I am grateful to reader @AM2, who early this morning alerted us to a report in the French website LaDepeche.fr stating that investigors who have been examining the flaperon found on Reunion have been unable to find any evidence linking it to MH370. Soon after, reader @Jay provided the translation below, which I’ve tweaked and edited using my high-school French and some online dictionaries. Thanks to both of you (and to Brock for his translation help)! Any corrections or suggestions from people who actually know the language would be very gratefully received.
MH370: At Balma, the Technical Investigation is Complete
The Toulouse experts of the Directorate General of Armaments have finished the survey of the flaperon found on Reunion. Nothing permits it to be 100% certified as belonging to MH370!
In Balma, near Toulouse, technical analysis of of the wing flaperon believed to belong to the Malaysia Airlines Boeing has ended. The Toulouse engineers have submitted their findings to the Paris Prosecutor’s Office, which is in charge of the judicial inquiry. At the moment none of their observations have been leaked. “The investigation team headed by the French to consider the flaperon concluded the first phase of its inspection work,” the Australian Transport Safety Bureau (ATSB) announced in Sydney.
Circumstantial evidence
“French authorities will, in consultation with Malaysia, report on progress in due course,” added the ATSB. Indeed, the judicial authorities remain silent and refuse to comment. According to our information, the experts have found no compelling technical element that would certify 100% that this piece belongs to flight MH370. “The expert conclusions are only the technical part of the criminal investigation, which is still going on,” so the case cannot be considered closed. For now all that is certain is that the flaperon, which was transferred from the island of Reunion to Toulouse on August 5, corresponds to a moving part of a wing of the Boeing 777. A representative of the American manufacturer Boeing quickly confirmed that after arriving at the site of the DGA Aeronautical Technical Center in Balma. If the deputy prosecutor of the Republic of Paris has stated that there was a “very strong supposition” that the piece belonged to the plane of flight MH370, which disappeared 18 months ago, that is based on circumstantial evidence.
First, the piece belongs to the aircraft model corresponding to that of Malaysia Airlines, a Boeing 777. In addition, no other aircraft of this type except that of the Malaysian company were reported missing.
Also, the trajectory of the wing piece that ran aground on a beach in Reunion matches the sea currents that link the search area of the wreckage of the plane to the French overseas department. Finally, the shells found attached to the flaperon belong to a species endemic to the southern Indian Ocean where the unit is believed to have disappeared.
According to a Toulouse aeronautics expert who requested anonymity, the element of the wing would not have floated for several months at the water’s surface but would have drifted underwater a few meters deep. According to Jean-Paul Troadec, former chairman of the Bureau of Investigation and Analysis (BEA), the state of flaperon, even if it is not intact, indicates that there was no violent impact with the ocean surface. “If this had been the case with the MH370, one would expect much smaller debris than a flaperon,” said the expert.
COMMENT
A couple of observations from me, JW:
- I find it odd that a piece of random debris would happen to have exactly neutral buoyancy, as floating for months just below the ocean surface would require. Unless it was tethered…
- Reader @Jay raises the question: “What about the maintenance seal that Malaysia claimed 100% linked the part to MH370?” Likewise, no mention is made of the discrepencies that Boeing and NTSB officials reportedly found between the flaperon and Malaysia Airlines maintenance records, according to the New York Times. Hopefully the French will soon issue a report clearing up these issues.
@Gavin
Your linked reference makes no mention of the FAA. I am pretty adept at Googling, and have found nothing attributable directly to the FAA, Boeing, or the NTSB. In the case of Boeing and the NTSB there are vague references by some unnamed person “close to the investigation”, but nothing directly from any of these sources.
Relative to your “absence of alternatives” manifesto do you believe the following?
1> The US moon landing was faked, and filmed in a Hollywood studio.
2> Princess Diana crash was not an accident, but planned by the Royal Family who did not like her?
With all this discussion of Drift Modelling, can anyone help me out here?
One thing that really ‘bugs’ me is that how is this drift modelling done?
Is it based on hard evidence like dropping satellite floaters into the ocean and tracking them, or is it simply done by computer modelling?
As you know, a computer program is only as good as the data fed into them – garbage in – garbage out!
Unfortunately, I don’t think there is a method of posting images to this post, so I will just have to attempt to explain what I mean….
If you have a look on Google Earth, note the area from above Reunion/Mauritius Islands up to the Maldives. Notice what the sea floor looks like – ridges running from North to South-West.
Where did Google get the data to plot this sea floor from?.. I presume from known plotted sea charts.
In my understanding ridges like these are carved out by years and years of strong water action, just like how canyons such as the Grand Canyon were carved out.
So please forgive me if I’m wrong, but wouldn’t this be more of an indication of where debris floating on the surface is likely to have come from?
These ridges must surely have some influence on the currents above?
@Gavin: As time goes on, I increasingly believe we are being manipulated by one more parties. We are so starved for information that any crumb that comes our way is taken very seriously.
What we need are clear, on-the-record statements from the official investigators. Instead we have official statements that are contradictory and/or ambiguous, or we have unverifiable leaks from anonymous sources claiming to represent the officials. This pattern of information disclosure suits the needs of whatever party wishes to spin the story in a particular direction.
So we have Malaysia stating unequivocally that the flaperon is from MH370, quickly followed by the French criminal judge contradicting this by declaring that the flaperon was probably (but not definitely) from MH370, followed by a leak from the NTSB/Boeing claiming the MAS maintenance records don’t exactly match the part, followed by an anonymous French expert making the incredulous claim that the part drifted completely submerged from the SIO to La Reunion.
These are just some recent examples in a long pattern of leaks and contradictory statements. A leak that the aircraft climbed and descended rapidly, but this was later dismissed as anomalous radar readings. A leak that a cell phone connected at Penang, but this was later dismissed as inaccurate. A leak that flights to the SIO were found on Shah’s computer simulator, but later this is dismissed as false, only to have the story resurface in recent weeks.
I’ve heard the claim that this confusion is because under ICAO protocol, Malaysia heads the investigation, and therefore controls the flow of information. My observation is that countries observe the ICAO protocol when it suits them. For instance, in the days after the disappearance, the US made official statements that put the plane in the Indian Ocean rather than in the South China Sea. But now, after the NTSB and Boeing have examined the flaperon, all we have is a cryptic statement allegedly attributed to them.
So now we are all waiting with bated breath for French investigators to release their findings. Don’t be surprised if the statement is less then unequivocal and parties use the opportunity to leak information in whatever direction suits them.
I am optimistically hoping I am wrong and we actually learn something soon.
@ALSM
Thanks for the drift model link from Metron.
Just to be clear this model differs from Henrik’s model, which I believe is stronger. The Metron model does not consider where debris has not been found. The Metron model makes an assumption relative to the ATSB probability distribution of locations based on the ISAT data, and then links in the drift model of Sebille. The result is to draw the “combined” probability distribution to the North along the 7th arc.
Henrik’s model does not rely on an ATSB probability distribution at all.
@Ron
Sorry Ron, I’m not quite sure where you are coming from re airspeeds, but I presume it was from my website.
Yes, the “Indicated Air Speed” (IAS) is 330 knots, but this is not the True Air Speed (TAS). A good converter can be found on this link: http://www.csgnetwork.com/tasinfocalc.html
To explain the difference, I like to think of it as: at 330 knots, the aircraft is pushing through the air at a rate of so many trillions of molecules of air per second. As the altitude increases, there is less of the molecules, so to achieve the same amount of molecules the aircraft has to travel over the ground much faster – and so the airspeed indicator will still show that it is doing 330 knots whereas in reality it is much faster. The converter (on the site I refer to above) gives a TAS at 35000 ft of 561 KIAS with a 330 KIAS. This is why aircraft fly at such great heights as they can travel much further on the same amount of fuel… and get there faster.
@DennisW
I’m very sorry, but I got my wires a little crossed. What I should have said obviously “Boeing”, but in checking it out further I now see in this article that it is also the NTSB:
http://mobile.nytimes.com/2015/08/06/world/asia/mh370-wing-reunion.html?referrer=
…and as to your question of 1 and 2… I have no opinion. At the end of the day it has no bearing on this subject.
@Victorl
Very well put! This whole thing has become very maddening and I think everyone just wants to see an end to it… I know I do!
@Gavin
Yes, Boing and the NTBS were mentioned, but no statements from either party have actually been made.
Relative to my questions, they were not selected at random, but were actually taken from an article published by Lewandowski in “Psychological Science” that has become an oft-cited work relative to what motivates and drives conspiracy theorists.
Jeff & Victor,
Yes the verbiage submerged is correct, as I used in my description in my post to airlandseaman. I’ve had 8+ years of French. I don’t believe Brock is entirely correct, it did not “bob” between waves, it was more blanketed or flanked, if you will, by the water, under the water.
I posted this hours earlier but it may be in a queue somewhere or did not post so excuse if a duplicate.
@Victori
It’s gut wrenching reading your description of only a minut amount of the bs you (and others) sift through to gleen pure factual information. All your efforts (and others) will eventually force this story too unfold so those that seek peace shall have it
They want more debris,just to make sure plane is in the ocean and not on land. Maybe someone is throwing the pieces inside?that piece looks like it hasnt been in water for long,by looking at the images.
@Michael Molinaro,
Thank you for your interest in my work on MH370. I have been diverted for a number of months due to family medical issues and consulting work.
Soon I hope to address several MH370 questions, including the following:
(1) Are the satellite image features (resembling but perhaps not actually typical ice contrails/distrails) real atmospheric disturbances?
(2) Could they have been produced by a high-altitude airliner?
(3) If the answers to (1) and (2) are affirmative, what do the features tell us about MH370’s route?
(4) What are the limits of possible routes between 18:22 and 18:40 UTC that are consistent with all the BTO and BFO data?
(5) Do the answers to (3) further constrain the answers to (4).
(6) Does the zone of 18:22-18:40 routes found in (5) also constrain the southerly routes that are consistent with post-18:40 BTO & BFO and the fuel remaining?
(7) What zone on the 7th arc is consistent with (6)?
In addition, I hope then to address the available acoustic/seismic data in greater detail. In principle, these data can provide a useful and precise method of identifying a small search zone on or near the 7th arc. The acoustic data are quite sensitive, and that is both a blessing and a curse. The difficulty here is not in detecting the delayed arrival of the impact noise of a B777 aircraft in the ocean. The hard part is to discern which of the many detected noise events should be associated with the aircraft impact event. There are small but important uncertainties in the SOFAR acoustic propagation speed from Point A to Point B in the SIO. It varies with location (primarily latitude) and with season. The resulting uncertainty in arrival time at a sensing station is on the order of several tens of seconds. So the problem then becomes that of finding a hypothesized event location and time that produces detectable noise events at multiple (preferably >10) stations at their calculated arrival times (within a window ~1 minute or so wide). For acoustic sensor arrays, the signal processing is complicated by the need to compute the relative arrival times at each sensor element (or to compute an arrival direction corresponding to the hypothesized impact location). For seismic (3-D) sensors, one also needs to synthesize the equivalent measured signal if arriving from the direction of the impact location. The acoustic and seismic data files can be quite large. The net result is that it is a significant mathematical exercise to probe the 3-D impact space (lat/lon/time) based on finding many (preferably all) SIO acoustic/seismic stations with events in the predicted time window. However, insofar as I know, this is the only method (besides the Inmarsat BTO/BFO data) that offers the possibility of providing an independent estimate of impact location. In my opinion, the work on this has not nearly been exhaustive. Kirill Prostyakov has led the way, but no one to date has produced a definitive positive or negative result. So far the difficulty, in my opinion, has been due to two factors: (a) it involves a mountain of work, and (2) the identification of associated noise events is more subjective than one would desire it to be. There is no known “fingerprint” of the impact event that will tell us that this is the sound of an aircraft impacting or ditching in the ocean. We can put some loose bounds on the duration of the impact event, and perhaps also some loose bounds on the frequency content, but that’s about it. Many naturally-occurring ocean sounds (such as a large piece of glacier falling into the water) will be similar.
In the end, I think the best one can do is (a) not assume any specific event location other than the SIO, (b) remove subjective judgment as much as possible from the event selection process, and (c) demonstrate an extremely high degree of statistical improbability that the selected noise events are not associated with the hypothesized impact event.
Victor – I think it’s called groundhog Day. It’s a continuation of what we have always had, and that is something that has prompted me often to haplessly state that this just doesn’t look right.
Number crunchers are maily in the flutter camp not surprisingly, but the flaperon itself might suggest there was a sharp onset of force. The back half of the flaperon represents the bulk of the aerodynamic force on the hinges, yet it’s gone? Sudden or cumulative? It looks sudden and if I had to bet I’d go that way, but out of Dennis’s book – maybe it’s a good thing I don’t have to bet….
Sorry, it appears my assumption that the thing that differed in 9M-MRO’s Log Books to the flaperons in the various articles on the Net, was not contributed to an AD, but to a repair made to it. As this post will require moderating if I give you more than one link, I will just leave it as one, but there are more on the Net on this subject if you care to look:
http://www.bloomberg.com/news/articles/2015-08-06/flight-370-probe-said-to-examine-repair-clue-to-tie-part-to-jet
MAS are in a hell of a spot. If they continue to admit that it is off 9M-MRO then they are in the gun for not keeping good maintenance records, and that would severely affect their resale of any of their aircraft… not to mention other possible repercussions. If they make a retraction, they suffer an even bigger loss of face from the families and the World.
@DennisW
As far as the NTSB and Boeing go, it was published in an American newspaper (New York Times).
Boeing and the NTSB would most certainly have seen it and so I’m sure if it was not true they would have very quickly denied it.
Have you seen anything of the kind? I haven’t, so the only assumption then can be that it is true… unless you have information to prove otherwise?
As expressed by many other commenters, it is frustrating how little official information we obtain from the French flaperon investigation.
Not even which side of the plane (left or right) this flaperon originated….
Considering that in the cost of searching for MH370 in the SIO, that ANY little bit of information would have been helpful, the French investigation results made public are deeply unsatisfying.
I agree, although if you spend a lot of time digging around on the Internet, you can generally find answers. The flaperon came from the starboard (RH) side of the aircraft, although I guess this was only deduced from the images in relation to the hinges… not from an official French Report (What report you say?… there ain’t one!)
@Gavin,
The fact the flaperon came from the RH side is because of a maintenance reference number painted on an access panel (657BB).
@Matty,
I don’t buy the flutter routine; the inboard ailerons of commercial aircraft generally are not mass balanced because they don’t need to be. They are located at part of the thickest section of the wing, so tendency to aeroelastic induced flutter would be minimal; remember these are attached to 2 hydraulic rams and they are not that big in the scheme of things.
OZ
@Matty/VictorI,
I posted this yesterday, here it is again if you haven’t seen it.
http://www.japantimes.co.jp/news/2015/08/23/asia-pacific/mh370-civil-investigators-fear-criminal-probes-hurt-efforts-make-flying-safer/#.Vdw8NTYViUl
This is why the information coming out of the French is so thin.
OZ
@OZ,
Parts do not have to be big to be susceptible to flutter. Think of the modified P51 that crashed at Reno. The likely cause was flutter of the elevator trim tab, a part measuring about 15 x 3 inches. It quickly detached and created an unrecoverable hi-G climb.
Any moveable control surface can be susceptible to flutter, independent of any aeroelastic movement of the wing.
Here is a paper which has a description of flutter testing of an Airbus aileron. It’s about the same length as the flaperon, but not as wide.
http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19840019767.pdf
Also, consider that at the likely time that flutter developed there was no hydraulic power, and only the RAT would have been providing limited electrical power.
@Fitzer-Flyer,
That’s a discussion paper from a 1982 on the L-1011; hardly applicable today!
You missed my point; a P51 (one of my favourite aircraft though) is not hydraulically powered; flutter is easily introduced by the fact it relies on aerodynamic stability alone (shape; then the cables, pulleys, arms and bearings which are the only attaching parts – the weak links!).
As for no hydraulic power (B777) consider this:
1/ The RAT also supplies hydraulics to the centre hydraulic system for essential hydraulics to flight controls.
2/ The flamed out engines just don’t stop rotating (it’s called wind milling). Windmilling engines will rotate the accessory gearbox on the engine which also drives hydraulics to flight controls (let’s call that RAT 2 & 3).
Not too much flutter in a powered flight control surface(particularly an inboard one)
OZ
@Gavin
Regarding the french expression “entre deux eaux” :
Originally the expression is “nager entre deux eaux”. Which directly translated is : swim between two waters.
It is used to describe a ship that has to fight currents to stay on it’s course.
It has become a figurative expression to describe a person that can’t make his mind between two opinions.
In the literal sense, it usually means : right under the surface of the water.
But it’s also often used in the figurative sense to describe a person that just woke up and is not completely awake yet. In this case it means “between two states”.
In the spoken language it is used loosely to also mean “floating partly submerged” like an iceberg.
Note : this sentence is not a direct quote from the aeronautical expert. The journalist is summarising what that person told him.
Finally, this article is not meant to be a scientific factual report… the reporter doesn’t seem to know more than the official statements made and so does the people he interviews.
I wouldn’t put too much faith in the sentence that says that the experts can’t be 100% sure it comes from MH370.
Hope this helps
Cheers
Sinux
PS: french is my mother tongue
@sinux, Far from helping, I fear that you are intentionally trying to muddy the waters.
@jeffwise
🙂 My aim wasn’t to muddy the waters! How could I ? Aren’t they already muddy enough 😉
Just pointing out that the expression in itself is vague and the meaning varies between speakers and between places.
One would have to ask the expert from Toulouse to be sure what he meant. Sadly he is anonymous…
The journalist says “According to the anonymous expert … it would have drifted, plunged between waters, at a few meters of depth …”.
I guess we’ll have to take his word for it.
@jeffwise:
I think you’re too harsh. Why don’t you offer him to remove the line with the iceberg?
@Gysbreght, @Sinux, If I came across as too harsh, it is perhaps because I have become frustrated by what seems like an onslaught of (as Victor has already pointed out) misinformation, an unknown proportion of which may be deliberate. Since the appearance of the flaperon these pages have seen an influx of new voices, which is great, but some of them have created a good deal more heat than light. The phrase “entre deux eaux” is not particularly ambiguous; the quoted source is clearly stating that the piece was floating several meters below the surface. This was discussed here and resolved. Perhaps Sinux was not around to read it. It’s an occupational hazard of this particular game that new people are constantly coming in and raising issues that have already been hashed out; our job is to welcome them. At the same time, we have to keep moving the discussion forward, and if I seem excessively paranoid it’s because I well know certain people who (mostly elsewhere rather than hear) are deliberately trying to obscure the correct interpretation of recent events.
All of this is in the context, of course, that we’re discussing reports that have yet to be validated, so all this squabbling may be beside the point anyway.
Sinux,
When I do a Google search for the words:
flottabilité “entre deux eaux”,
I see many references to objects floating completely submerged with neutral buoyancy, i.e., flottabilité nul. Can you please provide some references in which an object partially submerged like an iceberg is referred to as being “entre deux eaux”?
Lauren Hutton is having trouble commenting on the site — some problem with WordPress I guess — and asked me to put this up for her:
Jeff,
As a not new but very religious reader of (despite being a very infrequent contributor to) jeffwise.net, I’m quite curious about your statement:
“I well know certain people who…are deliberately trying to obscure the correct interpretation of recent events.”
That strikes me as a fairly startling claim and provokes a whole host of questions, not the least of which are who and why? Random troublemakers? State actors? To simply create a confusion? To mislead the world as to the true sort?
As someone who has come to admire you for your journalism, your remarkable ability to moderate what I think to be the most coherent forum on this topic, and your willingness to consider all scenarios (including your own spoof as well as its shortcomings), I don’t doubt what you say at all.
I guess what I’m wondering is if you know something more concretely that the rest of us just presume, and if so are you planning a piece about it? It would be quite a read…
Following up on Lauren Hutton’s post, here’s a translation of a paragraph from a German article about China Airlines 006, describing the damage incurred when the plane reached Mach 0.99 in a near-vertical dive:
Note that no damage to the wings is mentioned, and the accompanying picture doesn’t seem to show any. (http://www.spiegel.de/einestages/sturzflug-im-jumbo-jet-a-946913.html)
I seem to recall accounts of Silkair 185 which also mentions pieces of the tail being ripped free, but not of the wings.
@Scott, Thanks for your kind words. I don’t have definitive answers to the question motive–certainly not enough at this point to publicly point a finger–but I suspect that the answer may be “all of the above.” If I come across conclusive proof I’ll certainly say so!
@VictorI
In the spoken language the “between” is more often used “horizontally” if you will (see “nager entre deux eaux”).
In the article it is used “vertically”. And that might be why there is so much confusion.
Personally I’ve never used this expression in the “vertical” context although according to the dictionary it’s not wrong to do so.
I was also tired of seeing so much debate on that expression on diverse forums.
Taken out of the context of the article the expression has different meaning.
As Jeff mentioned, in the context of the sentence of the article, there is no interpretation possible : the flaperon was submerged under meters of water.
Case closed ! At least grammatically 😉
@scott
Really “scott”?? Take your 5 paragraphs and go play sh•• disturber somewhere else
How expensive would it be for Boeing to remotely fly a 777 hull and try ditching & high speed impact, see what happen in both cases to the flaperon?
Also they could try with some very advanced simulator they possibly have although it wouldn’t be as precise as real thing.
Sinux: Thanks for your posts and help understanding the language.
FWIIW…I came to the conclusion early on that the flaperon floated like an iceberg (at least part of the time), independent of the statement, because that is what the design, materials (aluminum, CF and honeycomb sandwich), physics and photos suggest to me. If all the voids except the sealed honeycomb cells filled with water, it would probably float with a net buoyancy of only a few lbs, or perhaps 5-10% of the ~100 lb weight of the flaperon. Given the relatively dense (heavier) leading edge, made mostly of aluminum, it probably floated leading edge down and trailing edge (consisting of mainly sealed honeycomb) up, partially above the water line.
From the NTSB Accident Report No. AAR86-03:
All the damage found on the airplane occurred during the descent and was caused by aerodynamic overload forces.
Wings and Engine Pylons. — The wings were bent or set permanently 2 to 3 inches upward at the wingtips; however, the set was within the manufacturer’s allowable tolerances. The left outboard aileron’s upper surface panel was broken and the trailing edge wedge was cracked in several places.
I have always been suspicious of the Malaysian RADAR thus I had the following private conversation with one well known contributor to get their feedback:
Could the radar hits that make the turn back path be of other aircrafts ?
These aircrafts would normally fly those paths but their identifiers were purposefully removed or the radar system in Malaysia could not capture their true identifiers ? So then on playback there are a bunch of hits that almost show a fly back over Malaysia east to west.
If I may ask those more familiar with RADAR playbacks if my scenario is feasible?
@airlandseaman: Have we found a reference stating the B777 flaperon was made from honeycomb? There was a reference recently posted which described the design, fabrication, and testing of an aileron for the L1011 with a composite structure that was not honeycomb and had “lightening holes” in the spars. Yet the B787 flaperon was made with the honeycomb-CF-aluminum sandwich you described. If made from honeycomb and the voids remained sealed, then floating is possible. If made in the manner of the L1011 flaperon that was described in the paper, it is less likely that it floated. Do we know yet which design was used?
@susie
I’m sorry, Susie, I don’t understand your reaction, though there is no need for an explanation.
Quite simply, I posted what I thought was an honest question from one journalist (though not remotely involved in aviation reporting) to another, who clearly has deep knowledge of the subject matter and many sources connected with the various investigations around this incident (official or not) and who, as we’ve seen in various posts, continues to make his own inquiries of various experts.
To me it was a simple and obvious thing to ask with no agenda other than genuine interest nor any intent to derail the other threads here.
@scott
My apologies if I inaccurately challenged your integrity and sincerity. It is knee jerk to defend the ones who chose to extrapolate the extraneous information with such a steadfast commitment and not clutter these pages with deterrent thoughts or journalistic profiling
Victor:
Yes, I have seen such references, but I will have to look for them. In any event, the skin is clearly honeycomb in the photos (about .50 to .75″ thick). I believe some of the ribs or spars may also be made from honeycomb sandwich stock.
@airlandseaman: Yes, I would be interested in seeing reference to the structure of the flaperon. I see evidence of honeycomb structure under the skin, but not elsewhere.
@Victor
Several articles make a vague reference to a honeycomb structure, but I could not find anything that actually details the internal construction. Odd really, given all the attention this part has received.
Victor:
Check with Don.
https://www.reddit.com/r/MH370/comments/3f3exn/why_would_a_flaperon_float_particularly_after_a/ctmy4mp
I defer to the signal data experts on all matters pertaining to the signal data. But when the topic turns to The extraction of probabilistic indications from sparse data, we’re on MY turf.
Nothing I’ve read does a better job of explaining rhe limits of Bayesian revision of estimates than does Stephen Philbrick’s wonderful paper, “An Examination of Credibility Concepts”:
https://www.casact.org/pubs/proceed/proceed81/81195.pdf
Even if you can’t do the whole 25 pages, the simple example (“Credibility and Marksmanship”, bottom of p.3) is not to be missed.
The bedrock principle behind the example is self-evident: if one data source leaves you with 90% confidence the answer is Los Angeles, and then an equally valid data source comes along and suggests 90% confidence in New York, it does not necessarily follow that the best revised estimate is Grand Island, Nebraska. That MAY be an appropriate revision if the estimate is to the expected impact point of an asteroid – but not if the estimate is the location of an international rock star’s upcoming lone US concert.
The Philbrick example – four shooters aiming at four closely-spaced targets – asks us to re-estimate the location of the next shot, given the location of the first shot. Bayes forms a new estimate by combining his a priori guess (middle of the four targets) with the new info (location of first shot), to come up with something in between.
Even better though, Philbrick demonstrates, is to make use of what is already known about the distribution of shots from each shooter. the shot’s location may tell us a lot about the identity of the SHOOTER, which may in turn yield a much better expectation for the second shot.
The actuarial profession would have me use this paper to price your insurance policies. I think it can help us use drift analysis to pressure-test the signal analysis – and vice versa.
I think that
1) those who say 1 piece of debris is insufficient to conclude much of anything are absolutely correct.
BUT:
2) the signal data experts have been telling us Los Angeles, the drift experts are telling us New York, and the Bayesian analysts are telling us this means we need to search Grand Island.
The danger in moving the search to the Bayesian analysis-indicated point is clear: the probability distribution functions for impact point which flow from each camp are (thus far) SO disparate (hardly any overlap), they may well render a compromise location the LEAST likely, not the most. S34 is such a northern outlier for the (CAD-respecting) signal data experts – and such a southern outlier for the drift experts – that we must first run to ground whether those two camps aren’t, for whatever reason, shooting at different targets.
Brock,
I’m afraid I’d give zero weight to the first of your points of evidence (WA locals), that one is extremely tenuous I feel.
For the rest – are all these models equal do you think? Each of these tools has a particular purpose. Does a larger number of inappropriate models carry more weight than a smaller number of appropriate ones?
I have great respect for Eric Van Sebille, and his numerous sensible comments on the MH370 debris drift problem. However his adrift.org tool was specifically developed to model ocean currents at the very largest (global) scale and over timescales of centuries, in order to track ocean-scale concentrations of plastic garbage input from coastlines. It is based on a statistical treatment of global drifter data, applied to a grid with cells approx 110km per side, The tool suggested that over timescales of 10-100 yrs, 6 garbage patches are formed in different oceans globally, and that over timescales of 100-1000 years, there is intermixing of debris between these patches. An appropriate tool to answer the question “does a substantial portion of MH370 debris injected at a very specific location in the SIO on a very specific date make landfall to the east within 16 months”? Maybe not.
I have looked in some detail into the other models that you mention, and have several observations on their various strengths and weaknesses. I may stick that into a subsequent post. However to cut this short, if we are really interested in answering questions about the movement of debris off the coast of Western Australia, the CSIRO models would seem to be good candidates :
– these guys are SIO specialists. Just look at the publication history of Griffin, Schiller and Peter Oke
– they were tasked specifically with modelling debris drift from points along the 7th arc and have models tailored to this goal
– they (uniquely) have access to all the data from the many drift buoys dropped by AMSA at various 7th arc locations and various times during the search
The latter point is important; despite the relatively short timescales involved, the data will provide important high resolution calibration between model elements and true drift. Eg if currents are doing x, and wind is doing y, and waves are doing z, but the buoys are doing w, the learnings in history matching are likely to be useful for forward prediction beyond the battery life of the buoys.
The buoy data may incidentally allow some calibration of the effect of leeway (as would the proportion of global drifters that have lost their deep drafted anchors). The fact that they were running with 1% and 2.8% in July 2014 does suggest that 1.5% is not a completely ludicrous number dreamed up to keep debris off WA.
But let’s take a step back. None of these models are perfect. There is of course no clear answer to the quantity of debris that ‘should’ have washed up along WA shores. This may be troubling for those who need to see things in black and white, (or who for example seem to expect a tidy, orderly, linear sequence of events during search operations that are, by their very nature, typically chaotic and subject to abrupt changes of direction as new data comes in and different opinions hold sway 🙂 ).
However the fact that there are serious models suggesting dispersal of debris away from the early search areas, and away from the nearest shore suggests that such happenings are at least plausible, and that those who used the lack of discovered debris as a rationale to write off a crash close to the SIO 7th arc were premature to do so.
@DennisW
Sorry to take so long in replying. There’s a big time difference here as I’m obviously on the other side of the world to you.
Yes, I’m well aware of the way the media often make mistakes – believe me they have made plenty of them in what they report of me over here. I now try to insist they run the article past me before they print, and then their editors go and change it and so still stuff it up!
The point is Boeing and the NTSB don’t like bad publicity about them, so I’m sure if someone was making false claims about them, I’m sure they would have something to say about it,- and coming directly from them, I’m sure the papers would print it. They are two powerful organisations that make big news when they say something… in other words they wouldn’t be ignored by the newspapers.
Over the years I’ve found that it is the little snippets of information that is normally more close to the truth.
The media is more focused on sensationalism rather than the truth. To them the truth is boring…just a fact of life, I guess?
Gavin:
Given the lid the French seem to have put on the investigation, I have to assume that everyone involved is bound by a very strict non-disclosure agreement.
@airlandseaman:
Looks like metron people have a 1d model memorandum published on aug 12th. That is what they say in aug 24th paper. Do you have a link to aug 12 work?
Re Dr Bobby Ulrich from Michael Molinaro … “I believe a great many MH370 followers would relish reading it.”
I strongly second that.
It should also be said: a huge thank you to the main protagonists here for their enduring work (they know who they are, they do the work).
Is there a possibility the flaperon left 370 well north of the theorised crash zone (fractured bolts etc), seemingly conforming to the drift models?
@ALSM: drawing upon all this knowledge you’ve accumulated on the density distribution of this specific, heavily damaged, and surely at least partially waterlogged flaperon – and how the ocean might orient it on its journey – what is your best estimate for the average percentage of windspeed we ought to assume this item drifted downwind (for reference: CSIRO used 1.5%, van Sebille used 0%)?