Minor MH370 Mysteries, #1: The Case of the Wayward Etihad A330 — UPDATED

UPDATED 1/29/16: Here’s an image from Victor Iannello showing how EY440 diverted from its normal flight path about two minutes after takeoff on January 7, when it was still climbing and at an altitude of 5000 feet:

EY440 Departure

Just to clear up any potential confusion, it seems most likely that this incident does not have anything to do with MH370, but it’s very interesting in its own right. What is the dynamic at work here? Is it part of a trend? If so, does it potentially represent a system-wide vulnerability?

Here’s another image from Victor showing the plane’s continued path over Malay Peninsula. He writes: “I re-examined the FlightAware ADS-B data and noticed that there is a gap starting at BIBAN and ending at Kota Bharu. The FlightRadar24 coverage looks more comprehensive than the FlightAware data, especially in the South China Sea (SCS). I have re-plotted the flight path such that each underlying FlightAware data point is shown, and estimated the path in the SCS from the FlightRadar24 video. The path does indeed seem to follow airways across the SCS. (It would be helpful to have the underlying FR24 data.) The route seems to be ANHOA-L637-BIBAN-L637-BITOD-M765-IGARI-M765-Kota Bharu-B219-Penang-G468-GUNIP-HOLD-Langkawi-B579-Phuket.”

EY440 Flight Path w data

ORIGINAL POST:

The case of missing Malaysia Airlines Flight 370 is an incredible strange one, as we all know. But what only the true obsessives know is that orbiting around the giant mystery is an Oort Cloud of lesser enigmas. I’d like to briefly diverge from this blog’s main line of inquiry to cast a glance at some of these issues.

My first installment concerns Etihad Airways Flight 440, which took off on January 7 for Ho Chi Minh City bound for Abu Dhabi. Scheduled to depart at 20:10 UTC, it actually left 13 minutes early. Then, instead of flying along its normal route, to the northwest, it flew almost due south, crossed waypoint IGARI, then flew along the Thai/Malaysia border to the Malacca Straits, where it flew in circles for an hour before finally heading off in the direction of Abu Dhabi. By this point, however, the plane no longer had the fuel to reach Abu Dhabi, so it stopped to refuel in Bombay and reached its destination many hours late, leaving some passengers irate. (Special thanks to reader @Sajid UK for bringing this to our collective attention via the comment section.)

This is all very strange, but what makes it interesting to the MH370 crowd is the fact that a portion of its bizarre route was an exact match with that taken by the Malaysian 777 when it initially took a runner. Had EY440 been taking part in some kind of experiment to recreate MH370’s route, perhaps to get a better understanding of the Inmarsat data or the radar data?

We may never know. Katie Connell, who heads up Etihad’s media relations for North America, was very friendly when I called her and asked her what had happened. She said she’d check with her colleagues at the head office in Abu Dhabi. “It was simply a scheduling decision by ops that was later adjusted,” she wrote me in a text earlier today. I wrote back, asking if her contacts had been able to explain why the plane had flown south instead of northwest, and why it had flown a holding pattern over the Malacca Strait. She answered: “No; I did not get into that level of detail. I go with what my folks said.”

So there you have it. Make of it what you will.

UPDATE: I should have pointed out that this topic has been discussed for quite a while in the comments section of “Free the Flaperon!” and “A Couple of MH370 Things.” One of the ideas mooted there was that the flight crew inadvertently entered the wrong route into the Flight Management System, somehow overlooked the fact that they were heading in the wrong direction (scary!) and then circled for an hour until they could get the proper flight plane figured out, filed and cleared. This would be embarrassing enough to the airline that they would prefer to call it a “scheduling decision that was later adjusted.”

UPDATE #2, 27 Jan 2016: I’ve received a clarification from Etihad via Katie Connell, who writes: “The standard route flown by Flight EY440 from Ho Chi Minh City to Abu Dhabi on January 7, 2016 was automatically amended by the Flight Planning System which calculated and filed an alternative route as the most favorable, due to high winds. Shortly after takeoff, a new route was re-plotted which required Flight EY440 to fly through Thai airspace. While awaiting the overflight clearances the aircraft went into a holding pattern which resulted in the aircraft needing to refuel in Mumbai prior to continuing its journey to Abu Dhabi.”

So it sounds like the problem was not a human mis-entry, but a faulty flight-plan solution by a computer, which then had to be fixed while in transit. Software bug? Non-optimal algorithm? It will be worth keeping an eye out for more incidents like this one. Here’s one that took place in December involving a Malaysia Airlines flight from Auckland to Kuala Lumpur.

UPDATE #3: Victor Iannello has directed my attention to a Wired article suggesting that hackers have disrupted flight plans in the past and could do so again.

Here’s a chart showing the path the flight took as it circled over the Malacca Strait, created by reader Oleksandr:

EY440 path

 

571 thoughts on “Minor MH370 Mysteries, #1: The Case of the Wayward Etihad A330 — UPDATED”

  1. @Paul Smithson: I did not mean to imply that the early reports were correct. I was simply trying to document the origin of the image with the loopy turn.

    You think more radar data was available than has been made public. I think it may be the exact opposite. I suspect that the only radar data that is available is the civil radar data from Bota Bharu and Butterworth (the short-range radar at Butterworth is part of the civil radar system) and the long-range military radar in the Malacca Strait from the radar site at Western Hill, Penang.

    I suspect that the turn at IGARI was not captured. This would imply that the older military radar site at Bukit Puteri was not operational. If true, Malaysia would certainly not want to make this known publicly.

  2. Oleksandr,
    Thanks for that clarification regarding the Curtain boom. If you’d like to write up a report I’d be happy to post it for everyone’s edification.
    As to your comment to Dennis, I think it bears emphasizing once again that what a person thinks is less important than how one thinks it. The whole point of coming together collectively to help each other understand the data is to try to understand what might have happened to the plane and, just as importantly, what couldn’t have happened to the plane–that is to say, to recognize what the bounds of possibility are. Dogmatic insistence that one’s favored scenario is the only correct one is poisonous to this effort, and I am very happy that it really doesn’t take place anymore, at least here. While I do favor a particular scenario, I also believe that all the ones you list (and some others as well) are still very much on the table.

  3. Daud’s misspeak on an aircraft’s progress across the Str Malacca was resoundly contradicted by Datuk Tengku Sariffuddin Bin Tengku Ahmad, Najib’s press secretary, who reiterated, “As far as they know, except for the air turn-back, there is no new development…” 11th March, The Guardian, 12th Mar NYT, NPR.

    Consider that with the ICAO MH370 SAR review/lessons learned comments concerning political interference.

    My concern for a motive, is Malaysia’s motive, which was manifested as repeated undermining of the efforts to locate the aircraft. From the initial events of 8th March thru to failing to make good on a public commitment to deploy resources for the sIO bathymetry survey and deep ocean sonar search.

    :Don

  4. Don, Victor

    I defer to you two who have done the detailed homework on precisely who said what/when and the extent to which the traces in the FI tie together with regard to plausible airspeed, heading, altitude, timing. Clearly, there was a bunch of contradictory announcements early on and I am not saying that one is more correct than the other – merely that there was either a) ambiguity in what Malaysian records showed or b) different folks had access to different information or c) the message going out had to be revised and controlled. I tend to go with “all of the above”.

    it is also clear that the Thai radar did capture an abnormal target coming back SW, starting 17:28, an on an intermittent basis thereafter. If so, why is the FI silent on what Thai radar showed [1.1.3 (e)]?

    Given the [nominal] coverage of Indonesian military radar it would seem highly improbable that they did not capture an abnormal primary radar target as well.

    If this sort of military radar record is not available we must assume that not one but several military defence radars were non-operational. Likely or not? I don’t know…but I’d trust Don to make an educated guess on just how likely this is.

    But the FI as it stands strikes me as an “authorised version” (albeit couched in more circumspect language than the political announcements of the time) than as a systematic analysis of what the available radar record from all sources shows…

  5. Given what is said about the radar ranges, it maybe that MH370 made a course change outside the range of several radars but its unclear even from Malaysian military radar that it actually crossed back over Malaysia. Hope if by digging deeper more information may come to light.

  6. @Stevan & @ DennisW – Maybe I don’t understand probabilities well enough for this to make sense but I look at the Ethiopia Air 961 and your CI scenario differently.
    On EA961 the hijackers had a motive – asylum in Australia.
    The highjackers didn’t believe the captain when he said there wasn’t enough fuel on board to reach Australia.
    The plane ran out of fuel and crashed into the water. In this case we have one unusual event – the highjacking followed by plausible events that led to a crash.
    In your CI scenario, we have an unusual event – the captain hijacks to plane to make a political statement regarding the court decision by flying to CI. (It needs to be triggered by the court event as he flew to Bali the previous day and Bali is only an hour or so from CI) This is then followed by a second highly improbable event – the plane didn’t arrive. If Captain Shah did intend to fly to CI, he would have landed at CI. The pax and crew would not have feared for their lives when the captain announced the diversion and didn’t have a reason to revolt. Therefore, I look at these as two highly improbably events both occurring on the same flight.

    @All – As for the reason the wreckage hasn’t been found I would say that one or more of the AP assumptions are incorrect. If we accept the the FMT was initiated manually but at some time during or after the FMT MH370 became a ghost flight there are plausible scenarios that I do not remember being proposed such as a higher initial altitude (FL380 or FL400) or instead of LRC, perhaps max TAS, MRC or ECON using the Cost Index of 52 was chosen. (Note, I have found two different units for the Cost Index and I couldn’t find if the 52 is kg/min or lb/hr.) As pointed out in the Dec.3, 2015 ATSB update, “When considering the scenario where the FMC maintained speed in accordance with the flight- plan cost index of 52 and the aircraft maintained a constant altitude, the initial speed would have been approximately Mach 0.83 decreasing to below Mach 0.80 as the aircraft gross weight decreased due to fuel burn.” Last May, Gysbreght posted graphs that showed that when endurance is constrained there might be only one or two combinations of speed and altitude capable of that endurance. If one were to pick a constant altitude and input a speed that decreased by Mach 0.005 per hour I think we might to predict a reasonable accurate location for the 6th arc at 00:11 UTC. A minute or two after that, we have a single engine burn rate of around 5 mt/hr for 3-5 minutes followed by a banked turn and spiral into the ocean. Now, while I agree that everything points to that turn being a left turn, I wonder if Mach 0.8 at 00:11 would carry enough speed for a right turn and still hit the 00:19 Arc?

  7. @Lauren

    My “CI scenario” is often mis-interpreted as a scenario with the prior intention of the PIC to land on CI. I do not believe that CI was ever a firm intended destination. It was on a menu of possible destinations that could be selected depending on how “negotiations” were proceeding in KL. This menu included Banda Aceh, Cocos, CI, Bandung,…and other airports that could conceivably accommodate a 777.

    My CI scenario should be interpreted as a flight path along the Southern Coast of Sumatra that just happened to end near CI when the aircraft ran out of fuel East of the Island.

    One might ask how an experienced pilot (i.e. Shah) could make a horrendous error like that. I think there are two contributing factors. Perhaps more that I have not thought of.

    1> low and slow

    The PIC was in no hurry, and flying low (to avoid radar) necessitated flying slow. These flight conditions are not at all normal for a 777, and fuel usage models for these conditions are not as accurate as the models would be for normal cruise operation.

    2> power cycling

    Something happened to cause the ISAT login at around 18:25. My speculation is that the EE bay was accessed by either the FE who was a passenger or by whichever pilot was locked out of the flight deck (most likely Hamid was locked out) or both. While in the EE bay breakers were manipulated in an attempt to discourage the PIC from continuing the diversion. It might be that the accumulated fuel usage was disrupted or compromised in some way along with the AES being rebooted.

    Lots of speculation, obviously. What I do want to emphasize, and my reason for posting, is to correct the interpretation of my CI scenario as a pre-meditated attempt to land on CI. CI was merely one of many options.

  8. Jeff,

    Thanks. Perhaps I should write a more detailed report, but I still hope to get some clarity from Dr Duncan as he is the best to comment his own work.

    What I know is that the intersection of the curve of possible sound origins, derived basing on RCS and HA01, with the 7th arc is around 100E, 28S – I have emphasised this earlier, including my TN-CTS-* notes.

    A few points are:

    – ATSB has discarded “Curtin event” on the basis of time inconsistency, but they did not consider that it could be caused by the seabed impact.

    – Dr Duncan and his team discarded the event based on the recordings at Scott Reef station, but it is not clear to me why they considered Scott Reef event to be of the same origin as RCS and HA01. Signal magnitude suggests that it was caused by a different event, while no data are available later at this station.

    – One argument against seabed impact was a thick layer of soft sediments along the 7th arc in the SIO, which was proven to be a common misperception, as shown by recently found ship wreckage.

    – When comparing the surface and seabed impacts, one has to remember that a much larger fraction of the kinetic energy goes to sound in case of the latter. Also one has to account for entrained water, which results in ~10x mass of B777 compared to its “dry” weight. My estimation of the maximum settling velocity of B777 filled with water was ~8 m/s if I recall correctly.

  9. Dennis,

    The paper you cited has numerous flaws. I saw it earlier, but I do not think it is worth of commenting.

  10. Jeff,
    I find it very strange that you are publishing an article on Feb7th without putting it or at least referring to it on your Web site. What’s wrong with that ?

  11. @Oleksandr

    “The paper you cited has numerous flaws. I saw it earlier, but I do not think it is worth of commenting.”

    That was not very helpful, Oleksandr. 🙂

  12. @DennisW
    Thanks for your recent insight, your objectivity, and a kinder, gentler DW. It feels like you have re-grouped and no longer waste time in some meaningless arguments. It’s an effective hybrid; brain power with additives of an unabridged ego and open mind.

  13. I am with Oleksandr.

    The signal to noise ratio in that data is so low, the author may as well have looked at sat pics of white caps, or find some shapes in the clouds.

  14. @MuOne

    True, but you would not expect anything else. The disturbance caused by the aircraft hitting the water is small relative to the aggregate of disturbances in a big ocean. The author of the article says as much.

    Any two such measurements could be easily discarded. The existence of a third independent disturbance emanating from the same point starts to get interesting, and a fourth independent disturbance even more so. Couple that with the fact that the disturbance location almost exactly coincides with a calculation I made long before I ever saw the sea state data, and there is reason to take notice. IMHO.

    Could it all be a coincidence? Sure. I am not claiming a slam dunk here by any means. Just seemed extremely interesting (and surprising) to me.

    I actually had second thoughts about referencing the paper since I was very sure how it would be received, and I really don’t need it to support my map pin.

  15. @DennisW and Oleksandr,

    The paper is interesting, but what it is missing is a probability that the alignment of the four “spikes” is not by random chance.

    If dozens of other similar patterns could be found within the same data, it’s meaningless. But if the probability is extremely low, it’s significant.

    That’s what the paper doesn’t get into, unfortunately.

  16. Assessment of Favored Routes from “Bayesian Methods”

    Figures 10.5 and 10.6 of “Bayesian Methods” (BM) are a pictorial summary of the best-fitting routes found by the DSTG. Gysbreht has already digested some of the information from the heading figure (10.5) in a previous post. What else can we glean from these figures?

    These routes allegedly all start from a point just East of Pulau Perak (18:01:49).

    Figure 10.5 shows the “control angle” (i.e., ground heading) distribution, and it is the easiest to understand. It shows that most favored routes are headed along a track angle of about 290+ degrees up
    until 18:39. These routes head towards VAMPI, MEKAR, NILAM, and IGOGU. There are a small number of routes that head more directly north – roughly towards IDKUT.

    In either case, the FMT happens as late as possible – just before the first 18:39+ SATCOM phone call.

    Afterwards, the plane heads South along a track angle of ~185 degrees give or take. There is a slight decrease in track angle along the various routes. This decrease is the opposite of what one would see if the plane were following a Great Circle (LNAV) route. Overall the best-fitting route is one that approximates a true track. In addition, there are a small number of routes for which the heading decreases significantly as the plane heads South, reaching to as low as 170 degrees. These routes are close to magnetic track routes.

    Figure 10.6 shows the velocity as a function of time, and here the behavior is rather complex. The figure exhibits multiple horizontal bands, some extending over the length of the flight, others over only a portion. The most extensive, continuous band over the entire flight path is at about Mach 0.79. This value is the speed appropriate for a true track route. Bands at Mach 0.81 to 0.84 are close to speeds appropriate for LNAV Great Circle routes. Clearly some of these are favored as well. Finally, there is a smattering of routes that reach as low as Mach 0.73. This Mach number is at the bottom of the range included in the study and is at the upper end of the speed required for magnetic track routes. Had the study included routes with still lower Mach number it would likely have found many more in this region of parameter space.

    It should be noted that the models all allow the speed to be changed aribitrarily at any time. That is why there are bands that cover only a portion of the flight.

    In summary, even though the BM routes encompass complexity not included in anyone else’s analyses, the most favored routes are still close to the ones that multiple people have derived based on less complex considerations.

    Comments/Feedback always welcome.

  17. Dennis,

    I agree that any expected signal in the wave height data would be minuscule compared to other disturbances. That was my point. The wave height data is incapable of discerning MH370 hitting the water.

    I am pretty sure, one could find “signals” of the level as per that paper for 3, 4 or more sites in corroboration ofany pin anyone puts anywhere in the IO.

  18. @Lauren H,

    For examples of routes with higher altitude and with higher speeds, see my post on February 6 at 6:23 pm.

    I concur that simultaneously matching endurance and range to Arcs 6/7 can in principle reduce the number of possible solutions. That is exactly why I have done this.

    One factor we would like to know is the engine PDAs. ATSB has refused to make this publicly known, claiming this is proprietary Boeing data. Furthermore, ATSB told me they would not release fuel reports from previous flights. I am submitting a FOIA request, but I am not optimistic about learning anything useful in the near term. Still, we can assume typical PDA values and not be too far off.

    If the radar track data are accepted, then the aircraft speed during that time exceeded ECON with Cost Index = 52. Possibly it was increased upon Diversion at 17:21. Perhaps it was even lowered again later in the flight, although this does not seem likely to me.

    Based on my analysis of FMT BFOs/BTOs, I believe a climb occurred then. This extends the range consistently with the higher average speed with a higher Cost Index. The result is a fairly small range of 7th Arc solutions overlapping the southwest end of the ATSB’s twice-extended Search Area. This SW region has not yet been thoroughly searched. If no debris is found there, then the next most likely region is the NE extension of the 7th Arc.

  19. @MuOne

    Without trying to belabor the paper (hey, I am not in love with it either) it is worthwhile to look at what is actually being measured at these stations. The measured parameter is pressure. Sea level height is not measured with a float, for example. The disturbances in the graphs are pressure changes, which are a proxy for water depth. If one actually looks at the data from one of these sites, the cyclical depth changes due to tides are very apparent. On the scale the author presented one is looking at pressure changes not wave height (which is filtered out).

    One of the questions troubling me was the propagation speed used by the author. I queried him, but received no response. Obviously the 950km/hr speed is not the speed of a surface wave disturbance. It does correspond to a pressure event of the type triggered by a tsunami. The propagation speed of this type of event is in that ballpark depending on a number of factors – water depth being the biggest.

    Anyway, it is what it is. I certainly understand and share your skepticism.

    My belief is that complex problems finally fall to a “preponderance of evidence”, and the disturbance paper, my terminal calculations, the location of the flaperon finding, and a reasonable(1) motive all reinforce one another.

    (1) There are many people who would call my motive suggestion silly (and worse), but it is better than no motive or causality.

  20. i have been thinking that a trick was done just after the FMT, even though it turned south at the correct time to make it look like MH370 flew south but a few moments later it turned and headed north. the end of flight location might be in the region of 39′-40′ N 84’E

  21. Dennis,

    Here are some more details of my skepticism regarding the paper you cited:

    1. Christmas Island AU Station. Disturbance is of the same order as other fluctuations in the plot. It is only 1 or 2 cm higher than the peaks 2 minutes earlier and 2 minutes later. You can take any other “peak”, for instance any of 3 peaks past 00:37. The level of noise depends on what was actually measured, with what accuracy, and what filter was applied to remove contribution of wind waves.

    2. Cape Cuvier Wharf station. The event lasted for 3 minutes and exceeded 15 cm above the “background” values. Do you believe that a splash 2,000 km away could cause such a displacement of water? My estimation is that this would require ~300 MJ energy coming to the stretch of 2 km along the shore at the Cape Cuvier Wharf. Given the distance of 2,000 km, the original energy should be of order 900 GJ, which is two orders higher than it would be released by the crash at 300 m/s speed.

    3. The spike at CI station should be the largest due to its proximity. It should be ~70 times higher than at the Cape Cuvier. But it is of the same amplitude, if not the lowest.

    4. Phase propagation speed depends on the depth and cannot be assumed as constant. This is in contrast to the sound speed. For the depth of 4 km, phase speed would be 713 km/h, which is considerably lower than 950 km/h derived by the author of this paper.

    5. The author assumed the impact at 00:19. The 13 minutes interval for the arrival of the first peak is already unrealistic due to wave propagation speed, and should the impact be a few minutes later, this would become impossible. Basically the author made a very strong assumption.

    6. The author concluded that “fluctuations have aligned coincidentally.” What fluctuations? I don’t see anything aligned at all. You can select a group of any other peaks and derive other locations. Why not? Perhaps some of them will also fall on the 7th arc or its proximity.

    7. The author’s calculations are quite approximate. I am not sure how sensitive his results to the input are.

  22. sk999:

    The horizontal time axis on figures 10.5 and 10.6 is labelled incorrectly. The time labels are one measurement to the left of where they should be. The turn should be between 1828 and 1839. The control angle at 1839 is the first point at close to 180 degrees.

    I’ve also commented earlier on Figure 10.7 and noted the significant trend in the BFO residual error. Although I’m not sure of the sign of the error, the trend could indicate that the highest probability path should have had more curvature to the east to eliminate that trend.

  23. @Oleksandr

    Thanks for the comments. I had some of the same concerns, but the physics of this paper are very far from my areas of experience or study.

    As I commented to MuOne, I do not believe we are looking at water “displacement” relative to these measurements. While the y-axis is labeled “height” that is not what is actually being measured. The transducers at these stations are measuring pressure which is produced by changes water height (s-waves in the seismic vernacular), but also by compression waves (p-waves in the seismic vernacular).

    Your assumption that the measured amplitude decays at a one over distance squared is not correct for this type (p-wave) of disturbance. It is probably closer to 1 over square root of distance or one over distance (still scratching my head over which applies here). Tsunamis decay at a one over distance rate so I am inclined to use that metaphor. In any case, your observation that the disturbance magnitudes do not exhibit decay with distance is well taken, and I have the same concern. Likewise with propagation speed. The physics here is far more complex than the paper suggests with respect to both amplitude and speed.

    Again, thanks for the comments. I am not going to get in the awkward position of trying to defend someone else’s work. I have more than enough difficulty defending my own. I offer it “as is”.

  24. @Gysbreght

    The most interesting figure in the DTSG document (to me) is figure 9.2 along with the associated commentary cut and pasted below.

    begin cut-paste//

    There are a number of paths that end in significantly different locations to the truth. These occur because in this flight the aircraft travels in a direction that is almost horizontally radial from the satellite. While the aircraft moves towards the satellite its initial dynamics constrain the plausible paths but once it passes through the point of closest approach and begins to move away then it is possible to make turns that result in different near-radial paths. The support of these ambiguous paths is disjoint because of the finite number of samples: the true underlying pdf has support all the way around the arc. Without dynamic constraints the location of the peak of the pdf is simply a function of measurement noise.

    end cut-paste//

    What this result tells me is that without dynamic constraints the flight path probability density function (pdf) is a continuous “fan”. The “spokes” shown” in figure 9.2 are the result of the sampling interval as stated. The true pdf is continuous. Likewise with MH370. I think the probability distribution is a continuous fan extending from approximately 40S to 10S along the 7th arc. In other words, it is possible to declare any location along the 7th arc from Sumatra to 40S as feasible.

  25. DennisW:

    This section comes across for me as pseudo-scientific bla-bla. The errant paths involve some near right-angle turns, so what are the dynamic constraints?

    What this result tells me is that selection of ‘plausible’ paths are made by the filter on the basis of BTO residual error, with very little weight given to the BFO. If you go from one arc to the next for a given groundspeed, considering BTO only, there are always two solutions that fit the distance/time equally well.

  26. @Gysbreght

    I pointed out some time ago (long before the DSTG report) that path bifurcations were mathematically consistent with both BTO and BFO. The bifurcation being about a radial line to the sub-satellite point. This observation was met with quite a lot of disdain on this site, and was essentially ignored and discarded. I did not pursue it further, since it served no useful purpose at the time.

    The DSTG report simply revived this “elephant in the room”.

  27. DennisW:

    A contributing factor is probably that the BFO is insensitive to east-west velocity, and therefore does not distinguish between paths that are symmetrical about a north-south axis.

  28. @DennisW: Regarding the bifurcation, if the BTOs and BFOs were exact, then for level flight, the BTO and Doppler shift are symmetric about the radial path to the subsatellite position, but the BFO is symmetric about the north-south azimuth. Therefore, only one branch of the bifurcation should be allowed.

    I think it is the now acknowledged imprecision of the BFO that allows the bifurcation.

  29. @Victor

    Yes, accept for those times when the satellite motion is minimal. Without satellite motion the BFO would be symmetrical about a radial to the sub-satellite point. The BFO error does indeed further expand the path possibilities.

  30. @DennisW: If the subsatellite is at its maximum or minimum declination, and the speed in the fixed earth frame is near zero, there is still the residual Doppler generated by the “pointing error”, which should (I believe) still be symmetric about the N-S axis. (I’ll try to find time to check this.)

  31. DennisW:
    The ‘motion’ of the satellite is greatest when it passes the equatorial plane. At that time the BFO is insensitive to airplane motion in any direction. The motion of the satellite is minimal at the maximum or minimum declination. At those times the BFO has the greatest sensitivity to airplane north-south velocity. Since the frequency compensation (delta_fcomp) compensates for the east-west component of airplane velocity, the BFO is nearly insensitive to airplane east-west velocity.

  32. @Gysbreght

    I agree with the substance of your remarks above. “Insensitive” and “nearly insensitive” are qualitative. It has been some time since I looked at this issue, but I seem to recall a longitude sensitivity relative to what Victor calls the “pointing error”.

    Not withstanding any of that, when the DSTG book is published by Springer you and Victor are both on my list of people for whom to buy a copy for your coffee tables. 🙂

  33. @Victor

    I’ll take another look as well. I am getting some CE credits over the next several days (through Saturday). So I will be out of action until Sunday.

  34. @Lauren H

    “This is then followed by a second highly improbable event – the plane didn’t arrive. If Captain Shah did intend to fly to CI, he would have landed at CI. The pax and crew would not have feared for their lives when the captain announced the diversion and didn’t have a reason to revolt.”

    you would have to personally know all the people on board to conclude something like that, out of 200 passengers there could always be several that could undertake something, they could yell if there is anyone who could break the door or know how to open it, that flight engineer might have listened to them and started fiddling with CBs in E/E bay…just one of infinite possible scenarios

    You cannot be sure that all of the >200 passengers would just calmly sit going back over malaysian mainland and to the SIO when their destination was Beijing…well be sure that this is written by someone who certainly wouldn’t sit calmly in such situation.

    @DennisW

    “My “CI scenario” is often mis-interpreted as a scenario with the prior intention of the PIC to land on CI. I do not believe that CI was ever a firm intended destination. It was on a menu of possible destinations that could be selected depending on how “negotiations” were proceeding in KL. This menu included Banda Aceh, Cocos, CI, Bandung,…and other airports that could conceivably accommodate a 777.”

    this is where I slightly disagree with Dennis since I firmly believe destination was australian territory, being it Cocos/CI or even mainland (if not for the loiter around Banda Aceh it could be reachable)

    although yeah that decision might have been brought during the flight, but after the loiter I am quite sure only australian territory was the option

    @Oleksandr

    I have sent that link to Dennis just as an interesting coincidence, without claiming it had no errors. But the coincidence is quite stunning if you consider how big SIO is (or even just part around 7th arc).

  35. At the very least the following entities know, and have known, the failure scenario of the Flaperon.
    #1 Boeing
    #2 French Investigators
    #3 NTSB
    #4 US high ranking government and military officials

    One could debate the confidence level of the scenario (feasibility, concept, basic, detailed) but given the time available for analysis, and the talent & tools available, it must be nearing highly detailed.

    Why the absolute silence from the French authorities regarding the Flaperon failure scenario?
    – Perhaps they are waiting to come to a consensus?
    – Hung up in some legal circus inside French Justice Department?
    – French authorities are running into difficulties with consultation with Malaysia?
    – French authorities are putting into a report that is taking a ridiculous amount of time?

    The public may someday be fed some reason(s) for the delay but the silence for this period of time is seemingly inexcusable and unprecedented given the relative low complexity and rich evidence. What is going on?

  36. @Victor@Gysbreght

    Had a chance to take a quick look. Both of you are correct, and I am incorrect. The bifurcation is about a N-S line, not a radial to the subset point. Don’t spend any time on it.

  37. DennisW,

    Thanks for your reply.

    Bayesian Methods Fig. 5.6 offers a nice illustration of the principle. The curves reflect the cosine of the track angle, and are symmetrical about the track angles 0° and 180°. They have the same value for all speeds at track angles 90° and 270°.

  38. Dennis,

    You are right with regard to amplitude: my mind stuck on acoustic waves. It would be linear decrease if wave propagates without dispersion and dissipation in all directions in unlimited space. But this would still imply an order difference in water levels spikes at the two stations I mentioned.

    With regard to phase speed, I considered long waves, which is likely incorrect for this case: expected wave length would be much shorter than 4 km. That means lower propagation speed. Hence, the time interval of 13 minutes is definitely insufficient to correlate the spike at the first station with the splash. Unless the splash occurred closer, but then forget about all other stations.

    With regard to how level was measured, I don’t think it actually matters. Probably pressure sensors, which typically have resolution of 1-2 cm of level equivalent.

  39. StevanG,

    “But the coincidence is quite stunning…”

    What coincidence? You arbitrary take a number of spikes, which may be just noise, and call it “stunning coincidence”. Take any other set of peaks at your choice, and you will get another “stunning coincidence”.

  40. It would seem there is an equal chance MH370 flew North or South. What are the end of flight locations now? Probably on the North 7th arc ?

  41. @Oleksandr

    I think the “stunning” the StevenG was referring to is the probability of two “blind squirrels” coming to the same 10km x 10km spot on the 7th arc.

    If you take my pick as a prior the probability of someone picking the same spot at random is about 1 in 2000 (given the length of the arc between CI and 40S). If both picks were done randomly the probability of a match would be about 1 in 4 million.

    Of course, that says nothing about validity. It does say something about being “stunningly” unlikely. Just a really weird coincidence.

  42. Dennis,

    That is absolutely wrong assessment. Here I agree with JS.

    Four spikes were arbitrarily selected to triangulate location and wave propagation speeds assuming the impact at 00:19. Select other sets of peaks and other speeds, and new locations may also be in the proximity of the 7th arc. Have you checked this?

    You have the following possibilities:

    1. Assume these spikes are due to a single event and the impact was >=00:19. Wave propagation speed makes it impossible to reach the first station within 13 minutes. That means the origin must be even closer to it. Then the selected spikes at the other 3 stations are not relevant to this event. No “stunning” coincidence.

    2. Assume these spikes are caused by a single event, but earlier that 00:19. Apparently the speed would be lower, but triangulation point would be different, not necessarily on the 7th arc. But it is irrelevant to MH370 as it likely was in the air by 00:19. No coincidence again.

    3. The 4 spikes in question are not caused by a single event. Again, no coincidence.

    Your fundamental mistakes are:

    – You assume that 4 spikes are of the same origin, rather than just noise or irrelevant events.
    – You ignore physics. This allows you to hide higher propagation speed than physically possible in your probability assessment.

    The other mistakes of lesser importance:
    – You ignore accuracy of calculations.
    – You assume the impact at 00:19.

    So, what is the coincidence you and StevanG are talking about?

  43. @Oeksandr,

    “Spikes”?! That’s one mother of a euphemisms!

    First graph has 4 more disturbances of same magnitude. The rest, bar the last, have barely distinguishable anomalies.

    I tend towards assuming the author had followed discussions and went to find corroboration of a near CI location.

    As I said before, one could pick any pin anywhere and find wave data corroboration of similar quality.

    I think, there’s nothing to see here, lets move on.

  44. MuOne,

    That is why initially I wrote this paper is not worth of commenting. Generally, I also feel this “discovery” was made in support of CI.

    In my opinion it is arguable whether these “spikes” are above the level of noise or not. But Dennis and StevanG continue calling it “stunning” or “weird” coincidence, while I do not see how this could be called coincidence at all.

    Perhaps those disturbances of water level (or some of them, or some other) can be correlated with Curtin acoustic data. I don’t exclude a possibility of some geological event irrelevant to MH370. If so, perhaps some conclusion can be drawn with regard to the relevance or irrelevance of the signal at Scott Reef to MH370.

  45. @Oleksandr

    I think you are missing my point. How the author arrived at his location is not at all important. The important thing is that is is the same location. I have no quarrel with your comments at all relative to the quality and the selection of the “signals”.

    Now, if the author saw my CI result, and deliberately homed in on it, your point would be well taken. However, if two independent people just happened to get to the same point at random then it is a rare coincidence considering all the available points to chose from.

    Once again, I am not trying to assign significance to the tsunami data or accuracy or anything of the sort. I am merely defending Stevan’s remark that it is indeed a remarkable coincidence.

    I think you are interpreting a coincidence as a flattering attribute. In fact, it is very unflattering. The definition of coincidence below.

    coincidence -a remarkable concurrence of events or circumstances without apparent causal connection.

    When I label the tsunami data as a coincidence it is about the worst thing I can say about it. Likewise with MuOne’s comments. You are both struggling with the English language here, not the validity of my statements. I agree with MuOne. Let’s move on here.

    This is precisely the reaction I expected from the tsunami data which is why I did not reference it previously, and I truly wish I did not bring it up at all.

    The reason I have made a long comment here instead of simply moving on with my life – miserable as it frequently is, is that clarity of expression is important. If we cannot agree on what words mean, then we might as well give up blogging at all. Said another way, I have another “coincidence” lurking in the background, and I am debating whether to even discuss it.

  46. in this strange MH370 event any “coincidences” are likely worth bringing forward for discussion. It may lead into linking other things together.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.