What We Know Now About MH370

It’s been more than six months since MH370 vanished, and in some ways we know no more now than we did in late March: no new clues have emerged, no more data has been discovered.  In a sense, though, we have come a very long way. For one thing, we now understand how many of the “breaking news” developments that occurred in the early days were actually untrue. (There were no wild altitude swings, no “fighter plane-like” maneuvering, and probably no cell-tower connection with the first officer’s phone.) What’s more, thanks no doubt to a drumbeat of public pressure, the authorities have released a tremendous amount of data and provided useful explanations of how that data is being interpreted. And finally, a spontaneous collaboration between technical experts and enthusiasts around the world has provided a trove of insight into avionics, aerodynamics, satellite communications, and a whole host of other topics that collectively shed light on what might and what might not have taken place on the night of March 7/8, 2014.

While a great deal of information has become available, it has not always been easy to find; much of it, for instance, has been exchanged via email chains and Dropbox accounts. For my part, I often find myself rummaging through emails and folders looking for information that I’m pretty sure I’ve seen, but can’t remember where. So what I’d like to do with this post is try to aggregate some of the most basic facts — a set of canonical values, if you will, of the basic data on MH370. Necessarily, some of this data comes with implicit assumptions attached, so as far as possible I’ll try to make these assumptions explicit.

Okay, on to the data. What we know now:

The bedrock data. In the wake of MH370’s data, there were numerous news reports concerning information leaked by anonymous sources from within the investigation and elsewhere that have subsequently been either disproven or inadequately verified. For the purposes of the present discussion, the following are considered the bedrock sources of information upon which our understanding of the incident can be built — the “Holy Trinity” of MH370 data:

  1. Up to 17:21: radio communications, ACARS, transponder, ADS-B
  2. 17:22-18:22: military radar track. This information is of uncertain provenance but has been endorsed by the governments of both Malaysia and Australia. Furthermore, it plausibly connects the prior and following data sets.
  3. 18:25-0:19: Inmarsat data, especially BFO and BTO values. There is some discussion as to how this data is best interpreted, but the numbers themselves are assumed to have been received and recorded by Inmarsat from MH370 via their 3F-1 satellite. The “ping rings” in particular are derived through relatively simple mathematics and should be regarded as established fact unless someone comes up with a specific mechanism by which some other result could be obtained.

Timeline. Courtesy of Richard Godfrey and Don Thompson, here is a basic timeline of MH370’s disappearance (all times UTC):

  • 16:41:43 MH370 departs runway at KUL runway 32R
  • 17:01:14 MH370 flight crew report top of climb at 35,000 feeet
  • 17:07:48.907 Last acknowledged DATA-2 ACARS message sent from plane
  • 17:19:29 Last radio voice transmission
  • 17:21:04 Plane passes over IGARI waypoint
  • 17:21:13 MH370 disappears from air traffic control (secondary) radar screens
  • 18:22 Last primary radar fix
  • 18:25:27 Inmarsat log-on request initiated by aircraft
  • 0:19 Final transmission from aircraft to satellite

A more complete table of values, including the location of the plane at each point in time, can be found here, courtesy of the inimitable Paul Sladen. And Don Thompson has created an impressively detailed breakdown of the sequence of events, with a special focus on radio communications between the aircraft, ground, and satellite, here.

More stuff after the jump…

Physical characteristics. MH370 was a Boeing 777-200ER. Its “zero fuel mass” (ZFM) was 174,000 kg. With 49,200 kg of fuel aboard, its takeoff weight was 223,200 kg. (We know the fuel aboard on takeoff at 16:41 thanks to Paul Sladen’s deciphering of ACARS data shown briefly onscreen during a CNN segment. Note that in a press statement Malaysia Airlines indicated that the fuel load on takeoff was 49,100 kg.) UPDATE: Thanks to the October ATSB report, we now know that the fuel remaining at 17:07 was 43,800 kg.

UPDATE 2: Don Thompson has rounded up four publications which contain a wealth of 777 technical information: Boeing 777 Flight Management System Pilot’s Guide, Qatar Airways 777 Flight Crew Operations Manual, United Airlines 777 Aircraft Maintenance Manual/Satcom System, and Honeywell Multi-Channel SATCOM System Description, Installation, and Maintenance Manual.

Communications. In addition to a traditional transponder for use with ATC secondary radar, MH370 was equipped with ADS-B equipment that was operational the night it disappeared. The plane was equipped with VHF and HF radios for voice and data communication, which could also be sent and received via a satcom system that relied on one low-gain and two high-gain antennae mounted near the rear of the aircraft. (Specs, courtesty of Don Thompson, here.) These antennae were connected to a Honeywell/Thales MCS6000 satellite communications system located in the ceiling beneath them; this unit received location and velocity information needed to aim the high-gain antenna and to precompensate the transmission frequency via ARINC cable from the Inertial Reference System in the E/E bay. After the plane disappeared from primary radar, Malaysia Airlines made three attempts to reach its crew via satphone, but the calls did not go through; Don’s signal analysis of the three attempted phone calls suggests that the high-gain antenna might not have been working properly, perhaps because the antenna was not steered correctly.

Wind speed and temperature aloft. Stare at this for a while if you want to. If you like your data a bit rawer, you can find historical radiosonde data at the website of the University of Wyoming. For a more granular idea of what the weather was doing on the night in question, Barry Martin has compiled a large table of reanalyzed weather-model data from NOAA here.

Speed. As part of his paper detailing his estimate of where MH370 might have gone, Dr. Bobby Ullich has produced an impressive analysis of MH370’s speed before it disappeared from radar. While I’m agnostic as to the correctness of Bobby’s conclusions, I think he makes an excellent point with regard to the plane’s speed, which is that it clearly accelerated after the diversion at IGARI. The ground speed before the turn was about 470-474 knots, after, it was around 505-515 knots. Given that the winds aloft at the time were somewhere around 20 knots from the east-northeast, this would be broadly consistant either with an acceleration in airspeed or with a steady airspeed in the range of 490-495 knots.

 

Bobby Ullrich speed values

In his ongoing analysis of MH370’s performance, Barry Martin points out that a likely speed for the plane to fly would be “Long Range Cruise,” or LRC, which can be selected through the flight management system. LRC is faster than the Maximum Range Cruise speed and 1 percent less fuel efficient. To quote a Boeing manual: “This speed… is neither the speed for minimum fuel consumption nor the speed for minimum trip time but instead is a compromise speed somewhere in between. It offers good fuel mileage but is faster than the maximum range cruise speed.” LRC is given as a Mach number, and varies with weight. At MH370’s takeoff weight, LRC at 35,000 feet would be Mach 0.84, which translates to 481 knots in a standard atmosphere. At the time, however, the temperature was 11 deg C higher than that of a standard atmosphere, so its true airspeed would be 494 knots.

It’s worth noting as well that Brian Anderson has devised an entirely different means of calculating airspeed, based on the observation that between 19:40 and 20:40 the plane reached its point of closest approach to the satellite; by calculating this distance, and estimating the time at which it occured, one comes up with a groundspeed that turns out to be, by Brian’s (and other’s) reckoning to be in the neighborhood of 494 knots. Brian observes that “by removing the wind vector, the answer becomes about 486 knots TAS.”

Richard Godfrey has run the numbers for the early part of the flight and come up with slightly different figures from Bobby Ullich.

The last ADS-B data shows a speeds around 471 to 474 knots. Last calculated Ground Speed was 474.3 knots. The average Ground Speed required to follow this path from the turn back point and get to Pulau Perak by 18:02:37 for the start of the Beijing Radar Trace is 510.7 knots. The difference between 474.3 and 510.7 is accounted for by an 18 knot head wind that becomes an 18 knot tail wind after turn back. The wind in the area was around 18 knots at the time. This would make the Air Speed 492.5 knots. The Ground Speed required to get from the start to the end of the Beijing Radar Trace by 18:22:12 is 503.6 knots.

He adds:

The major turns and turn back flight path occur at borders between Malaysia, Vietnam, Singapore, Thailand and India. Indonesian Airspace is carefully avoided in the Malacca Strait. The major turns are just out of range of the Malaysian, Thai and Vietnam radars. The Satcom Login at 18:25:27 is just 14 seconds after reaching NILAM which represents the point just out of range of the Malaysian and Thai radars.

Performance. As the plane flew along, it burned fuel, and thus became lighter. As a consequence its optimum altitude — that is to say, the altitude at which it would experience the greatest fuel efficiency — became higher, and its LRC at a given altitude would become lower. Additionally, as the plane moved to higher latitudes, the air would have gotten colder, which would reduce its true airspeed for a given Mach number. All these factors would tend to gradually reduce the measured ground speed of the plane, which is indeed what we see geometrically for straight-line flight through the ping rings. For more on aircraft performance, see Barry Martin’s excellent Analytic Fuel Flow Analysis.

The Satellite. From 18:25 onward the sole evidence we have of MH370’s fate comes from the analysis of a handful of electronic exchanges between the plane and Inmarsat satellite 3F-1, which occupies a geosynchronous orbit above the equator at 64.5 degrees east longitude. Its position was not fixed; two years before, due to the fact that its hydrazine thrusters were getting low on fuel, Inmarsat had begun to let its inclination slowly increase. By March 7/8, it had reached an inclination of 1.7 degrees. Paul Sladen has published a table of ephemera. Here is a chart produced by Duncan Steel, showing the progression of the subsatellite point during the course of MH370’s final hours (click to enlarge):

duncansteel.com:archives:362

The Search. Via Don Thompson: As announced at a JACC press conference 28th April, on the occasion of the end of surface search, “Australia has been coordinating the search for 41 of the 52 days since MH370 went missing. In this period, more than 4.5 million square kilometres of ocean has been searched. There have been 334 search flights conducted, an average of eight a day for a total of over 3000 hours.”

On September 24, 2014, the ATSB announced that “over 106,000 square kilometres of the wide search area have been [bathymetric] surveyed.”

Inmarsat Raw Data and ATSB report. For two months after MH370 disappeared, members of the press and the general public begged and pleaded for the authorities to release the raw data logs of transmissions between Inmarsat and the missing plane. On May 27, 2014, they finally did.

In June, the Australian Transport Safety Board released a report (later updated) that provided even more useful information, this time explaining how the raw data had been interpreted. More recently, Inmarsat’s Chris Ashton was the lead author of a paper in the Journal of Navigation explaining how the company conducted its analysis.

Thanks to these documents, we now have a much better understanding of what transpired, and have the wherewithal to undertake a critical assessment of the official investigation–which, as I described in my last post, seems to be paying off.

Burst Frequency Offset is a measure of how the signal received by the satellite from the airplane has been shifted by various factors. You can measure how closely a prospective route matches the values recorded from MH370’s actual flight by using Yap’s BFO calculator.

End of the flight. The BFO data associated with the final “half ping” at 0:19 is anomalous in comparison to the preceding pings; it values that could not be generated by any combination of speed, location or heading that is physically possible for a 777. The data is compatible with a steep descent into the ocean at an acceleration of 0.7 g, which Mike Exner, Victor Iannello and others have interpreted as a spiral dive resulting from the fuel tanks running dry. There is some dispute at present as to whether fuel exhaustion would result in such a dramatic maneuver. While plans to enlist a professional-grade simulator are underway, John Fiorentino reports that he has already researched such an experiment, and says that the plane did not spiral dive but instead descended wings-level in a phugoid oscillation, that is to say, with the plane pitching down and gaining speed, then pitching up and losing speed, then pitching down and gaining speed, and so on. I’ve excerpted his report here.

More to come…

 

 

 

344 thoughts on “What We Know Now About MH370”

  1. Nihonmama: why, you clever thing, baiting me out into the open like that ;-). And then parsing my statements like a pro…

    I begain by addressing one point and point only, which I will repeat: there aren’t any indications of US prescience re the SIO location as per the allusions of the Sandilands article. I don’t know what the truth of the matter is; I know only that there isn’t any evidence to support US foreknowledge.

    As for the US being aware of military radar data and other military activities on the part of the Thais, the Malaysians, the Indonesians or the Australians for that matter, of course we can expect that they were informed at some level, as the provision of information or the installation or the sale/purchase of military equipment nearly always involves some sort of quid pro quo between the parties involved. I am glad that Victor outlined things so succintly, as he illustrates how things actually function in this domain.

    In conclusion (i.e., I am torching this particular horse), I think the initial US response is quite telling, in that, whatever level of concern was inititally brought to the matter, it seemed to ebb quite quickly, as if they had grown quite comfortable in knowing of the disposition of MH370, even while the general public remained perplexed. There IS a telltale “pall of silience”, and it could very well represent a “paralell investigation” into a convoluted terrorist threat; I stated as much a few weeks ago. But then it could be along the lines that Victor has alluded to: the US promised to keep quiet in exchange for a few extra cartons of cigarettes and another case of whiskey.

    BTW, I forgot to mention that I had a flight with a 777 airframe a couple of weeks back, where i distinctly recall the Captain exiting the cockpit to tap down a floor hatch in the forward galley. It could be wholly unrelated to anything, but I fly again on Sunday and will be sure to pay attention as to whether this perhaps has become SOP on flights in general.

    Nihonmama: I’ll ping you in a sidebar exchange.

  2. @Rand:

    LOL.

    The alternative name for this entire unfortunate event could be Quid Pro Quo. Who got what and who got got will unfold in due time, I suspect.

    The Captain tapped down a floor hatch on your last flight? By the galley (on the 777) is where the E/E hatch is found.

    I look forward to the sidebar. 😉

  3. Nihonmama – Mike Exner’s line – “Not even close to true. The ISAT data is excellent, and in any case, the best we have. Pls stop the propagation of such fiction.”

    We’ll know before too long how excellent it is I guess. I was thinking that if you have no technical visibility of what happened on the plane or to it, including the SDU, or any of it really, and you know there was human interference, it can’t be excellent data?

    Sounds to me like the pressure is building?

  4. Greg Yorke – Questions about the confidence level with the SDU never seem to go anywhere? Another one of Nihonmama’s elephants.

    Victor – It was never lost on me that where Inmarsat had the plane loitering for an hour(Aceh) is possibly the most hard line Islamic enclave in SE Asia. Where they join the data(radar to Sat)they also have the plane making landfall. And because of incredibly lax laws the Maldives is a money laundering hub for global Islamism. Actually the laws seem designed for it?

    Rand – the US certainly did seem pretty relaxed about it all, then they come out with the aircraft-centric Khorosan group?

    Nihonmama – the Chinese govt care a lot more about what their people think than they care about their people. The finger pointing may have been for show?

  5. @Matty – Perth: Just to be clear regarding the loitering around Sumatra. There are multiple paths satisfying the BTO and BFO that require various loitering durations. For instance, the path calculated by the IG requires no loitering and ends around 37.7S latitude. In my Banda Aceh scenario, there is time for a landing/takeoff and the path ends around 34S. There was about an hour of loitering required for the end point of 30S, which is where the ATSB report recommended search should be centered. (It seems the priority search area has moved south since then.)

    This is why the radar data around Sumatra is so important. It would help narrow the search area and also provide important clues for determining the cause of the disappearance.

  6. Matty: I hear you, and Nihonmama, too. I have just been considering whether perhaps a justified quiet pursuit of suspected perps would not by now have given way to their surfacing or perhaps a drone strike made public. Khorasan does appear to have nodes of connection in Malaysia, so it’s certainly not far-fetched to consider their involvement, and then given their emphasis on weaponized aircraft. It’s certainly on the table, and again I think it’s good that we continually serve up different possibilities, so that others here can impeach whatever we happen to argue for the day.

    Yes, the Indonesian radar data…

  7. Hi Jeff,

    I have completed two new white papers related to my previous work in fitting a unique route for MH370 (“The Location of MH370”, September 25, 2014).

    The first Addendum is “Error Sensitivity Analysis”, dated September 28, 2014. It can be downloaded at:

    https://drive.google.com/file/d/0BzOIIFNlx2aURUh3ckc2SkRDSEk/view?usp=sharing

    The second Addendum is “Comparison of MH370 Route Synthesis Methods and Results”, dated October 5, 2014. This third white paper provides my responses to the Independent Group’s critique of my first white paper. It also has a map showing the recommended search areas (by me and by the IG)along with the ATSB bathymetric survey. My addendum can be downloaded at:

    https://drive.google.com/file/d/0BzOIIFNlx2aUOEd1UmtlbFVJN00/view?usp=sharing

    Bobby

  8. @Matty, Rand, Victor:

    “Questions about the confidence level with the SDU never seem to go anywhere? Another one of Nihonmama’s elephants.”

    +1. Sure seems that way, doesn’t it?

    As I read all of your comments, I’m looking at a silver bracelet that I bought down in Koh Samui (Thailand) some years back. I wear it regularly and always get lots of comments about it – because it’s covered with elephants, all in a line.

    Elephants, like dolphins, are the seers in the animal kingdom. And you want them around — because they warn of, and protect others from, danger.

    The SDU, elusive Indonesian radar data, Khorasan (which as, Jenan Moussa pointed out “‘Khorasan group’ as name doesnt exist on ground in Syria. U.S used this name to refer to ‘Wolf Unit JaN’, elite unit of AlQaeda in Syria” — and a whole slew of other hanging chads.

    Elephants.

  9. @Matty – Perth: Mike Exner is correct that the BTO and BFO data is very clean and is well fitted by relatively simple aircraft path models. It is hard to see that failures in the SDU that has to compensate very large inherent Doppler offsets would result in such coherent data. Similarly, some malevolent power on-board trying to emulate signals that gave the impression of a simple Southerly path would have a very hard job – in practice impossible.

    The area that Go Phoenix is now scanning corresponds in my modelling to close to due South tracks from the final turn just before 18:40. These tracks are good fits to the BFO data, allowing for the errors. There are of course other possibilities but the range of all well-fitting tracks is relatively small. If the track taken after the final turn was very simple (for whatever reason) then the search could be short. This is all very speculative, but it’s going to be a long wait on this blog otherwise.

  10. @Dr. Bobby:

    Thanks much for posting your whitepapers /responses. Reading and have tweeted your link.

  11. @Matty, @Richard,

    I have a lot of respect for Mike Exner, for getting a lot of this discussion moving in the early days, but I take issue with any characterization of the data as “clean,” “accurate,” or “excellent,” by him or anyone else. I don’t want to misconstrue Mr. Exner’s comment, but even he admits it’s the best we’ve got.

    The last time we have both confirmed location and BTO data is at 17:07. This is a little over 30 minutes and a little over 100 miles from KL.

    There is not a single, published, data point with both a BTO value and a verified location outside of this 30 minute window (60 minutes, if we count the ground time.) Not for this plane, not for any other plane. There is no logical explanation for the absence (in the public forum) of BTO data for prior flights or from the later sample flights.

    The values within this window correlate poorly to the trip. They fluctuate for some yet unknown reason. They are rounded and reduced for some other unknown reason. The subsequent radar locations are no help either – nobody even knows what the true coordinates are because the image is erroneous, and nobody will even admit who’s radar it was.

    The work of the IG and Exner is excellent, and it truly is the best we’ve got. But the reality remains – they’ve taken 30 minutes of poorly correlating data and extrapolated it over the next 6 hours and thousands of miles. The data simply isn’t “excellent” for that purpose.

    @Richard – I know you disagree, so this is a light-hearted challenge: show me why you think there’s a good correlation between the BTOs and distance through 17:07. I’m aware that the value increased to 15,600 by 17:07. Can you statistically show that this increase is dependent on the distance variable? I can’t – a simple regression test shows otherwise.

  12. @Rand, Matty:

    Language is a beautiful thing.

    “Upon receiving the raw data, the Malaysian authorities immediately discussed with the US team how this information might be used. The US team and the investigations team then sent the data to the US, where further processing was needed before it could be used.”

    http://www.duncansteel.com/archives/751#comment-4241

    How this information might be USED.

    Hishammuddin Hussein’s highly nuanced and indirect communication style (informed by both his Malaysian and British education), is something to behold.

    See also the Japanese. #Nemawashi
    (http://japanese.about.com/library/weekly/aa080597.htm)

  13. @JS – Clearly, the models being used by various parties outside the official investigation cannot be verified since the data (BTO, BFO and navigation data) from other flights has not been released. So we are only able to use the available information to build up the elements of a model, with the background of basic physics. On the specific point you raise, the released method for interpreting the BTO numbers can be used to derive the distance of the aircraft from the satellite at 17:07UT, and that corresponds to the known position of the aircraft. We aren’t trying to derive the model from the data, so I am not sure where regression is relevant. Could you please elaborate on that point?

  14. @Matty, Rand, Victor:

    TODAY:

    Al Qaeda attempted to hijacking the warship PNS Zulfiqar

    “The attackers were former Pakistan Navy officers-turned-jihadists,” said its spokesman Osama Mahmood. The plan was to hijack PNS Zulfiqar to launch an attack on a U.S. aircraft carrier.”

    FLASHBACK (with the caveat that I am NOT a consumer of FOXNews, which has not challenged MCInerney. And he has never backed down from his assertion):

    Lt. Gen. McInerney has a source confirming Mh370 is in Pakistan – sources have BOEING info

    https://twitter.com/nihonmama/status/447476462631649280

  15. @Bobby: very well written article.

    I must admit to siding with the IG initially on the “Occam’s Razor” aspect (if a pilot has the capacity to turn toward an airport, why no descent?), but your rebuttal helped me understand that this could just as easily have been a simple waypoint choice – from which no intent to land need be inferred.

    More fundamentally: point #7, Table 4.1, p.8 should be required reading for anyone paid to turn data into information. Bravo.

    MOST fundamentally: I fear we’ve all been denied some fundamental truths – and in our starvation, we are gnawing away at each other’s work. I hope I can convince both you and the IG – and everyone else – to take a break from skeet-shooting each other’s speculative analysis, and join up with me in my quest for the members of the JIT to emerge from the shadows, identify themselves, and open up their agencies’ MH370 models to public scrutiny.

  16. @Brock McEwen

    “take a break from skeet-shooting each other’s speculative analysis, and join up with me in my quest for the members of the JIT to emerge from the shadows, identify themselves, and open up their agencies’ MH370 models to public scrutiny.”

    I couldn’t agree more and Bobby Ulich is a very intelligent man who’s put much work into this.

    The problem is though, some people simply aren’t looking for cooperation, and some claim to “know” the “unknowable.”

    That is why I am currently pushing the ATSB very hard re: this idea of “narrowing the width of the search area.”

    The IG and Exner along with Dr. Ulich seem to think it’s a good idea and use basically the same scenario to come up with their recommendation.

    For myself, I believe If adopted, this recommendation increases the chances the plane will never be found. The idea is contrived, specious and without merit.

    I offer very good reasoning (I believe) to support this in my two responses to the IG, Exner and Bobby.

    No skeet shooting here, just concern.

  17. Dr. Bobby Ulich:

    Hi, I’ve read Duncan’s post on your White Papers and am very interested in your analysis. But it’s really difficult to download the files in China, probably because of Google Drive, probably because of the browser I use. Would you please share your papers on Dropbox or somewhere else? Maybe a direct download link to the pdf files? Thank you very much for sharing.

  18. @Richard – Sure. Several reasons.

    First, just because you get the right answer does not mean you have a good model. If I ask my 6yo what 2 times 2 is, and he adds, he’ll get the right answer. However, when I ask him for 8 times 8, and he says 16 instead of 64, it becomes clear that he is using the wrong method. Here, the model predicts the 17:07 location. But that is not evidence that the model or the data is properly understood, or that it can predict beyond 17:07.

    Second, you have many other variables which all trend upward between 16:40 and 17:07: airspeed, altitude, flight time, etc. You have others that trend downward, but remember a correlation can be inverse – for example airspeed is inversely correlated to air density. Those variables are fuel remaining, air pressure, temperature, etc. Finally, you have variables outside of the plane, such as the satellite position and velocity.

    The regression helps show that your dependent variable (BTO) is correlated to your independent variable (Distance from plane to satellite). If they correlate, you can discuss causation. But right now, you have I think 19 points, and they don’t correlate. When you don’t have a correlation, the obvious conclusion is that some other variable is influencing the result.

    In this case, you have a somewhat fluctuating BTO, even when the plane is stationary. Statistically, you have no evidence that this fluctuation, alone, did not cause the result that matched the location.

    Others have gone as far as to speculate that signal reflection is the other variable, which makes a lot of sense. But my point remains – notwithstanding the fact that the model correctly predicts ONE verifiable location, we’ve extrapolated really far with only a few points that don’t tie together well.

    That becomes more disturbing when we consider that the data that would prove or disprove this model is already known, but being intentionally withheld. In my opinion, withheld information is generally adverse to the party withholding it.

  19. Richard – I’m personally glad that we have an IG and I appreciate all the effort, but it has just occurred to me that having an IG at all is a funny old arrangement?

    The fact that the data is coherent doesn’t say all that much to me. IF…..an act of state terror required that a clean route needed to be created I’m inclined to think some software could be written to do it. Inmarsat are claiming to be able to model the BFO data with a bare sniff, so how would a resourced, determined, planned endeavour to imitate a route be “impossible.”

    Alternately, my current understanding is that the numbers are not replicable or verifiable in a hard way, but are considered good data because it looks coherent? If it only has to be coherent I’m thinking there would be geeks out there to do it in a canter? Not quite the same I know but air to air missiles are using doppler shift to hit planes. It would have been thoroughly investigated over the years with substantial budgets as a means of tracking aircraft beyond radar, suddenly it’s up there with a lunar landing.

    I hope you are right actually, and that it shows up soon but it needs to for any of this to be right? Some of the talk out there reminds me of PM Abbott burning himself when they thought they had it in the hairs. I’m just not as confident as you.

  20. Nihonmama – that Pakistan story is amazing – in that it wasn’t a major story? “Details are now just beginning to emerge” they say. Why wouldn’t they be coy about it, or the Malaysians for that matter? What a shocking trend?

  21. @Matty:

    Amazing indeed.

    And here’s the kicker: WSJ appears to have been first out with it THREE WEEKS AGO ago — on Sept 16 – then Reuters and other followed. To your point, how this was not all over the news (worldwide) is THE question.

    Here is the very detailed Newsweek piece where I first saw this story — dated 10.07 (and apologies it was left out of the earlier post) http://t.co/w3UM5VZQZB

    Again, here’s the Twitter convo from 03.21, wherein I referenced the Jakarta Post article (with the then unnamed USSOCOM ‘Wolfhound’)
    https://twitter.com/nihonmama/status/447077518214840320

    You get me now?

    Way too late, must hit the bed. The birds are beginning to chirp!

  22. @ALL

    ATSB has NO plans to “narrow the width of the search area”

    The latest analyses indicates that the underwater search should be prioritised further south within the WIDE (emphasis mine) search area for the next phase of the search.

    nvestigation number: AE-2014-054
    Investigation status: Active

    Flight path analysis

    Published: 8 October 2014

  23. Nihonmama – and you have the June 10 Al Qaeda attack on Karachi airport in Pakistan also. Was that an attempted hijack? And a similar one in Libya shortly after? Seems like an upswing in that sort of activity.

  24. @ALL

    Investigation number: AE-2014-054

    Flight path analysis

    Published: 8 October 2014

    Analysis varied with their consideration of BTO/BFO at 0011 and 0019. The last time associated with a log-on interrogation request was at the 0011 arc whilst the 0019 arc was a log-on request initiated by the aircraft which had an effect on the BFO.

  25. John Fiorentino posted October 8, 2014 at 7:03 AM:
    “ATSB has NO plans to “narrow the width of the search area”

    Of course not, the ATSB understands airplane dynamics.

  26. Perhaps some have missed this in the ATSB Summary:

    The simulator activities involved fuel exhaustion of the right engine followed by flameout of the left engine with no control inputs. This scenario resulted in the aircraft entering a descending spiralling low bank angle left turn and the aircraft entering the water in a relatively short distance after the last engine flameout. However when consideration of the arc tolerances, log on messages and simulator activities are combined, it indicates that the aircraft may be located within relatively close proximity to the arc.

  27. @MikeExner

    “Perhaps some have missed this in the ATSB Summary:”

    No, I don’t think so, at least I didn’t.

    A “descending spiralling low bank angle left turn” is NOT a “spiral dive.”

    And, “it indicates that the aircraft *MAY* (emphasis mine) be located within relatively close proximity to the arc.”……..DOES NOT indicate within 1nm as stated by the IG.

    I’ve seen this in sims, though no one wanted to listen. I published a max bank angle of 25 degrees……a bank angle is indicative of a TURN. A LOW BANK ANGLE is indicative of a WIDE TURN. — NOT A RAPID DESCENT as apparently indicated by the “affected” BFO value.

  28. @ALL

    “though no one wanted to listen.”

    I take that back, JEFF WISE was nice enough to post an excerpt from my response here on his blog.

    Thanks again Jeff!

  29. There is room for debate over the diameter of the spiral, but the best fit to the BFO data shows a ROC = -4765 ft/min at 00:19:29 and -15,258 ft/min at 00:19:37 (last data). This corresponds to a change of 10,493 ft/min in 8 seconds, or a vertical component of acceleration = 21.9 ft/sec^2 (0.68 G’s). If the BFO values were accurate, it means the decent rate was 173 mph at 00:19:37. This could only result from a much steeper spiral dive than what ATSB assumes. In direct conversations with ATSB a month ago, they were still not on board with any spiral dive. Clearly, they are just coming up to speed on the meaning of thee 00:19 BFO values. They were not even looking at this until we brought it to their attention. The vertical Doppler signal is so huge that even if the AES OCTCXO did drift a few Hz while the power was off from 00:15 to 00:16, it would only change the results a few percent. In any event, it is not important if the spiral diameter was 2Nm or 5NM. The important thing is that the plane did not glide for 50-100 NM. ATSB clearly agrees with this now. This is very important to limit the search width to that which is necessary, and no more.

    There are other error margins needed due to the uncertainty in the ephemeris, exact altitude at 00:19:37, BTO Bias calibration error, BTO jitter, etc. All these factors add to about +/- 10NM…maybe a tad more. But that is a lot better than a 100 NM width, which is where we were a couple of months ago. Note that the Fugro Equator is now surveying the 7th arc in a band (covering the IG spot) about 20NM wide so far, compared to >50 NM last month when it was further north.

  30. @MikeExner”

    The important thing is that the plane did not glide for 50-100 NM.__ ATSB clearly agrees with this now.__”

    Not so sure either is completely true.

    (BTW, 50 miles is really not that far in the scheme of things, neither is 40, or 30 or 20. None of which is 1.

    And yet the search will remain in the WIDE search area.

    “All these factors add to about +/- 10NM…maybe a tad more”

    But that is NOT 1nm as you stated in your report is it?

    “This is very important to limit the search width to that which is necessary, and no more.”

    The same pertains to the search length according to the ATSB and common sense.

  31. John: You misread the statements. We never suggested the search width should be reduced to 1NM. What we said was that the plane spiraled in, probably in a very steep dive with a 1NM radius, right over the 7th arc. But there was and still is an uncertainty of about +/- 10NM in where the arc is, for all the reasons explained in the previous note.

  32. @MikeExner

    Sorry, but you’ve misread MY statements.

    I NEVER said you indicated “the search width should be reduced to 1NM”

    Where do you get these ideas?

    Perhaps you should read my responses, which I will bet you really haven’t.

  33. For me the important thing is that the ATSB is not fixated on any ‘unique’ scenario, but keeps an open mind towards all scenarios that are compatible with the evidence.
    The BFO in the very last transmission received from the airplane at 00:19:37.443 may be evidence of high V/S, but perhaps it is not.

  34. John: What did you mean by “…But that is NOT 1nm as you stated in your report is it? …”. Sure sounds like you believe the IG suggested a search width defined only by the spiral width. Please ack that we did not ever suggest this.

  35. @Matty – You said, “Not quite the same I know but air to air missiles are using doppler shift to hit planes.”
    Depending on your definition of “hit,” this might not be accurate. I’m pretty sure that in the case of a near miss, (i.e. not a direct hit) both air-to-air and SAM’s use doppler to detonate the explosive rather than locate the target. As soon as the doppler shift is detected, it means the missile is now going away from the target and is therefore as close as it will get so it is detonated. The doppler could be audio (of the target’s engines) and/or radar reflection. A direct hit is not always required to bring down the plane.

  36. @MikeExner

    One more example as a point of reference and then I’m done with this particular issue for now.

    “And, “it indicates that the aircraft *MAY* (emphasis mine) be located within relatively close proximity to the arc.”……..DOES NOT indicate within 1nm as stated by the IG.”

  37. From paragraph 5.3 of this Technical Paper from Inmarsat engineers:

    journals.cambridge.org/nav/mh370

    (copy/paste into the URL window of your browser)

    “This means that we can use the Logon Request message information from 18:25:27 and
    00:19:29, but it is prudent to discount the measurements between 18:25:34 and
    18:28:15 inclusive, and the one at 00:19:37.”

  38. @ALL

    DUNCAN STEEL SAYS…….not “much actually new” in ATSB Report

    New ATSB Report Appears
    Duncan Steel : 2014 October 08

    Well, it’s a new (update) report from the ATSB, but it doesn’t contain much actually new, because the Independent Group had already reached essentially the same conclusions and published them here a couple of weeks ago (and in preceding posts). One wonders whether the two might be related?

    SO, I guess the IG has backed off its idea of “reducing the width of the search area?

    NO< that can't be right, can it??

  39. @Matty – Perth – As they might have said in the 1950s ‘I am not now nor have I ever been a member of the IG’. The IG will have to speak for themselves, my interest has only been to understand what is going on, not second-guess the investigation.

    >so how would a resourced, determined, planned endeavour to imitate a route be “impossible.”

    Hmm – I think we are in the realms of Mission Impossible for the bad guys to hack into the onboard system during flight and upload new code that guarantees a seamless transition to fake data that replicates another aircraft path with all the correct signal characteristics (given that a lot of the contributions to the BFO are outside the aircraft). How would they know that the code worked in test without looking at the recorded BFOs and BTOs in the Inmarsat log? If I was organising such an operation I would just turn the system off, not rely on a complex spoof that might not work.

    >… my current understanding is that the numbers are not replicable or verifiable in a hard way.

    They numbers are certainly replicable, in that they correspond to a small set of southern routes (within the errors). Even if the model was completely proven on other flights, there is a single set of data from MH370 so it would not possible to verify the data itself (what would be used for that verification?).

    I will leave the Aussie politics to you. But as you say, the proof will be if and only if the aircraft is found. If it never is this story will run longer than Amelia Earhart’s.

  40. @JS – thanks for the clarifications. I think we are stuck, since a full detailed verification using aircraft data independent of MH370 is not available. The model cannot be derived or proven from the MH370 data (since there are only a few MH370 points with BTO/BFO and navigation data).

  41. @Richard – we are stuck, not because the data doesn’t exist, but because it’s being withheld. There should be months or years worth of BTO data from 9M-MRO with verifiable ADS-B locations.

    @John and Mike Exner – does it matter how wide the search area is? Wouldn’t the search start at the highest probability location for any possible dive/glide scenario, and wouldn’t the highest probability location always be the last known location (00:19 in this case, not 18:22)?

    I understand the differences of opinion but does it impact the search in any meaningful way? Considering they’re probably going to keep expanding the search area until they find the plane, regardless of how big the initial search area was?

  42. @Richard – sorry, but I can’t leave this one alone. You wrote:

    “Hmm – I think we are in the realms of Mission Impossible for the bad guys to hack into the onboard system during flight and upload new code that guarantees a seamless transition to fake data that replicates another aircraft path with all the correct signal characteristics (given that a lot of the contributions to the BFO are outside the aircraft). How would they know that the code worked in test without looking at the recorded BFOs and BTOs in the Inmarsat log? If I was organising such an operation I would just turn the system off, not rely on a complex spoof that might not work.”

    I disagree with some of your conclusions: that the transition was seamless; that the signal characteristics were even available; and that turning the SDU off would be better for certain perps than a spoof that sent everyone to the SIO; or that a perp would need to “upload code” rather than just feed the SDU bad location data from a laptop.

    At the core, though, I think what you’re suggesting is that a spoof could not generate such coherent data. But let’s turn this all around because I think you’re introducing a major bias.

    If that’s true, then a similar model should be able to differentiate between a real set of values and a fake set. If I picked 7 BTO values and 7 BFO values, could the model determine that those values were fiction, OR would the model do its best to identify a straight line path that fit my made up values?

    In other words, is the path driven by the data coherent because it’s real, or is it coherent because the model specifically set out to find the most coherent path to fit the data? I think the latter, within reason. The whole purpose of finding a “fit” is to make the data appear coherent. You can’t go back afterwards and say that because we made it look coherent, it must be a bonafide set of data.

    So I would challenge anyone with a model (and the time) to produce a best-fit path for a set of bogus BTO and BFO data points. My belief is that as long as there aren’t wild swings in either, you’ll get a flyable route if you try hard enough.

  43. JS and John: the width of the search area is actually of primary concern, as, of course, the width of the search area is fundamental to defining its area. From here, there is km2 cost associated with the time-sonsuming use of rare and fragile sensing devices/towables that will be deployed in Phase II of search. There are resource constraints inherent to the search that are likewise of primary importance, all compounded by the element of time and the relative interest level. Consider the loss of a towable submersible; it’s not as if one can run down to your local DIY store for another. And then there is is the issue of funding, now provided from the Aussies and the Malaysians, which is a dereivative of the interest level. Yep, defining the search area is damn important, which is why the IG has attempted to refine continually its analsyses to better define the search area.

    Thousands of km2 of ocean, 2000km from shore, with the water thousands of meters deep…there are basics elements inherent to the search that need to be viscerally apprehended.

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.