The ATSB has just issued a lengthy and detailed new report explaining its latest thinking regarding the underwater search for MH370 in the southern Indian Ocean, available here. For the most part the media are reporting that its basic point is that the current search area is the right one. While that’s true, there are some more interesting points buried within in it, and within its companion volume from Australia’s Defence Science and Technology Group (DSTG) which explores the logic in further depth.
Here are my bullet points—I invite readers to add their own, or to correct or elaborate upon my points.
— One of the most jaw-dropping revelations in this report is that after 18:01:49 there was just a single radar return point. Note only does this contradict data shown to passenger family members soon after the disappearance (see Victor’s note below, and the image above), it also raises questions about the reliability of that piece of radar data. Since it was detected at the far limits of the radar equipment, it is relatively inaccurate, and as a stand-alone piece of data it is much more likely to be anomalous.
— The report reiterates that the only way to deliberately depower the SDU is by pulling circuit breakers in the E/E or isolating the left AC bus from the cockpit, but offers no explanation of why this might have happened prior to 18:25.
— It turns out that the time to recycle SDU is not 2.5 min but only 60 seconds. This is particularly important when it comes to laying out an end-of-flight scenario that presumes fuel exhaustion.
— The report says that after fuel exhaustion ditching not possible, with or without a conscious pilot. There has been a great deal of debate about the possibility of ditching in this forum, and I hope (but doubt) that the report will lay the issue to rest.
—At last, we know the cost index for the initial portion of the flight: it was set at 52. Of course there is no reason to assume that the later portion of the flight was conducted at this setting but it helps us calibrate likely flight modes.
— Overall point: With these two documents, the Australian authorities have shed a commendable quantity of light onto the subject of how they have determined the likely flight paths that MH370 took after it disappeared. It is heartening to note that they have greatly shrunk the length along the 7th arc along which the plane might plausibly lie: by my reckoning, from about 630 nm to 380 nm. And the “fried egg” of maximum probability is smaller still, only about 150 nm long. However, I find it baffling that, given the incredible level of effort poured into figuring out how the plane might have traveled prior to fuel exhaustion, there seems to have been basically zero time spent figuring out how the plane would likely have traveled after the fuel ran out. Frankly, I was expecting a lot of analysis along the lines of Brock McEwen’s work on this topic. As it is, it seems that instead of examining flight modes they took a guesstimate from accident-investigation experts and added a fudge factor. The result is that, while this latest analysis shrinks the search’s target area on one dimension, it makes it fatter on the other. In current southern search area, they’ve looked about 18nm inside, 30nm outside 7th arc. According to the new report, they should expand the search box to a width of 80 nm, symmetric around the 7th arc. This is not progress, and I think the ATSB can do much better (and hopefully will in a future report.)
Some additional points from Victor Iannello
Before I could post the above thoughts, Victor emailed me some observations of his own, which I include here unedited:
“There are some very strange results reported starting on page 17. Here are some comments related to that and the BFO bias:
1. As the attached graphic illustrates, if we are to believe that primary radar data exists every 10 seconds up to 18:02 and then only a single capture at 18:22, the slide presented to the NOK on Mar 21 at the LIDO hotel is false or includes data not used in the report.
2. There is a statement that the ground speed observed by the radar prior to 18:02 is relatively high and implies the aircraft would be at low altitude. While this would be allowed from Ma number and available thrust considerations, the indicated air speed would be extremely high, the airframe would be stressed, and the fuel efficiency would be incredibly wasteful. This is not consistent with the fuel calculations after 18:22.
3. The groundspeeds they calculate from the radar data have tremendous variability, even after the Kalman filter is applied. I estimate the peaks to be about 550 kn (!) We need the raw radar data to see what the hell they are doing.
4. The bias term was observed to be time-varying and modeled with a SD of 25 Hz. But there is also a statement that “Substantial effort was made to characterise this structured bias. It was found to have a geographic dependency but it has not been possible to determine a quantitative function to compensate for this change in bias.” This implies the drift might not be a simple OCXO drift issue. In fact, perhaps it is the correction term that has a geographic dependency rather than the oscillator drift.
Overall, I am not sure their work adds much value over the deterministic approach that the IG and others has followed. The PDFs and assumptions on heading/speed changes of previous flights are practically irrelevant.”
He later added:
“Their approach of randomly spaced turns, accelerations/decelerations, and climb/descents will always favor straighter, more constant speed paths that fit the BTO data. Since commercial flights are relatively straight to conserve fuel, it predicts those flights relatively well. If MH370 flew relatively straight, it should work there, too. But there are no guarantees the flight was near straight. For instance, if the flight flew south, a circle “loiter” above Sumatra would be ranked low, as would a curved path that followed the coast of Sumatra. There is bias in their model that is not acknowledged. Also, fuel calculations are only indirectly included by limiting the range of speeds. The model seems overly complicated for the value of what is produced. It seems developed more to impress than to enlighten.”
UPDATE: 12/4/15
Reader Paul Smithson asked “how much of the newly-defined priority area has already been searched?” In the image below I’ve outlined in black the area already searched (via Richard Cole) on the “fried egg” map. (Click to enlarge) The original 120,000 sq km search area is outlined in red; the new 120,000 sq km search area is outlined in purple. As you can see, almost all of the high-probability area has already been scanned. As more and more is searched, the probability density of the area being scanned will decrease, so that search becomes ever less fruitful. The effort expended between now and the end of the scheduled search will, by my seat-of-the-pants estimate, increase the probability distribution coverage from aroun 85 percent to 90 percent.
Initial reactions:
1) More of a question: am I reading the report correctly when I infer that they are now considering the logon process initiated by engine 2 flameout to be taking only 2:00 now, and not the 3:40 +/- 10sec they had previously given? If so, did I miss the explanation they gave for this change?
2) Relative to the Oct.8, 2014 report, I don’t see any explicit change in inputs to the fuel limit determination in either of the two new papers – yet the ATSB is as we speak searching impossibly beyond the original paper’s explicit limits. Has anyone spotted ANY description of – or explicit CHANGE in – the current working fuel limit?
3) the “Bayesian” paper – which I’ve only skimmed – seems merely to re-present the Oct.8 report in much greater detail – with an addendum folding in the flaperon information. The assumptions on which their probability distribution is based:
a) the radar data is authentic,
b) the ISAT data is authentic, and
c) the floating debris record tells us almost nothing further
…complement my work very well – because everything I’ve done in the past few months has been an attempt to TEST the appropriateness of those assumptions. My work suggests – particularly if Florence de Changy’s buoyancy or Jeff’s barnacle/water line reporting are accurate – that those assumptions are mutually incompatible – they can’t all be authentic, or they’d have found deep sea and/or other surface debris by long before now.
4) The assertion that zero surface debris (OTHER than the flaperon) by Day 508 throws no suspicion whatsoever on the entire SIO theory is galling in and of itself – and becomes even more so when one realizes the paper is being proudly presented on Day 636. Either the 10,000 or so pieces of debris are HIGHLY correlated – in which case, pieces would have been spotted during March 2014 air & satellite searches, as well as during the intensive air search conducted in and around Réunion Island in August 2015 – or they are WEAKLY correlated – in which case, Western Australia would have been hit by late 2014, per every drift expert I’ve ever asked who hasn’t changed his story.
5) The WIDTH (distance from the 7th Arc) of the latest search zone is consistent with neither Mike Exner’s description of his flight simulator results nor with the common view that the BFO value at 00:19 indicated a roughly 15,000’/min descent rate. Has the IG stopped endorsing one or both of these, or is the ATSB’s new search width bogus?
In short: the science continues to be stretched to fit the observations, rather than revisiting the core assumptions. What else is new?
6) If the search area width is being expanded to accomodater “controlled glide” scenarios, why is the probability distribution being made MORE symmetric? Does anyone actually think a pilot ran out of fuel, waited a couple of minutes, then – within a minute either side of the 00:19 pings – turned around almost 180 degrees, and tried to maximize range back in the direction from which he CAME? Into a SIDE wind? Nonsense.
@Brock
I have so many with these reports that it would take a multipage post of chronicle them. What is new you ask? For me it is the path bifurcations shown in Figure 4 of the AE-2014-054 document and the additional commentary to be found in the “Bayesian_Methods…” document.
From page 62 of the latter we have:
Begin cut-paste//
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//
Wow. Another way of stating this is that the plane could be anywhere on the arc without dynamic constraints. I wonder where I have heard this before. Oh wait,…
Page 7 of “definition of…”
“The environmental data mentioned above, was provided by the Australian Bureau of Meteorology (BOM). It must be recognised that this data has its own limited accuracy, and a suitable probability distribution must be applied.”
Read this in the context of recently reported “massive and pervasive” IT systems breaches at the BOM and classified recommendations to wholly replace all its systems at a cost of several 100s of millions of dollars, one may question the reliability of that input data to the modelling.
@Brock
Re 60s: page 13 footnote 6:
“Previous advice, included in previous ATSB reports, identified this time as 2 minutes and 40 seconds. Further testing conducted by the manufacturer of the SDU has shown that approximately 60 seconds is the correct time.”
@DennisW: Jeff added some more comments of mine above. I believe there is bias in their model (intended or not) that favors straight, relatively constant speed paths. That is why their hot zone aligns with previous studies that assumed AP control.
The biggest problem with this study is the hot zone was already searched with no result.
@Victor
To your latest comment, and I am not trying to be mean here, I was scanning the references and noticed that [11] cited Bowditch “American Practical Navigator”. While I have a copy, I don’t think I have touched it in a couple of decades. I was curious what they could possibly be referencing from Bowditch.
Found this on page 38:
The direction of the Earth’s magnetic field is not aligned with the geographic poles and so angles measured with a magnetic compass are not the same as angles measured with respect to the geographic poles. The difference between the two angles is referred to as magnetic declination [11].
Speaking for myself, I would put this factoid in the common vernacular category, and never consider citing Bowditch. Don’t know why I found this so humorous.
Bowditch once visited Reunion Island. Very appropriate to cite his book.
Unfortunately, having cited Bowditch, the book bungles the description of how magntic and true headings on a modern aircraft are determined. p. 40:
“Note that aircraft instruments measure magnetic angles so this form of control [constant true heading] requires the aircraft to correct the measured magnetic heading angle for magnetic declination. This can be achieved with a correction table since the Earth’s magnetic field is well-characterised and varies slowly.”
Somehow the book has never heard of inertial reference systems. Further, the MagVar tables (which are used to go the other direction) are valid only for a particular epoch and are often out of date and inaccurate. I was hoping that the book would provide information on what epoch was installed on MH370, but to no avail.
Useful information that the book does provide include: 1) more detailed information on BFO errors; 2) quantitative coefficients for the OU processes that govern speed and heading control, including winds; and 3) various other facts and bits of information.
The biggest problem I have is with Figure 10.7, which shows the BFO residual measurements. The claim is that these residuals are consistent with the errors; however, that is not true. The model assumes that the errors are uncorrelated random Gaussian, but even independent of the Gaussian assumption, there is a strong correlation at the 99% confidence level. This problem is nothing new (it has been commented on here before) but seems to get glossed over or ignored. Why?
Does anyone have a contact with the ATSB? Has anyone been able to contact someone working there? We have so many great reports being done by 3rd parties that it would be great if they would use or at least discard. We were able to contact scholars, experts in all fields, Boeing etc… but not the ATSB?
Regarding the odd revelation regarding the radar data, I offer the following (sure to be wrong) conjecture.
It seems likely that two military primary radars tracked MH370: one at Gong Kedak on the NE coast, and one at Western Hill, Penang Island, off the NW coast. The conjecture is that all the data used in the book came from the first radar. The last point at 18:01 near Pulau Perak is 280 NM away – rather long range, but perhaps feasible. Data from the Western Hill radar may not have been used because of some problem with them. While the plane was detected up to 18:22, the position data may have been too inaccurate to be useful, other than to say that the plane appeared to fly a straight track up the Strait.
@sk999: The DTSG report clearly talks about a “penultimate” capture at 18:02. There is no mention of any data that is ignored other than ignoring the capture at 18:22. (For what it’s worth, I believe that Western Hill was the radar station that captured the targets in the Malacca Strait.)
The radar data is the final bit of evidence we have that (arguably) captured the definitive position of MH370. For this reason, all aspects of this data set need to be thoroughly understood.
@SK999
You have more than earned the right to be wrong. Conjectures are welcome.
I have been puzzled by the radar data from the getgo. I have never been able to make any sense of it.
@Jeff
I wrestled with this post, but could resist no longer. Your comment about putting the controlled ditch to rest based on the recent ATSB reports is about the dumbest thing you have ever said. You set a new bar on that one.
@Victor, how much of the newly-defined priority area has already been searched? And has the IG’s best estimate area (which seems to lie about 25NM north-east of the centre of the “fried egg”) been searched yet?
@MuOne: re: logon timing: thanks for the reference – appreciated.
1) the ATSB perhaps should have nailed that factoid down a little better before setting the SW point at which the Fugro ships – for 13 months – turned around, instead of searching a bit further. Ditto the fuel limit.
2) in and of itself, knocking 1:40 off of the logon time has two offsetting impacts:
(Full disclosure: winging it tonight – going from memory – always dangerous)
a) since MH370 now has only glided sans fuel for 2:00, instead of 3:40, it could conceivably make it another 10nmi or so in glide distance after crossing the 7th arc.
b) since MH370 now must endure to ~00:17:30 UTC (instead of the former ~00:15:50), its range limit is set by a path which must be slower than before. I defer to the avgeeks, but would expect this to require only a modest reduction in fuel limit-establishing speed; however, a very slight speed reduction leverages out a significant EASTward shift in the intersection point between the fuel limit and the 7th arc.
It would not surprise me if the net impacr of a) and b) is a shift EAST in the western border of fuel-feasible Impact points.
Finally got the GEMS report released via FOI:
https://drive.google.com/open?id=0B-r3yuaF2p72bDhUVTNobFRxeVk.
Even if it actually IS the original report (it is date-stamped Dec.1, 2015), and even if it is the ONLY information AMSA/ATSB ever received from GEMS – it is still pretty damning on its face:
– Figure 80 shows a major current system leading AWAY from Indonesia, which flies in the face of GEMS’ own conclusion
– By Oct, 2014, the ATSB priority search zone had already moved to the extreme SW end of the wide area
– extrapolation of Figs. 78 and 79 suggests strongly that, even with the error, Indonesia would be a very weak conclusion
– the report’s failure to map end-points back to their generating start-points should have raised a huge red flag: otherwise, how do we know the shoreline indication wasn’t generated by virtual debris starting its journey from a point 2000km NE of the actual search zone?
– truncation of all probability density West of 60°E longitude should have been yet another red flag
Why is this flawed-yet-trumpeted study important?
– publication of this study was politely requested by me (and others) 12 months ago; why was it suppressed?
– if the flaperon narrative is taken at face value (more on this later this week), this suppression may have caused key evidence to be missed/destroyed due to poor/late notification
– PRETENDING to believe in a wrong shoreline would have rid the Abbott government of the need to search legitimately plausible shores
– not searching plausible shores would have rid the Abbott government of having to explain why nothing was found there
I think former PM Abbott should be asked to explain why this study – and none of its more rational contemporaries (CSIRO, UWA, etc.) – was allowed to (mis)direct shoreline search resources.
What does the professor know about airplanes, or physics in general? Or any member of his team?
The authors express an opinion about ditching that illustrates their total ignorance of the subject.
The authors express an opinion about the effect of airplane attitude on the fuel available to the APU after main engine flame-out, for example in phugoid motion, that ignores physics.
In figure 4.2 of their book they have the airplane flying at a ground speed of 200 kts but never ask if the airplane is capable of that.
The tragic aspect is that the ATSB seems to be entirely sold on this statistical hocus pocus and that way will fail to find the airplane. What a waste of time and effort by respected scientists.
@VictorI:
Regarding the radar data, it is apparent that the Factual Information only shows DCA Civilian Radar Data from Kota Bharu. About the Malaysia Military Radar data the Factual Information writes:
So apparently the data available to the Australian Defence Science and Technology (DST) Group ended at 1802:59 UTC, plus a single point at 1822:12 UTC.
The “Military Radar Plot from Pilau Perak to Last Plot at 02:22H” shown to NOK at Lido Hotel, Beijing, Mar 21, 2014 does not exist officially. Apparently it is a military secret that should not have been shown, and someone is probably regretting that it was shown.
Regarding the odd-shaped turn past IGARI was have previously theorized that perhaps there is a gap in the radar data, during which a 240 degree right turn might have been accomplished. I have determined the distance travelled in the ‘dip’ in the ground speed plot in figure 4.2 of the “Bayesian” book and compared it to the distance travelled in the same time at approximately constant speed. The difference is 20 NMi, and at about 500 kt that would correspond to a gap of missing radar data of about 140 seconds. Could the statistical scientists have missed that?
Hi Paul, great question, I’ve added a graphic to show the how much of the newly defined search area has already been scanned. Yes, the IG area has already been searched.
@Gysbrecht: Even if we ignore for the moment the “secret” military image shown to the NOK, the FI contradicts the DTSG report relative to the military radar captures after 18:02 (which is referred to more precisely as 18:02:59 in the FI and 18:01:49 in the DTSG report). There is absolutely no justifiable explanation for this contradiction.
As for the wild speed fluctuations, when I was preparing my report on the radar data, I found that by deriving the speed from short segments of flight there were unrealistic swings in speed. Instead, I opted to find a steady Mach number that, when adjusted for temperature and winds, matched the overall flight path and timing. This produced unexplained “time offsets” for parts of the flight. In some ways, I resolved one problem (speed variations) by introducing another (time offsets).
The sharp turn after IGARI and the associated time anomalies remain unexplained. In my analysis, I introduced a time offset to match the measured time and position. You have suggested that the true path at the turn might not have been as depicted, and explained the time anomaly on that basis. The DTSG has introduced a sharp dip in speed at the turn, from about 500 kn to 200 kn and then up to 550 kn, while admitting that their estimated results are “not an accurate representation of the aircraft speed” due to a “mismatch between the assumed linear Kalman filter model and the high acceleration maneuver performed by the aircraft”. These “artifacts” are then ignored. Clearly, the DTSG was more interested in getting to a final answer than fundamentally questioning the data in its possession.
The net result is that the radar data has significant anomalies that remain unexplained.
Of course, we would know much more if the raw radar data set was released by Malaysia. For this very reason, it is unlikely to occur without a tremendous amount of public pressure.
Please join me in demanding that all the raw radar be released so a proper and transparent analysis can be performed.
@VictorI:
You already have an answer from Malaysia: wait until next March Interim Report.
I suspect the ATSB hasn’t got more data from the Malaysians than those used by the DTSG. Perhaps we could ask them to release a few points after IGARI? I would be glad to join you in a request to the ATSB, but my own mails to them never produced more than the polite obligatory standard response.
What plausible explanation might there be for no radar data being available to the Australians between 1802 to 1822? If it’s a military secret, what is the sensitivity that does not apply to 1802 or 1822?
@Gysbreght: Hoping that the raw radar data is released in the March report is not acceptable to me. Suppose it’s not there? Do wait another year for the next report? This is a classic case of stonewalling.
I agree that the ATSB and the DTSG probably have the same data sets. On this basis, there should be many requests to the ATSB to release the raw radar data.
@Bruce Lamon: The raw data between 18:02 and 18:22 might contain additional anomalies that make us question the integrity of all the radar data. Already, we have identified unexplained anomalies at the turn after IGARI. If there were national security issues at stake, Malaysia would not have shown the radar image to the NOK back in March 2014, essentially putting it in the public domain.
@Bruce Lamon:
RE: “What plausible explanation might there be …”
Perhaps not the same radar installation after 18:02 ?
@Gysbreght: Then why include the capture at 18:22, which was ignored in the analysis by the DTSB?
I remember that in the information chaos reigning in the first days after the disappearance of MH370 a high-ranking military officer had to deny that he had told a newspaper that military radar had recorded the diversion of the flight.
@Gysbreght: Yes, Air Force Chief Gen Rodzali Daud first said military radar captured the plane flying over the Malacca Strait. Then he denied ever making such a statement. Later, officials said that military radar captured an aircraft that might have been MH370. And finally, they conceded that MH370 was captured on radar in the Malacca Strait.
If there was primary radar data every 10 sec up until 18:02 as now claimed, it would be IMPOSSIBLE to misidentify the radar targets as MH370 in the aftermath of the disappearance.
@VictorI: Thanks for confirming my recollection. With that history in mind, I wouldn’t be surprised if the military authorities are now sitting on their data, refusing permission for publication of raw data. I wouldn’t rate that as a “national security issue”.
@Jean: funny you should ask about ATSB contacts, because just last night, I was e-mailed by Peter Foley himself. (I won’t call him a “contact”; my work has generally excoriated the agency he helps run, so I wouldn’t expect to be on his holiday card list.)
Background: Dr. Richard Cole has been painstakingly plotting and tweeting out search progress maps. His data suggested a small, thin stretch – roughly 20kms inside what he calls merely “the 7th Arc” – has in fact been missed. I asked the ATSB about this last week; they replied earlier this week to say nothing remained unscanned.
But last night, Peter Foley e-mailed me to say that this reply was incorrect. Once his ops team returned from Havila Harmony testing, they “rechecked”, and found that Richard was in fact correct: a small strip remains in their job jar.
@All: I don’t need to cross-reference the map. I’d bet my last dollar this lone missed strip is within the Dec. 1 “red zone”.
Here’s what my data suggests (search progress through to mid-Nov/’15, courtesy VesselTracker’s free trial license):
https://drive.google.com/file/d/0B-r3yuaF2p72Wm5aTmRpOE1OSFE/view?usp=sharing
(Powerpoint file, 19Mb; unfortunately, it seems you must download to view properly.)
The sliver has not been marked. I prefer viewers find it for themselves, to illustrate the odds of MH370 being in that patch.
If MH370 is found in that sliver, I hope mine is not the only spider sense that will be tingling…
(Apologies, Richard – it was Peter, not you, who referred to the 7th arc ambiguously – above was misworded. Your descriptions have always been impeccably precise.)
Here is a slightly more detailed map of the search progress with the new ‘Bayesian’ analysis added as a background. I have shown a +/-40nm box which matches the thick line in the new map. The outside boundary of the current bathymetric area matches the new search area.
A large area remains to be scanned inside the 7th arc.
http://www.recole.plus.com/MH370/4-12-15.jpg
Figure 4.3 in the “Book” can be compared with the Lido Hotel figure. In the Lido figure, note that there is a high density of points in the lower right portion of the aircraft track, but then, just before the track reaches Pulau Perak, the density of points drops dramatically. This point matches with the beginning of the yellow line in Figure 4.3. The yellow line is the “Book’s” extrapolation of the flight path past the last good position at 18:01:49. The white line in this figure plots the actual radar data up to 18:01:49 and then is some ill-defined representation of the radar data out to 18:22:12 (the last radar point.) If I squint properly, I can sort of make out that the curves in the white line of Figure 4.3 from before the start of the yellow line kind of match the curves in the points in the Lido figure.
Maybe someone can do some Photoshop magic and make a graphical overlay of the two figures.
If the Lido Hotel radar track after 1802 was shown by mistake, might it have reflected Thai radar from Phuket that Thailand confidentially shared with Malaysia? And for that reason Malaysia withheld it from Australia?
A few other weird to me excerpts from “Bayesian Methods”:
1. “The radar data contains
regular estimates of latitude, longitude and altitude at 10 second intervals from
16:42:27 to 18:01:49.”
And altitude? I was convinced back in the Duncan Days that military radar could not reliably estimate altitude, but if DST Group agrees, why mention it, and, if not, what are those altitude estimates?
2. “[T]he ground speed observed by the radar prior to 18:02 is relatively high and implies that the aircraft would have been at low altitude.”
I thought just the opposite was true, i.e., mostly because of the difference in air density, the higher the speed, the higher the implied altitude.
(This sounds like the crazy explanation given for abandoning the effort to retrieve surface debris in March 2014.)
3. “The 18:22 radar observation was not used quantitatively because the latitude and longitude derived from it are likely to be less accurate at long range and the aircraft may have manoeuvred prior to 18:22.”
If I read Fig. 2 of Victor’s 9/24/15 paper correctly, at 1822 MH370 was within the range of Phuket and Western Hill military radar exclusively. Victor deemed it unlikely that Kota Bharu military radar tracked MH370 after 1745. If Western Hill had tracked MH370 at 1802, it presumably could have continued to do so until 1822. This leads me to speculate that 1802 may instead have been tracked exclusively by Bukit Puteri–its range limit roughly coincides with MH370’s position at 1802 per figure 2 (and appears to cover the IGARI turn).
The point is that Bukit Puteri is roughly as far from MH370 at 1802 at Phuket is at 1822, yet the 1802 location is accepted while the 1822 position is rejected.
Overall comment: the radar data is a key basis for DST’s Bayesian priors and it seems ridiculous that they are on the verge of publishing a book on the subject with apparent critical gaps in their access to that data. More ridiculous still if their uninformed conclusions are guiding the searchers.
Talk about coitus interruptus.
Never thought we’d wait this long for a report that strays so far from the given logic & path of the IG.
Not a ditching? Where’s the debris!?
Just think of the buoyant materials. We should be awash by now. Only the Reunion flaperron? Why would it wash up & not ONE of the 478 shoes, nor seat cushions & other sustainable buoyant materials? KAL007, TWA800 & other flights that rendered their spoils of lost life. Makes no sense at all. Forensically, it’s like entering a crime scene where everyone died of gunshot wound’s with no blood, or ballistic evidence. No point of entry, or exit. No evidence… Only neighbors that heard things (the Maylay’s, but I did..I think) saw things (Maylay radar, but really didn,t..I think).
I’m not sure if the Maylay’s are covering for sheer incompetence or steering themselves away from the single largest lawsuit ever to be filed in the history of aviation. Someone will be held accountable for the lost lives without a single explanation!!
The sheer lack of debris points at one thing. Controlled landing. Why is the biggest question.
Bruce Lamon,
The Thai military radar data were released 10 days after the disappearence of MH370. By that time, there were plenty of figures available showing the last radar point, identified as coming from Malaysian military radar. It seems highly unlikely that the Lido plot was showing Thai data to the exclusion of Malaysian data.
Let me digress a bit. [The following is complete conjecture on my part.] Malaysia is a small country, population-wise, compared to its neighbors – Thailand, Vietnam, Philippines, and Indonesia. (I omit Singapore, as it is a special case.) As Andre Agassi famously said, “Image is Everything”. Malaysia wants it to be known that it should be treated on an equal footing with its neighbors. No way is it going to admit that Thai military radar is superior to its own. [End of conjecture.]
I could be wrong.
If end of flight was a ditching and someone who wrote a quote from the Boeing manuals a B777 – something like:
is designed to float indefinitely
then it should have been noticed by satellite imagery.
If end of flight was a crash then as @ChrisInDallas says
Where is the debris ?
Maybe it landed on land somewhere else ??
Thanks sk999. By “Thai military radar data,” are you referring to Air Vice Marshal Montol Suchookorn’s abbreviated description of what Thai military radar showed? (E.g., http://www.huffingtonpost.com/2014/03/18/thailand-malaysia-flight_n_4985619.html.) I’m not aware that Thailand has released any additional or more detailed military radar information.
I believe the impact of changing the reboot time from 3:40 to 2:00 has the following verifiable impacts on the SW limit of a proper search:
1) a handful of additional nautical miles’ worth of post-arc7 glide distance is now available (now 100 seconds more spiral glide available after crossing Arc7)
2) the assumed southwestern-most point from which MH370 could conceivably have sent the 7th arc signal moves EAST, by roughly QUADRUPLE that distance. Because (trigonometry):
For each fuel limit-generating path, we now have only 120 seconds “tacked on” to the end (between the 6th and 7th arcs) of each, instead of the former 220 seconds. If, accordingly, you shave x nautical miles off of each “speed=y” path, it must now take a more easterly bearing in order to intersect Arc7 at the same 00:19:29 time. I did this in Google Earth, and found that the performance limit moves East roughly 4nmi for every 1nmi of range transferred from pre- to post-Arc7 flight legs.
If I’m correct, then anyone trying to justify the new search zone out to 40°S by referring to the new reduced logon time has got it precisely backwards – if this gives MH370 10nmi more range outward from the intersection point, it moves the intersection point 40nmi east, resulting in a net 30nmi CONTRACTION of the indicated western search limit.
All feedback warmly welcomed.
And I ask again: has anyone spotted any indication whatsoever in its Dec/15 report that the ATSB has CHANGED its working fuel limit from that described in its Oct/14 report (which truncated its primary search zone at 38.3 degrees S latitude – citing its fuel limit explicitly)?
If the ATSB has been deliberately low-balling MH370’s range, it has been deliberately short-arming its search. We need to demand search leadership publish its fuel model in full detail. Those of use who are LEGITIMATELY trying to unravel this mystery have had quite enough of their obfuscation.
What is IG’s position ? Hypoxia ? Suicide ?
@Brock
Thanks for your efforts in obtaining the GEMS report. It certainly looks as though the notion of debris from the current search area first making landfall on the coast of Java is completely wrong and the chance of any debris going there at all is extremely small.
I searched for some more information on Indian Ocean currents and found that the situation is far more complex than it would first appear, including seasonal variation and different effects at surface and sub-surface. http://www-pord.ucsd.edu/~ltalley/sio210/Indian/ from UCSD provides some good information and references. See particularly the link to the graphic “Schematic of Indian Equatorial Currents (Tomczak and Godfrey)”. In the figure on the right (Southwest Monsoon season Sept-Oct) it seems remotely possible that debris from the current search area could start by travelling north and then west with the Southern Equatorial gyre, then northward up the coast of Africa and then eastward (with the clockwise current) towards Indonesia. I recall Prof. Pattiaratchi saying something to this effect, that it was possible given enough time for debris to cross the equator heading north up the African coast.
Overall though, the chance of just one piece of debris being recovered so far and in a reasonably intact state following the suggested crash (not ditch) after fuel exhaustion must be close to 0.
I looked at some limiting cases to better understand the results of the DTSG. The location on the 7th arc is primarily a function of the speed and the time the plane turned south. DTSG assumed that the speed was randomly distributed between M0.73 and M0.84, and matching the BFO bounds the randomly distributed turn south between 18:28 and 18:40. So, using their assumptions, the northern limit of the search zone should be determined by a turn at 18:40 and a flight at M0.73,and the southern limit be determined by a turn at 18:28 and a flight at M0.84.
I assumed FL340, using the appropriate meteorological conditions, I accepted the radar capture at 18:22, and used M0.83 until the turn south. I allowed turns at handshake times and matched the crossing of each ping arc. And other than using the BFO to determine the times for a turn south, I otherwise ignored the BFO, which is consistent with a BFO with a standard deviation of 25 Hz. This was consistent with the value used by the DTSG.
The result is the northern limit is -33.7,94.4 and the southern limit is -39.8,84.5. Since the PDFs used by the DTSB show no preference for any value in the allowed interval of speed and turn time, the PDF for the end point should be relatively flat between those two points.
Instead, these end points are in the “low probability” area of the search zone, as defined by the DTSG report.
This suggests to me that there is an inherent preference for straight flights in their supposedly random sampling of possible flights. That is the reason for the well-defined hot spot around 38 S latitude. It is no coincidence that it matches the predicted end point of the IG and others that assumed a (straight) flight on autopilot.
This bias for straight paths can be seen in Chap 7 of the DTSG report. Look at Figures 7.2 and 7.4, and the accompanying text in Section 7.3., in which paths are calculated for which the BTO and BFO are ignored completely. There is a hot spot for the end point corresponding to a straight flight starting at 18:02. This proves there is a preference for straight paths. The assumption is that the plane flew straight between maneuvers, so a flight with few maneuvers flies relatively straight.
The bottom line is that I believe the DTSG report suggests a much smaller hot spot because of the bias for straight paths. The complex math masks this bias.
Correction: I was projecting the impact of a reduction in pre-Arc7 range along the FUEL arc, when I should have been projecting it along the 7TH arc. Apologies.
Correcting this reduces the gross movement East in the intersection point from a factor of 4 (as I’d wrongly asserted above) to a factor of only 1.7 or so. For example, a 10nmi shift in range from pre- to post-Arc7 shifts the intersection point 17nmi eastward, for a net move east in the indicated search area of 7nmi.
No change in the basic principle: that the ATSB’s fuel limit of record has been breached by their own search, and the only thing their Dec.1 “clarifying” report gives us merely WIDENS the extent of this breach.
VictorI,
Minor point of clarification – the DTSG does not use a value of 25 hz for the BFO error, it uses 7 hz (p. 29). 25hz is used as a “prior” on the distribution function for the BFO bias (p. 28; 47). The way it is used is subtle – it’s just an initial guess at the range of BFO biases to search; the actual range for all the simulations that end up in the search is undoubtedly much smaller.
@sk999: Thank you for that note. My interpretation is that the 25 Hz refers to the drift and the 7 Hz refers to the noise. Over a short period, the BFO data would have a variation of 7 Hz about its mean. Over longer periods, the mean bias would also drift.
To show that the BFO has little effect on the PDF, look at Fig 10.3, which compares the PDF with and without including the BFO. Eyeballing it, the range of end point latitudes is about -35 or -41 for no BFO and -34 to -40 when the BFO is included. And it is acknowledged that the main effect of the BFO is to limit the window for turn times to the south. Note that the range with the BFO matches my estimate of -33.7 to -39.8 that I calculated using a much simpler methodology, although I would estimate the PDF to be flatter in that range than they predict.
I maintain that the peaking in their result around a latitude of -38 is due to the preference for paths with fewer turns. This is seen in Figure 10.4, in which the best estimate for the number of turns is 1, i.e., a single turn to the south.
Does anyone see a periodicity in the BFO error shown in Figure 5.4 of the Bayesian_Methods paper, that is of similar magnitude as the Schuler period of 84.4 minutes?
https://en.wikipedia.org/wiki/Schuler_tuning
Hi, reviewing the report from my perspective, few things/questions to emphasize:
“This model was calibrated against flight paths of commercial aircraft.”
– since U-turn, aircraft used more military behavior
“Power interruption to the Satellite Data Unit (SDU)”
“• the cycling of the left generator and backup generator switches with the bus tie isolated (all
switches are located on the overhead panel in the cockpit)”
– overhead panel, so no magic or EE-Bay?
“18:27:03 In-Flight Entertainment system (IFE) sets up a ground connection via SATCOM for a SMS/email application.”
– not possible for passengers anyway on this plane??, but not used (nor SATVOICE) by pilots in danger too
“the aircraft may have exceeded its design envelope when it was descending and experiencing phugoid oscillations”
– or controlled flight too
“The ATSB performed a basic trajectory analysis of an uncontrolled, but stable aircraft.”
– but not controlled
“The sequence considered in the end-of-flight section above favours a no active control scenario.”
– vs cycling on overhead panel
“Entering zero for the cost index results in maximum range airspeed and minimum trip fuel. This speed schedule ignores the cost of time.”
– 7hrs instead of 5hrs ??
“Applying the assumption that a series of stepclimbs[8] had been performed during cruise, produced a range greater than that required to reach the region of interest on the arc.”
“8 Step-climbs require changes to the auto-flight system at the time the climb is initiated”
– !!!
“Ditching considerations – impossible”
– Almost no debris, except flaperon media star – crash highly unlikely ??
– as we are currently in some kind of war, almost everything is possible, except things prohibited by international law against humanity ??
@DennisW said to @jeffwise: “I wrestled with this post, but could resist no longer. Your comment about putting the controlled ditch to rest based on the recent ATSB reports is about the dumbest thing you have ever said. You set a new bar on that one.”
It is true that an aircraft can land without engine thrust, in contradiction to the statement in the ATSB report. In the case of MH370, I think the real question is whether the aircraft can land with only the RAT supplying the electrical and hydraulic power.
As a plane slows for landing, the power generated by the RAT diminishes. In the US Air 1549 landing on the Hudson River, pilot Sullenberger was gliding with both the RAT and APU operating as the plane had plenty of fuel.
Based on the comments of the ATSB, it appears they believe that the RAT alone cannot provide sufficient hydraulic power to adequately control the plane at the slow speed required for a successful ditching.
Perhaps the conclusion of the ATSB, which is in consultation with Boeing, is correct.
@VictorI:
With only the RAT supplying electrical and hydraulic power, the flaps cannot be extended. Therefore the airplane cannot slow down very much without stalling and losing control. I guess that the comment about ditching is from the DTSG, rather than Boeing.
@VictorI:
See for example https://en.wikipedia.org/wiki/Gimli_Glider