look at it, a great circle path from NOLIC to ENDOR would have overflown LOSLO–instead it went a couple of miles to the south (and thus avoided a 2nd incursion into Thai airspace)–the path looks a little herky-jerky, and as you noticed, it cuts the wrong corner at ENDOR, even though it is very close.

    And speaking of handflying, here is something you guys might find interesting that I noticed today. If you look at some of the published radar from the very first leg (from FR24 I think), the initial flight out of KUL is perfect: after they got the clearance to go direct to IGARI, the alignment of positions is perfect. Clearly the aircraft is in LNAV mode. Length of radar track is approximately 85nm. https://i.imgur.com/livVTqO.png

    Some other, sporadic positions were published as the plane closed in on IGARI. These, however, are not perfectly aligned. The headings of the points seem to switch significantly after my point #7 (it counts down from 9 to 1):

    9 (025.1T)
    8 (025.0T)
    7 (021.5T)
    6 (026.5T)
    5 (026.5T)
    4 (025.3T)
    3 (025.8T)

    After #2, the apparent turn to BITOD begins. Here is the chart–the distance from 9 to 1 is about 43.5 nm. https://i.imgur.com/iGDKzaa.png

    Note that the original coarse was 025T, speed 469 knots. Thus the hijacking appeared to occur some time between points #7 and #6. Since the distance between the two points is less than 4 nm, we can pinpoint the time to a 30 second period. Splitting the difference, the distance to the 17:21:03 position is about 28 nm. At 469 knots, then the time of the hijacking was approximately 17:17:24 give or take a few seconds.

  421. airlandseaman says:

    Warren: You note: “And speaking of handflying, here is something you guys might find interesting that I noticed today. [We noticed it 4.5 years ago.] If you look at some of the published radar from the very first leg (from FR24 I think), the initial flight out of KUL is perfect: after they got the clearance to go direct to IGARI, the alignment of positions is perfect. Clearly the aircraft is in LNAV mode.”

    Yes, of course it was using LNAV to IGARI. We’ve known that from day one. SOP.

    Here is a better plot of ADS-B and the KB and BU PSR plots for a direct compare: http://bit.ly/2UyhqFo

    It was flying in LNAV following WPs to IGARI, but not after IGARI. If you look at the ADS-B path (white dots) flown by LNAV, it is straight as an arrow. That is very different from the wandering path between 17:30 and 18:02 (teal and violet KB and BU PSR data)…which is the point we have all been trying to get you to acknowledge. Do you finally agree?

  422. Warren Platts says:

    Yes, I agree. I never did disagree. It was never my point that the aircraft was in LNAV mode post-IGARI. Well, perhaps I should say I was agnostic, because in my mind it didn’t matter. However, the radar maps you provided clearly show the aircraft was not in LNAV mode. Nonetheless, it appears there was an attempt to steer for ENDOR and OPOVI (and possibly NOLIC or MATBA–apparently ABTOK was not used as a waypoint–in the Khota Baru vicinity).

    Now that I concede your point, can you see mine? That it is possible to navigate via waypoints and yet not be in explicit, formal LNAV mode?

  423. airlandseaman says:

    Warren: Yes, of course you can dead reckon to any WP by aiming for one on the Nav Screen (MFD). And you will pass very near it if you are paying close attention. That is what may have happened.

    But you seem to completely miss the import of learning (via the KB and BU PSR data) that the plane was flown by hand from 17:30 to 18:02, not being controlled by the AP. We were not sure before the leaked PSR data. You keep dismissing this as irrelevant. But it is not. This is a far bigger deal than you seem to understand. It means the plane was not under any FMC programmed route (to the SIO). It means someone was alive and flying the plane, not just along (dead or alive) for a ride. Combined with the high speed and altitude at KB, it means there is virtually no possibility left that this was an accident.

  424. DennisW says:

    @Niels

    @DennisW
    From the DSTG book and my (short) communication with Dr. Ian Holland it is clear that fig. 5.4 represents one of the flights for which there were BFO modelling problems (the alleged “geographic dependency”).

    I also had a conversation with Holland and challenged his geographic dependency assertion pointing out that oscillator behavior is neither stationary nor ergodic, and his use of normal stats in not appropriate. His response is highlighted below.

    We are aware that the oscillator behaviour is strictly speaking, not stationary and ergodic. Fig. 5.4 from the DST Group book indeed indicates this. The first paragraph below that figure states “The mean bias is different between flights and even within a single flight there is evidence of structured variation.” The next paragraph in the book explains that the structured bias variations happen over a timescale of minutes rather than hours, but for MH370 the values are only available approximately hourly, and that is why we did not use a coloured BFO noise model in our trajectory analysis.

  425. Victor Iannello says:

    @Warren Platts: Nobody doubts that MH370 might have been in LNAV mode at some time before 19:41 to the end, and was flying towards a waypoint. The difficulty is deciding which waypoint it was, especially since it may have been a custom waypoint. You may feel strongly about a particular waypoint, but it’s only a guess. The likelihood of any particular guess to be correct is small. That’s why the search has been conducted by progressively searching the seabed along the 7th arc.

  426. DennisW says:

    @Niels

    Basically using BFO for anything other than concluding the aircraft went South after the FMT and came down rapidly at end flight is a serious mistake. Are you not paying attention?

  427. airlandseaman says:

    Dennis: Would you agree with this summary:

    BFO Data
    •Horizontal vs. Vertical sensitivity to Velocity
    •Subject to many small error sources, mostly well calibrated
    •What Doppler extracted from BFO data does tells us:
    –MH370 is on the southern half of the 7th ARC
    –At 18:40 MH370 was either:
    •turning to the south (or very recently turned ~south), or
    •descending at ~2400 ft/min while continuing WNW
    –Broad constraints on the path when combined with other data
    –There was a high rate of descent at 00:19:37, 2 min after Fuel Exhaustion
    (controlled or not)
    •What the data does not tell us:
    –Precise location on the 7th ARC
    –Navigation Mode

  428. David says:

    Accident investigation quality. IATA offers to assist.

    Quote, “IATA also believes not all accidents and incidents undergo proper investigation because of a lack of funds, said Gilberto Lopez Meyer, IATA’s senior vice president for safety and flight operations.”

    https://www.ainonline.com/aviation-news/air-transport/2018-12-12/iata-sees-role-association-accident-investigations

  429. Marijan says:

    Victor, about your question:

    “The underwater search experts tell us there is little possibility that the debris field was missed in the area previously searched, and for that reason are reluctant to search those areas again. With this in mind, where would you suggest we conduct the next search? How do you suggest that we proceed?”

    Here is my suggestion:

    1. Search the areas that were missed during the previous search (shadow zones, low probability detection areas…)
    This is a crucial step in order to move on. Recent experience with the search for ARA San Juan showed that only the thorough search, without leaving any gaps can be successful. The probability that the debris is within those areas is not low. Actually, it is extremely high, at least if we follow the probability density function.

    2. Search all other remaining areas defined by the ATSB as Priority Search Area in the First Principles Review (25nm from the 7th arc)
    We are still not absolutely sure that the wreckage is not within the ATSB priority search area and we should not move on before that.

    3. Perform surface drift study for the whole Wide Search Area defined by ATSB
    This needs to be done to pinpoint the areas where debris would drift west and not east towards Australia

    4. Perform new analysis and define the search area based on all available results

  430. Victor Iannello says:

    @Marijan: The same individuals that advised that OI go back and investigate a previous contact in the San Juan search are advising that all contacts from the previous underwater search of MH370 are clear. The actual data gaps are very small (0.5% of the area).

  431. Warren Platts says:

    @airlandseaman: “This is a far bigger deal than you seem to understand. It means the plane was not under any FMC programmed route (to the SIO). It means someone was alive and flying the plane, not just along (dead or alive) for a ride. Combined with the high speed and altitude at KB, it means there is virtually no possibility left that this was an accident.

    That is certainly a grave allegation. It is hard to believe that any human being, let alone an experienced pilot would be capable of such an act. It is incomprehensible. Yet the evidence speaks for itself.

    The important question for our practical purposes, is: What are the implications vis a vis the search? I can see at least two: (1) a glide scenario is certainly well within the realm of possibility–and therefore, if there were good reasons for searching below 25S, those reasons have not gone away; (2) there is a possibility the search could be constrained if we can reasonably deduce a destination or path based on waypoints or past simulations or whatever. Really, this is a job for a team of the best FBI profilers we have.

    @Victor: “Nobody doubts that MH370 might have been in LNAV mode at some time before 19:41 to the end, and was flying towards a waypoint. The difficulty is deciding which waypoint it was, especially since it may have been a custom waypoint.”

    I totally agree. Yet the number of “established” waypoints is much smaller than the number of random points. It is worth looking at, I think. For MRC flying, what do you think should be the Mach number that we should use?

  432. DennisW says:

    @ALSM

    Yes. Agree totally with your summary points regarding BFO.

  433. airlandseaman says:

    Dennis: Thank you for your response. In case you are wondering, that summary was from a slide in a presentation to OI… this week 1 year ago. The summary seems to have aged well.

  434. ArthurC says:

    I don’t want to deviate the conversation, but I was wondering about something.

    Is there a visual aid somewhere, that anyone on this site might be aware of, that might plot points or bands of certain or probable locations of the aircraft, based on actual facts?

    I have seen bits and pieces of it and I wonder if a larger-scale of that exists somewhere.
    To clarify, we have a clear picture of where it was during the transponder “on” timeframe.
    We then have some radar data, a cell phone registration that could pinpoint a range, then all the BTO/BFO analysis that could give us a band of calculated probable positions for each ping.

    If I had the knowledge, I would tackle this myself, but I am merely an observer.

  435. Warren Platts says:

    Is there a visual aid somewhere?”

    Google Earth? Get that loaded up, and then you can download the arcs provided by @airlandseaman here: http://bit.ly/2Mg5E1O Then do a file open on the .kmz files. Radar plots also provided by @airlandseaman here:

    https://www.dropbox.com/s/zu6dk2mphud5dff/MH370_Radar_Files_2018-04-11_MLE.zip?dl=0

    I am currently working on program to compile a .kml file showing all known Indian Ocean waypoints. Will post it here when completed.

  436. Victor Iannello says:

    @Warren Platts said: Yet the number of “established” waypoints is much smaller than the number of random points.

    In the post on great circle paths, I identified several paths that go towards specific waypoints like airports. However, standard waypoints include those that are defined by integer latitudes and longitudes. For instance, the standard waypoint 45S04 is located at (-45,104). (There are also ways to enter these coordinates as a custom waypoint.)

    The bottom line is that even with the assumption of a great circle path after 19:41, we’d only be guessing which final waypoint was used.

  437. TBill says:

    @ArthurC
    The Geoscience Australia slide show is among the best overall graphics:
    http://www.ga.gov.au/news-events/news/latest-news/the-data-behind-the-search-for-mh370-phase-two-data-released

  438. Warren Platts says:

    @Victor Ianello: “standard waypoints include those that are defined by integer latitudes and longitudes.”

    Yes, that Russian database has all of those listed. Interesting how they work out the nomenclature:

    45S04 -45.000000 104.000000
    4504S -45.000000 4.000000
    45W04 -45.000000 -104.000000
    4504W -45.000000 -4.000000
    45N04 45.000000 -104.000000 (aka KD90W)
    4504N 45.000000 -4.000000
    45E04 45.000000 104.000000
    4504E 45.000000 4.000000

    Your point is well taken: there are not quite a million of them, but there are a hell of a lot of them. Still, the behavior of the aircraft did demonstrate a seeming predilection for the 5-letter types of waypoints. Of those in the SIO, they are few and far between. Might be worth checking out.

  439. Marijan says:

    Victor, despite expert’s opinion, I don’t think that any area of reasonable size (ATSB states 200mx200m and 100mx100m) should remain unexamined before moving on to the next location. There is no guarantee that the plane is not within those areas no matter how small their share is in the total search area.

    I don’t think that the search was ARA San Juan was particularly efficient. Science works, and it worked well for San Juan. Hydroacoustic analysis placed the location of the implosion just 20km away from the actual event. However, it took much longer to find due to the choice of search strategy.

    ATSB, other agencies and individuals did an extremely thorough work on defining the possible location and, in my opinion, it would be the unwise not searching it completely.

  440. airlandseaman says:

    If I may but in, Marijan, and note the following:

    1. The “SJ miss experience” is noteworthy, and it has been discussed by many, including the experts. But the consensus is that there is only a 3% chance (average) MH370 was missed inside the searched perimeter. There are records to back this up. This 3% cannot be dismissed, but it must be weighed against the probability that 370 lies just barely beyond the outer perimeter of what has already been searched.

    2. Expanding the search width to, say ±30 nm would arguably increase the probability of detection (vs. width) to a number closer to where we already are inside the perimeter (97%). Thus, expanding the width marginally may be a better place to start. But the cost, logistics and efficiency of these 2 “next step options” are very different from each other, and the third option below.

    3. The third option: Searching the “next block” NE on the 7th arc (say, S25 to S20; ±25 nm) is also in the same family of “next step options”. In fact, it can be argued that all 3 of these “next step options” have about the same chance of success, given how difficult it is to calculate and assess the odds analytically. But the third option has several advantages over the others. (a) much less cost, (b) much more efficient (faster), (c) logistically easier, (d) more favorable WX.

    Given the choices, I would search these in reverse order, starting with option 3.

  441. DennisW says:

    @ALSM

    Another way to look at the search choices is to assume the choice has a negative result. Then ask what negative result yields the most information. Searching wider below 25S and not finding anything yields little aditional info. Likewise completely closing any small holes in the search yields little info if the plane is not found. In both cases we are left with the 25S to 20S elephant in the room. A negative result of a search +/-22nm from 25S to 20S yields a huge amount of info if the result is negative. It virtually guarantees that we need to search wider and fill in any significant holes.

  442. Trip says:

    Two questions on waypoints
    1. Are waypoints part of the flight plan and can they tell us anything? Do we know the waypoints from the original flight plan? The course input is obviously not listing the SIO as an end point but when exactly did the plane deviate from the flight plan indicating an override or manual takeover.

    2. Are any of the aircraft course settings transmitted to ground before or after takeoff? Does disconnecting the transponder have any effect on automatic or manual controls? It seemed in some of the original discussion people were talking about seeing a change in the autopilot program from the ground.

  443. Richard Godfrey says:

    A month ago, I posted the following comment: “Our priority is to check the search area between 25S and 20S at +/- 25 NM.”

    http://mh370.radiantphysics.com/2018/11/04/crash-debris-from-lion-air-jt610-provides-clues-about-mh370/#comment-19935

    A month ago, @DennisW posted the following comment: “Richard’s work has convinced me that a terminus North of 20S is very unlikely. Based on this logic it is more than 80% probable that the wreckage is located between 25S and 20S latitude on the 7th arc, and within +/- 25nm of the arc.”

    http://mh370.radiantphysics.com/2018/11/04/crash-debris-from-lion-air-jt610-provides-clues-about-mh370/#comment-19907

    @Victor Iannello states in the current post dated 30th November “This suggests future subsea search efforts should proceed along the 7th arc, starting where the last search ended near 25S latitude, and continuing farther north. (In a previous blog post, I showed that an automated flight ending along the 7th arc at 22S latitude is possible.)”

    @airlandseaman states in a recent comment dated 13th December 2018 “The third option: Searching the “next block” NE on the 7th arc (say, S25 to S20; ±25 nm) is also in the same family of “next step options”. In fact, it can be argued that all 3 of these “next step options” have about the same chance of success, given how difficult it is to calculate and assess the odds analytically. But the third option has several advantages over the others. (a) much less cost, (b) much more efficient (faster), (c) logistically easier, (d) more favorable WX. Given the choices, I would search these in reverse order, starting with option 3.”

    It appears that there is agreement among a number of contributors, that the next step is to search 25S to 20S at +/- 25 NM.

  444. DrB says:

    @airlandseaman.

    You said: “But the third option has several advantages over the others. (a) much less cost, (b) much more efficient (faster), (c) logistically easier, (d) more favorable WX.”

    If the missed/uncertain areas in the previous searches amount to even 3% of the 180,000 sq. km covered to date (and I think it is actually less), that is only 5,400 sq. km to be revisited. From 20-25S at +/-25 NM, the new area is roughly 72,000 sq. km, which is about 13X larger in area. How can surveying that 13X larger area be lower in cost or faster or even logistically easier?

    They might not use 6 AUVs at once, depending on the size and separation of the missed areas, but you only need to keep 1 or 2 operating to beat (in time and in cost) 6 AUVs covering an area that is 13X larger. We know OI can operate Seabed Constructor in the search area for many days (at least 30, I think) with no external support, so there is no logistical advantage in 20-25S.

  445. DennisW says:

    @DrB

    My last recommendation for a search of 25S to 20S used well-known (and well-used) elements of game theory. The question is not if the plane will be found. The important question is the value of the info gathered from a negative result. Your recommendation is basically brain-dead. There is little to no info associated with a negative search of a wider area or a negative search of “pockets”.

    Good f’ing grief. I did not think this post was necessary. The 25S to 20S search absolutely must be done first.

  446. Barry Carlson says:

    @DrB wrote, “How can surveying that 13X larger area be lower in cost or faster or even logistically easier?”

    You forgot to factor in the time/distance required to traverse to each potentially missed area, along with the time launching and recovering any AUV’s utilized in a rescan. It may only be possible to use one AUV in many situations, and the overall cost measured against time and distance involved in getting to each location means that that option will not stack up against 6 AUV’s gathering new data.

    Option #3 is by far the most sensible.

  447. DrB says:

    @Barry Carlson,

    You are addressing the wrong question, which is not which option is the most efficient, or even “sensible”. Mike claimed that searching an area 13 times larger was lower in cost and could be done in less time (I.e., “faster”). I believe those claims are both false, and you have presented no evidence to the contrary.

  448. DrB says:

    @DennisW,

    Game theory has its place, but it cannot be applied here because there is no reliable information (or even a consensus of opinion) on the relative probabilities of any of the areas being proposed. Your guesses are probably much different than my guesses. Game theory did not find ARA San Juan, nor would it ever have done so.

  449. Marijan says:

    A year ago Victor asked contributors of this blog the same question: “Where is the next best place to search?”, and I refrained from making comments thinking that I don’t have enough supporting arguments. This year, after the successful search for ARA San Juan, I will not let things go so easily.

    It is very simple.

    Areas 34-35S and 38-39S are still the best places to start. The former agrees well with every piece of evidence we got so far. Later one as well, except the debris arrival time to the African coast. Still, this area is supported by Simon Hardy’s theory, which captures with simplicity and clear reasoning. I give them equal probability.

    Airlandseaman, combine options 1 and 2 from your list and do one area at a time. Start from the 7th arc, fill the data gaps first, and expand to 40nm as it was originally planned.

    20-25S does not agree well with debris arrival times. Eventually, the aim should be to find the wreckage as soon as possible, with a minimum total cost, not the minimum cost per square mile searched.

  450. DennisW says:

    @DrB

    The role of game theory in this instance is to pick a course of action (a move in the game) that reduces the entropy of the solution space to the greatest extent possible. Searchng 25S to 20S at +/-25nm is huge. It basically covers all the reasonable 7th arc possibilities with respect to latitude, and then supports the notion that +/-25nm is not wide enough if the aircraft is not found.

  451. airlandseaman says:

    Bobby: I’m surprised you don’t understand the trade space. It has little to do with area. To me, and others here, it is obvious. Searching 600+ dots scattered up and down the 7th arc, many requiring an ROV deployment in 4000-6000 m water, would take more time, money and risk than mowing the lawn w/ 8 AUVs in a warm, rectangular virgin area half the distance from land.

  452. Niels says:

    As we don’t have detailed information on the sonar scan coverage for different latitudes I think it is hard to come up with a recommendation.

    I can only say that before the OI search I would estimate the probability that the crash location is in the S28 – S36 latitude range (near the 7th arc) to be 80 -90 %
    And that the a priori probability to find the wreckage in the S20 –S25 latitude range would be only a few percent, as you would need a path with at least one kink to satisfy BFO constraints*. The BTO / BFO data clearly has a smooth signature. On top of this, even with a kink in the path, one has to justify a couple of BFO errors outside the 7 Hz range*, which all together makes such as path not very likely.

    *Regarding the BFO error discussion I refer to the careful work of @sk999, see:

    Amsterdam (MH16):
    https://docs.google.com/document/d/1WtTqG4n_NX1IlkQo_peyJ6Lt7WTIrlcnz7WyzGDkTz0/edit

    KL (MH371):
    https://docs.google.com/document/d/1ILWlxwNGu9dbCsP_H_gf03IT12mi-hu-arF6QLE96q8/edit

    From the paper “Amsterdam Flight: MH16”:
    “The one counter-example of well-behaved BFOs provided by the DSTG (The Mumbai-KL flight, Fig 5.4) is clearly an outlier. The biggest excursions happen around 22:20, when the flight makes a turn at MEKAR, departs airway N571, and heads straight to KL. Unfortunately we do not have the raw BFO data for this flight, so replication of the DSTG results is not possible.”

    I think that @sk999 nailed it well.

    The problem with the DSTG analysis is not only that they admit that there are issues with their analysis (the “geographic dependency”), the bigger problem is that so far they have not been transparent regarding the data and detailed calculations they have used. I therefore tend to ignore their results.

  453. flatpack says:

    @Victor Ianello

    You asked:- “Why do you believe the captain was interested in dumping fuel? The fuel endurance indicates that little or no fuel was dumped.”

    A number of things;

    Firstly, it was a specific issue that ZS commented on re the sim. So at the very least he was interested in it to the extent of trying it out and reviewing the result.

    Secondly, as the potential search area is shifted north along the arc, there may be a point where too much fuel has been ‘saved’ by the putative holding pattern and some needs to be ‘dumped’ in order to meet the expected fuel endurance. Fuel may alternately be used in a less efficient manner.

    Thirdly, contrail management, that is changing the appearance of the flight.

  454. David says:

    @Dr B. Your 39th paper supposed a left engine failure at MEFE for a final log-on at 31.57˚S on 7th arc.
    In the parlance of Victor’s 19th August, 2018 post about Boeing simulations a left engine failure prompting the final log-on was with the aircraft in a “normal” electrical configuration.
    Had your paper’s been of the “alternative” configuration that log-on would have been at right engine failure, 8 mins earlier. That is 3% of the flight time from your 19:41 FMT.

    To realise the same distance, aircraft speed for that leg would need to rise from 443 knots by 13, reduced just a little by there being no speed reduction from single engine operation over the last 8 minutes. Since 443 was MRC speed at the altitude of your assessment that would require a corresponding increase of fuel for the right engine or an altitude change.
    The other option, shortening the distance while maintaining MRC fuel consumption and arc times, would lead to a log-on farther up the 7th arc, affecting the BTO/BFO fit. Conceivably that would result in the CMT route not being optimal.

    Earlier I have raised this incompatibility of the “alternative” configuration with postulated routes based on the “normal”. In this instance I notice that in your Fig.13 a simple swinging of the 6th arc position onto the 7th arc, as an approximation of that 8 mins flight time reduction, would lead to a crash site about a degree (60 NM) farther north, so the difference need not be trivial.
    If however the “alternate” configuration is integral to the analysis this does not arise. The benefit in doing this is that there can be more confidence in outcomes.

    To expand on that, in just one in six “normal” configuration simulations are descent rates similar to that derived from the final BFOs; Case #5. However as Gysbreght noticed that was a particular case, the left engine failure apparently being timed to be simultaneous with an autopilot induced nose down at stick shaking. Consequently the pitch down was more severe than had it been in a glide at stick shaking. Also, whereas the descent timing should have been around 2 minutes after left engine failure it was about 16.

    About that and the 4 “alternatives” that Boeing simulated Victor noted that, “For the five simulations …. with high descent rates, the descent rate of 15,000 fpm and the downward acceleration of 0.67g occur at different times, and are not predicted to occur at the time of the log-on.” As to the “alternative” configurations also differing from the BFO-derived descent, at least they all do have high descent rates and their timing is closer, their maximum flight time of 7½ mins being less than the above 16, which was just to log-on. Thus they offer a better prospect of compatibility with the simulations actual results.

    In short, assuming that the electrical configuration would be “normal” after Victor’s post is likely to be more risky, based on those simulation outcomes.
    I think the simulations have serious limitations besides those in the actual descent (e.g. lack of relight) being more in the nature of experiments as with Case #5 than validations. They also included variations with start speed and time between engine failures. One example (#6, which Victor illustrated) had that time being less than 2 mins, as you pointed out at the time, without explanation. Hence they might not present a good datum to compare with the likes of your 443 knots and 8 mins between engine failures.

    Yet all the same I note that the SSWG and many others accept their suitability despite their limitations.

    You noted under Victor’s post, “That may mean that the alternate electrical configuration simply cannot explain the data we have.” Still, it is the more compatible with the outcome of the Boeing simulations – and there are no alternatives offering to those as examples.

    Acceptance of the “alternative” configuration as default, which is my proposal on balance, extends to other solutions too. For example it will shift the reachable southern 7th arc point farther north.

    As to how important this is, that will vary with the time between engine failures, which itself can vary with such as the left air conditioning pack being off. Zero time = no effect is one extreme. 16 mins (Case #1) is another.

  455. Don Thompson says:

    @Neils wroteAs we don’t have detailed information on the sonar scan coverage for different latitudes I think it is hard to come up with a recommendation.

    Geoscience Australia has provided public access to the entire Phase II data: side scan imagery, geo-databases, the lot. It’d be a significant task, and require much storage capacity, to recreate a working environment so as to review. The Phase I, bathymetry survey, is relatively accessible but it’s simply depth information at 30-50m lateral resolution (res dependent on depth).

  456. TBill says:

    @David
    I would ask: what are all of the possible alternate electric configs that the pilot might have used (eg; consistent with turning off SDU at IGARI). Not sure if there is only one possible alt config, that ATSB/Boeing studied, or if that was just one of several possible alt configs.

  457. DennisW says:

    @Niels

    Ask yourself the following question.

    Which of the following would you prefer to know. You can only pick one.

    1> The plane was not found searching 25S to 20S +/-25nm

    2> The plane was not found search an additional 10nm on each side of the arc from 38S to 25S.

    3> The plane was not found searching all the “holes” from 38S to 25S.

    The problem with the DSTG analysis (and your analysis) is that the AES reference oscillator has been incorrectly modeled. Likewise the Inmarsat statement of +/-7 Hz would never be used by anyone at NIST or in the business of manufacturing oscillators.

  458. Warren Platts says:

    4> The plane was not found searching an additional 50 nm to the south of the area already searched from 40S to 30S.

    If there is a consensus at all, it is that MH370 was not an accident. Therefore, glide scenarios are possible, and thus, the good reasons for searching to the SW never went away just because the wreckage was not found within 50 nm of 7th arc.

    Presumably, if there was a glide scenario, it would not involve doubling back on the path just traversed under power, therefore, there is no need to search further to the north of the previously searched area.

  459. DrB says:

    @airlandseaman,
    @Barry Carlson,

    Here are some back-of-the-envelope calculations on the times needed to search 20-25S and to revisit the missing/marginal areas farther south.

    1. Time to search 20-25S +/- 25 NM

    Let A = search area = 72,000 sq. km

    Let R = search rate for 1 AUV = 167 sq. km per day

    Let N = number of AUVs deployed = 6

    Let T1 = time required to cover the search area A = A / (N*R) = 72 days

    2. Time to revisit missed/marginal areas in entire previous search

    Let a = area of previous search = 180,000 sq. km, including ATSB/Fugro and OI

    Let n = number of missed/marginal areas (assume 100 as a guess)

    Let f = fraction of search area that includes missed/marginal areas that need to be revisited (assume 1% to be generous, although OI says only 0.5% missed, and ATSB says about 1% missed)

    Let m = total area of all missed/marginal spots = a*f = 1,800 sq. km

    Let d = average distance between missed/marginal spots (assumed to be uniformly distributed) = SQRT(a/n) = 42 km

    Let s = average speed of mother ship when transiting from one missed spot to another = 10 kts = 18.5 kph

    Let t = average transit time to move from one missed spot to the next missed spot = d/s = 2.29 hours

    Let c = time to prosecute one missed spot after arriving at location = 1 hour for deployment and descent + scan time to cover the average area of one spot (m/n = 18 sq. km, which only requires 1 AUV) + 1 hour to ascend and be recovered
    = 1 hr + 24* (m/n)/R hr + 1 hr = 1 + 2.6 + 1 = 4.6 hrs

    Let T2 = total time needed = n*c + (n-1)*t = 459 hrs + 227 hrs = 28.6 days

    So we see that for 100 spots the time for revisiting missed areas is a little bit less than half the time to cover 20-25S out to +/- 25 NM. This demonstrates doing 20-25S takes longer and costs more, contrary to your claim.

    With fewer missed spots, say n = 10, it is more efficient to deploy multiple AUVs at each spot since the spot area is larger. However, this is somewhat offset by the longer transit distance from spot to spot. Using 4 AUVs deployed 1 hour apart, the 180 sq. km spot area in this case is covered in 6.5 hours of operation from each AUV, for a total elapsed scanning time of 9.5 hours. So, for n = 10 spots, c = 1 + 9.5 + 1 = 11.5 hours. The total time needed T2 = 10*11.5 hrs + 9*7.24 hours = 115 + 65 hrs = 180 hours = 7.5 days only, which is much shorter than the northern search.

    For a very large number of missed spots, say n = 300, the total time is about 45 days; this is still shorter and cheaper than the northern search. For even larger numbers of missed spots, a different strategy can be employed by having the AUVs cover multiple nearby spots rather than always recovering the AUVs before moving the mother ship to the next spot. This improves the efficiency and reduces the time required.

  460. DennisW says:

    @Warren

    Your suggestion is about twice the area of 25S to 20S +/- 25m

  461. DrB says:

    @David,

    My CMT fit to the 19:41-00:11 satellite data and ending near 31.6S has a very tight constraint on the route (i.e., direction and speed). There are combinations of electrical configuration, altitude, and speed setting that provide a match to the needed speed profile and to reach fuel exhaustion in the vicinity of the 6/7th Arcs. The freedom to move the route noticeably up and down the 7th Arc is not available. It’s not simply a matter of keeping the total distance traveled constant.

  462. David says:

    @DrB. “The freedom to move the route noticeably up and down the 7th Arc is not available. It’s not simply a matter of keeping the total distance traveled constant.” Thanks, I understand that.

    Putting it another way, you assumed a ‘normal’ configuration. What broadly do you think would be the effect of assuming the ‘alternative’, at the outset, on your outcome?

  463. David says:

    @TBill. “Not sure if there is only one possible alt config, that ATSB/Boeing studied, or if that was just one of several possible alt configs.”

    In his August post Victor said, “In the second group (Cases 3,4,6,10), the electrical configuration was in an “alternate” configuration in which the left generator and left backup generator were isolated with switches in the overhead panel. In this case, when the right engine is running, power is supplied to the left bus from the right bus through the bus tie breaker.”

    Does that help?

    Fossicking around still, on a related topic you and I both have raised the possibility of manned powered flight post the 7th arc log-on, though discussion surrounding an active pilot has in the main been about gliding then.

    I believe there are plenty of places at the 7th arc where the SDU could have been depowered and rebooted by a pilot, with fuel left. Had the left IDG been powering the SDU he could have cycled that; in unison with either tie breaker or the right IDG had that been operative also.

    Had the right IDG been powering the SDU he could have cycled that or either tie breaker.

    Then he would have needed to dive a minute later when the SDU rebooted, leading to the final BFOs.

    Subsequently he would have halted an APU auto-start to obviate an 8th arc at MEFE, though had the bus tie breaker been isolated already (ie the left engine had been powering the SDU with that isolated) then that would be unnecessary.

    At MEFE he might have put the nose down.

    That is a ‘how'(leaving IFE non-connection aside) but my musing is well short on the ‘why’ unless somehow it was related to the 18:25 log-on reason.

  464. David says:

    @TBill. 3rd last para, the end of the first line, “…..though had the LEFT tie breaker…”

  465. Victor Iannello says:

    @DrB: If I understand your calculations, I think you are using an optimistic value of “R” for the coverage rate for missed areas. Most of those areas are missed because of challenging terrain. The R might be an order of magnitude lower than for flat terrain.

  466. Niu Yunu says:

    Is it correct that we don’t know where MH370 is because the FMT was not captured on radar ?

    Assuming straight A/P flight from FMT to MEFE@ARC7, the BTO data would tell us where MH370 crossed ARC7 if we knew the FMT location, correct ?

    If this is true and in case this was a deliberate diversion with the intent to keep MH370’s final location a secret, the perpetrator would have to be ABSOLUTELY sure the FMT was not captured by radar. IIUC the FMT must have occurred between 18:25-18:40. That’s only a few minutes after leaving radar coverage in any case! (and my personal bet is that the FMT occurred at or shortly after the SDU relogon, I don’t think this is just coincidence)

    So how could the perpetrator be so sure that the FMT was not captured by radar, given that the entire plan stands or falls on this !? He could not just have “guessed”.

  467. DennisW says:

    @DrB

    Assume:

    1> The aircraft terminated on (+/-25nm) the 7th arc between 20S and 38S.

    2> It is estimated that there is a 1% chance of missing the aircraft in an area searched.

    So the probability that your suggested search finds the aircraft wreckage is estmimated by:

    probability wreckage was missed * fraction of arc searched = 0.01 * (13/18) – 0.0072.

    The probability of finding the wreckage by searching 25S to 20S is estimated by:

    probability of not missing wreckage * probability wreckage is in 25S to 20S = 0.99 * (1 – 0.0072) ~ 0.98.

    It is obvious where to search. Of course, you can muddy the water like TBill and others, and object to the assumption 1> above. However, unless you fundamentally object to 1>, you are left with these conclusions.

  468. David says:

    @DennisW, Dr B. I do not think it is a question of which, somehow competitive, but more that if none offers a sufficient incentive, none will be searched. Because it is the best on offer does not mean it is good enough.

    So it is a matter of looking into them until one emerges as attractive, genuinely, enthusiasm aside, to potential searchers. Currently I think it advisable to work on all in the hope that at least one will emerge as providing the incentive. It may take more time and more information yet.

  469. David says:

    Further to that, developments in the last several months have included ACARS traffic log release, the extension of drift analysis, Igari analysis leading to turn back speed, height and manual control deduction, release of some Boeing simulation data, new fragmented debris findings…..

  470. Ventus45 says:

    @Warren,

    You said:
    Presumably, if there was a glide scenario, it would not involve doubling back on the path just traversed under power, therefore, there is no need to search further to the north of the previously searched area.

    I disagree, for three primary reasons.
    (1) Overflying an area and “clearing it” of any ships in twilight is vital for “the vanish”.
    (2) I don’t want to attempt a ditch with the rising sun in my face, or even at 90 degrees, thank you very much. Any pilot knows why, but essentially, (being in the left seat) you want the sun back over your right shoulder, so that the aircraft shadow is visible to you out at your ten-o’clock position, as you ditch, (vital for “alighting” gently, with “minimum ROD”, and lastly,
    (3) You also want to align on a wave crest, which would be roughly 330-150 degrees (heading 330 obviously).

Leave a Reply