[Note: the following is a translation from an article published by Spiegel Online on Friday. Thanks to @littlefoot (Sabine Lechtenfeld) for her assistance.]
by Christoph Seidler
What happened to the missing Boeing 777 of flight MH370? The aircraft with 239 people on board has been missing since March 8, 2014. As the end of July, a piece of wreckage washed up on the French island of La vunion, giving investigators that the wing flap might help solve the riddle–among other things, because it could show how the piece of debris made its way across the Indian Ocean. And that, in turn, should provide a clue to at least the general area that the plane went down.
Now computer calculations by German oceanographer suggest that perhaps the search for the Malaysia Airlines aircraft has been carried out in the wrong area. Until now, says Andreas Villwock of the GEOMAR Helmholtz Centre for Ocean Research in Kiel, it’s been assumed that MH370 had crashed far south of the equator at 35 degrees latitude. Computer models by his colleagues show, however, that the debris “probably comes from the eastern equatorial Indian Ocean,” far from the nothern end of the current search area.
Over the course of several weeks, the Kiel researchers Arne Biastoch and Jonathan Durgadoo used a computer model to reverse the track of the debris across the Indian Ocean. Their model used daily current data that had been obtained in past months by French colleagues. The key question: Which path did the piece of debris take across the Indian Ocean, which scientist liken to the inside a pinball machine because of its chaotic eddies and turbulence?
Researchers understood from the start that a computer simulation can’t calculate a precise location of the crash site, but at best only point to a broad area. This coming Tuesday, they will present their results in detail at a press conference in Kiel. However, this much is already clear: The reseachers’ results are contrary to the Australian search strategy.
The current seabed search is focused on an area of ocean west of the Australian coast, and not in the equatorial region, which the simulations now seem to point to.
A total of 120,000 square kilometers will be scanned, an area as big as Bavaria, Baden-Württemberg and Hessen put together. Two ships of the geotechnical company Fugro are carrying out the task on behalf of the governments of Australia, Malaysia and China. So far, they have examined around 55,000 square kilometers with sonar, without tangible result. The simulation seems to suggest that large parts of the southern region were scoured in vain.
Meanwhile, the wing flap that washed ashore has been examined in a laboratory in Toulouse, France. According to Malaysian reports, the wreckage definitely belongs to MH370. French investigators speak of a “very strong suggestion.” details the wreckage belongs definitely MH370. French investigators spoke of a “very strong presumption”. Tha’s why Australia expresses confidence that the wreck will be found.
Two German scientists believe that so-called goose barnacles could provide clues to the crash location. These colonized the piece of debris as it drifted across the Indian Ocean. Geologist Hans-Georg Herbig and the biologist Philip Schiffer, both from Cologne, have identified the small crustaceans in photos.
Both are experts in the small animals, which are part of the barnacle family. Herbig and Schiffer have compiled the first genetic fingerprints for five barnacle species from different parts of the world’s oceans. They have also found that each subspecies dwells in specific climatic zones dependent on the latitude.
But because the French authorities have provided them no sample material of organisms from the wreckage, they can’t apply their insights to help narrow the search for the crash site. “I’m pretty frustrated because despite making requests through various channels, no one has replied,” Hans-Georg Herbig told SPIEGEL ONLINE. “We have tried many things, but got no response.”
In summary: Researchers from GEOMAR Helmholtz Centre for Ocean Research in Kiel will officially present their findings next Tuesday–however they have already give a first clue: Computer analysis suggests that a a piece of debris from flight MH370 that washed up on La Réunion orginated from the “eastern equatorial Indian Ocean.” This is far from the area where two ships are currently searching for the wreckage of the Malaysia Airlines aircraft.
We’ll have to wait for the details of GEOMAR’s drift analysis, but if it is based on a reverse-drift, deterministic methodology, it is inherently flawed.
ongoing disaster
CNN says, spanish manufacturer could not match with certainty the flaperon with MH370.
Something crazy with the industry?
So CNN is now reporting that the Spanish subcontractor to Boeing that supplied the flaperon is back from vacation, and the company reports that the numbers found on the flaperon part cannot be used to definitively link the flaperon to 9M-MRO.
If Boeing, the NTSB, or the BEA truly believed the subcontractor would be able to trace the flaperon to 9M-MRO, would it have allowed vacation to impede that determination? In my mind, it was always a specious excuse.
So the story we are fed is that we have a part that was found without its ID plate, has a numbered internal part that is useless for aircraft traceability, and was maintained in a way that is not exactly consistent with MAS maintenance records.
The facts continue to suggest that the investigators have found reasons to delay reporting the results of the investigation to date and have deliberately planted a seed of doubt as to the provenance of the flaperon. We can only guess why.
In the meantime, the press continues to blindly report the drivel of news it is being fed without any real analysis or questioning. This environment allows speculation to run rampant.
We should all be demanding that the facts of the case be disclosed. After one month of investigation, surely much more is known about the flaperon than has been released.
http://www.cnn.com/2015/08/31/world/mh370-investigation/index.html
@VictorI:
The article says they are targeting a broad area. At the same time, I looked at their publications, and none of the three has either educational background or recent publications in the area or probability/Bayesian methods.
However, recent research made by Metron Corp. shows the maximum likelihood of the crash should be to the north of current search area, between S22 and S27, provided uniform prior distribution. That is pretty much consistent with the details released in Der Spiegel so far.
@Alex: First, Metron acknowledges that its calculations were only to demonstrate the capabilities of using a Bayesian approach to help guide the search effort, and the specific results should not be used for a number of reasons, including the simplified drift model. They are trying to get hired to be part of the search effort, as they were for AF447.
Second, S22 to S27 is not consistent with “eastern equatorial Indian Ocean”. The GEOMAR results seem to be quite different than what Metron has calculated.
Again, we’ll know more tomorrow.
@VictorI:
First, Metron and German scientists are in the same position – they both do not know prior distribution of probabilities that Australian government uses so far, they could only guess. From Metron’s perspective, uniform prior within wide search area would mean S22 to S27. Bivariate alpha distribution assumed in the work would mean much smaller extension to S30 from current area.
Regarding the second question, I do not think there is a clearly defined term “equatorial Indian Ocean”, it is subject to interpretations. But we shall see tomorrow.
@Victor
Yes, pretty much as soon as I finished reading the Metron report, I was of the opinion that Dr. Stone was fishing for business. At the end of the day, life marches on, and Dr. Stone has to worry about Mrs. Stone and all the little Stones.
Of course, the results of flaperon being found are going to be used together with previous results used by Australian government, and what was used before was constrained autopilot dynamics and data error optimization. Based on those two, Australian government assumed certain prior distribution. That is now subject to be converted to posterior distribution. Of course, the results for posterior distribution are going to depend very much on the prior distribution. And nobody except the Australian government know the prior so far.
If the research made by German oceanographers does not take constrained autopilot dynamics and data error optimization into account and bases the results on flaperon being found only, their research would not be useful
@Alex: You did not read what I wrote. Metron has advised that their results were only to demonstrate capabilities, not to produce useful results. The underlying drift model from Erik van Sebille that they used was never meant to be used in this way. I would not put too much weight on what they report as results because even they do not.
@DennisW: I don’t fault Metron for fishing for business, because as you say, we all have mouths to feed. But we have to be careful to distinguish between a sales pitch and technical conclusions that are meant to guide the search. Metron will only produce those kind of results if they are paid to do so, and with additional expertise from those that understand the satellite data.
@VictorI:
I did read it. Yes, that is true. However, among other things Metron stated there is a summary saying that finding flaperon is a major event and the search area should be changed to reflect that new evidence. Of course, without numerical results it would be extremely difficult to redefine search area.
@Victor
We are on the same page. However, Stone et. al. will be able to wade through the ISAT data without breaking much of a sweat.
My guess is that they would apply convex optimization algorithms to it such as the ellipsoid. I have been too lazy to even go there. I used to have people for that.
I don’t expect much to change tomorrow. It will be a reverse drift model. I am pretty sure of that. The refusal of the French to supply barnacle samples or even to acknowledge the request is telling.
It keeps getting worse instead of better. No apparent convergence whatever. I have absolutely no idea how to proceed at this point. We have all been relegated to the role of observers it would seem.
@alex: I believe precisely the opposite: if an unconstrained reverse drift probability distribution is wholly incompatible with the CAD/DEO probability distribution (immaterial region of intersection), then either the ISAT data(‘s interpretation) or the Réunion flaperon are shown to be bogus.
As I’ve said before: a Bayesian compromise is only appropriate to the extent the probability distributions overlap.
@Dennis, why should the French investigators supply a German paleonthologist with barnacle samples? I can understand the interest of Prof. Herbig since he has apparently worked on the genetics of the barnacles and found out that the different sub-species aren’t really globetrotters. They prefer to dwell within in the limits of certain areas. Thus, it would be VERY interesting to learn about the genetic make-up of the flap-dwelling barnacles. But this isn’t just any other scientific project. It’s also a criminal investigation. The French are probably not even authorized to give out samples to scholars.
Let’s hope, though, they’re looking into the matter anyway. If the barnacles don’t like to stray from their natural home all that much the flap population could tell really interesting stories.
@littlefoot
Certainly the French are under no obligation to share the barnacles. There seemed to be a lot of them, so I reasoned that there was little harm in getting additional qualified opinions. I can understand if the French don’t want third party input to potentially contaminate their investigation. So it goes.
@Brock
Well said, however, people will believe that even minute overlap supports the current search area. We have already seen this noise coming from the ATSB.
@alex
“eastern equatorial Indian Ocean” surely would include West of Sumatra…funny, the ATSB mentioned that coast ages ago 😉 but then it was pointed out that debris could not drift from the current search area that far North; however, it may be the case that if the flaperon separated from the wing N of the Equator it could have drifted S and then SW to Réunion.
In my admittedly frequentist opinion (IMAFO) no reverse drift model will be informative with only 1 data item; several identifiable debris items arriving at Réunion, Mauritius, Madagascar or E coast of Africa et cetera within a few months of each other might be a useful basis for such models. On top of that, as has been said many times by others, taking the current search area as the prior is unrealistic given the very few data we have, their somewhat doubtful provenance and the assumptions that needed to be made about the flight (albeit by expert and very well-intentioned analysts).
@Brock:
CAD area does not go to the north of S35. However, even the initial search area (60000 km^2) extends beyond S35 to about S32.5 further to the north. That means they do not consider CAD as a strong indication of the limit being reached. Provided that, further change to the north (of S32.5) could be done as well.
@AM2: If they say equatorial, it does not have to be so far to the north as west of sumatra. It could still be much to the south and still equatorial. Let’s wait until tomorrow
@alex. Agreed, we should wait but its already Tuesday here.
@AM2, I can assure you it’s not Tuesday in Germany just yet. No reason to suspect delaying tactics and obfuscation in this case.
🙂
@littlefoot
Your post reminded me of a card I got from a “friend” on my last birthday. It shows three old codgers sitting on a bench with leaves blowing by, and trees at an angle.
The first codger say “It’s windy.”
The second codger says “No. It’s Thursday.”
The last codger says “I’m thirsty too. Let’s get a beer.”
Reflects accurately on this investigation as well.
@Dennis, let’s get a beer sounds like a good idea (‘goes to the fridge’) It’s awfully hot here even at 8pm.
I prefer Weizenbier wth a lemon – with alcohol, please!
@Brock:
DEA area also has a northern limit – it goes to the north only to S30. I guess that is the reason Metron considers uniform prior and calculates posterior only to S30
For more information on Metron, visit the blog on the right side of the following post:
http://tinyurl.com/o8qwp3r
@littlefoot & Dennis
A Shandy perhaps but prefer wine. Coffee right now though 🙂
Perhaps the flaperon is an ‘unapproved’ third party component that fell off a barge or something. There is a large concern about unapproved parts in the aviation industry although I suppose it more likely to happen with higher frequency replacement parts such as seals, pumps, electric motors, valves …..
The smell is getting worse. When what was ostensibly a piece of aeroplane showed up a month ago my old sparring partner Spencer said – “I’m calling it, that’s Mh370” – and we all waited for the inevitable confirmation that was going to take 24 hours. Since then:
(1)We have no ID plate when the adhesives are rated for worse.
(2)No other apparent means to pin the flaperon to MH370?
(3)A heap of marine biologists on a pretty short lead.
(4)A flaperon that doesn’t scream out 16 months of exposure.
(5)Conspicuous media management and or misinformation with no end in sight.
It falls into French hands in a French territory that happens to be a tourist destination at the commencement of the “vacation” period? Not everyone sounds to be a fan of barnacle forensics but it could be all we have left – so where are those little buggers right at this moment? Anyone know?
And while I’m still in conspiracy mode:
(6)Maintenance records that don’t match.
What I’d like to see is this:
Get another aircraft to take off from KU when 3-F1 is at the same position it was when 9M-MRO lifted off – fly the course the experts believed it did – and see if they come up with the same data.
It would need larger fuel tanks though, otherwise we will definitely then have one to recover in the SIO!!!
A similar thing was done here in New Zealand a few years ago when a helicopter disappeared after dropping off radar. After a huge expensive search was called off, someone tried flying another helicopter to the same point guided to that point by a radar operator and found there was an error in the radar. When they got to that radar position, they found the helicopter!
@Matty: I would describe you as being in “tell it like it is” mode (as usual). Well done.
(7) Official drift analyses that were way wrong, then suppressed, then retroactively-perfect-yet-wilting-under-scrutiny.
@Gavin
I believe Inmarsat has performed benchmark tests with multiple aircraft flying existing routes, but I don’t think any of these routes took to the SIO.
….
@JeffWise
Will the 2 French families be given any inside information by the French president that could be leaked out to press? I think not.
But will the French govt. give assurance to the families that the Malaysian govt. version of events is correct so far? I doubt that too.
@everybody
What if the French govt. and the agencies involved in the search are playing up the mystery of the flaperon to create doubt in the minds of the perpetrators, and their sponsors?
Does this look like GAME THEORY in action?
@Matty @all
for a break, and being in “conspiracy” mode too … May 2014 WHCA dinner (2015 great too)
https://www.youtube.com/watch?v=fJOHRYhl67M
(is something funny there, these days?)
taking stock
Many people here seem to be frustrated with the lack of results in regard of the flaperon.
But is this part really so crucial to the independent investigation? It does make a sad difference for the Next of Kin, but would it change, what we have to look for?
Many people might think, flaperon from MH370 – game over – plane crashed in nowhere without rationale (besides a very cruel assumption about AP-mode). This thinking is quite wrong, because it does not take the crucial questions in our independent investigation away from us. Even if we think it is somewhere at 34S 94E which is reasonably disputed yet, the decision for us would be nonetheles, whether it was
A) accident, rare mechanical failure or
B) deliberate,
and if deliberate if it was
(a) pilot based accident reaction,
(b) pilot based nefarious,
(c) non-pilot based nefarious act,
(d) outright sabotage
After all technical discussions we have to admit that there are some levels of criminal investigations that need to be addressed and are not the usual daily diet of tech folks. If we speak about a nefarious act or sabotage – and i think most of us do – we might revisit the question of outside support for the deviation. There i a pattern behind everything concerning this flight, and that is the impression of a power behind the scenes, that is directing things. I have no idea what kind of power that might be, but i think of mafia syndicates like in mexico or elsewhere, who have infiltrated administrations on the highest level and do their activities from within.
Now the first sign of such pattern was the delay of SAR alarm by HCM ATC due to outright false information from MAS HQ. This supported willingly or unwilingly a deviation that could not be stopped by jets from Butterworth, because their reaction time window became less than 10 minutes. The deviated plane also run away as fastg as possible and copied the way american military aircraft would fly along FIR borders if they didnt want to be detected. After the radar signal abruptly ended, the png scenario started incidentally at the same time . What people often forget here, all spoof possibilities discussed do apply to southern scenarios as well.
The point is, there is this pattern of buying time, of delaying reactions and investigations, of looking into wrtong places, following false leads etc.etc., from the very beginning of the event thru to the finding of the flaperon.
So a criminal investigation would have to keep in mind this posible external assistance and then lok for organizations who are capable to execute this.
into question? It’s obviously wrong.
geomar press conference
there will be live coverage from the press conference by the NDR.de (german public tv for northern region) at 12:00 a clock CED (german time) GMT+2
Christmas Island
Seems to be empiric reverse drift model with empiric data from mercator on a daily basis
According to Geomar: The highest probability area covering the 7th arc might be , where the arc is closest to Christmas Island
http://www.geomar.de/index.php?id=4&no_cache=1&tx_ttnews%5btt_news%5d=3972&tx_ttnews%5bbackPid%5d=185&L=1
Live now here: http://www.ndr.de/nachrichten/schleswig-holstein/Geomar-Pressekonferenz-Neues-zum-Flugzeugabsturz-,geomar188.html
http://www.ndr.de/nachrichten/schleswig-holstein/Geomar-Pressekonferenz-Neues-zum-Flugzeugabsturz-,geomar188.html
geomar
movies and pictures beneath the german language article in this link
http://www.geomar.de/news/article/wo-ist-mh370/
For those that prefer English:
http://www.geomar.de/index.php?id=4&no_cache=1&tx_ttnews%5btt_news%5d=3972&tx_ttnews%5bbackPid%5d=185&L=1
Is the intent to prove that the La Reunion flaperon did not belong to MH370?
Warning: A completely unsubstantiated rumor follows.
From poster SLFGuy on pprune.org:
From a source in Paris not 10 minutes ago, (source is in media not airline industry so take this as you wish).. the next 72 hours will see a ‘showstopper’ announcement regards the flaperon.
http://www.pprune.org/rumours-news/565335-flaperon-washes-up-reunion-island-34.html#post9101795
Along Inmarsat’s 7th arc, Geomar’s probability distribution peters out at roughly 19 degrees south latitude.
Geomar applies all the right caveats (most notably, that one piece of debris is insufficient to determine source with any certainty). But still: forget “interpretation” – Geomar’s probability distribution appears to be incompatible with the Inmarsat data ITSELF.
Just imagine if this thing is concluded not to be from MH370. Would almost be as bad as the 37.5 hz frequency pinger debacle.
The probability map from GEOMAR is certainly interesting. http://www.geomar.de/fileadmin/content/service/presse/Pressemitteilungen/2015/map_mh370_particles_la_reunion_de.pdf
QBrock
Exactly: They won’t criticize ATSB, but what they want to say is, either the flaperon is true, then the INMARSAT data are false (flasely interpreted or mutilated somehow), or vice versa.
Its not really Geomar alone but a french (Mercator)/german(geomar) initiative. I have a strong assumption that it is not just a coincidence that Mercator sits in Toulouse , where the investigation of the flaperon takes place , and that the german french initiative to develop this drift model started a month ago at once with the advent of the flaperon. It seems to me, that Germany and France function as a kind of neutral power, who are not involved in asian and pacific politics, and that both dont give much on the INMARSAT data.
Given what we’ve deduced about the flaperon’s expected profile whole drifting, what is the appropriate percentage of wind speed we should add to a drift algorithm?
I’ll open the bidding at 0.0%, to two sig figs (90% submerged, nothing for wind to catch).
David Griffin (CSIRO) confirmed to me via e-mail that adding waves to their model had negligible impact.
@Brock
That was said so by the scientists on the press conference. wind with zero and waves negligible. But latter will be added in next week for refinement.
@Brock/CA
The statement that either the ISAT or the Geomar data must be wrong is simply not true. It is possible to construct a path to the CI region that is compatible with the ISAT data. In fact, there are many such paths.
The ISAT data in and of itself is not deterministic. What places the plane in the current search area is NOT the ISAT data, but the ISAT coupled with an assumed AP flight mode. I have been trying to get this point across on this blog for a very long time. I don’t know how else to say it. There is no reason to be suspicious of the ISAT data, and the Geomar results do not contradict it.
I said earlier that I will not evangelize the CI theory, but will only defend it. I will stand by that statement. I have published several links to the analytics and supporting information.