After a Boeing 737 operating as Flydubai Flight 981 crashed in the Russian city of Rostov-on-Don Saturday, preliminary accounts suggested that the plane had clipped a wing or struck the ground with its tail while attempting to land in stormy weather. Indeed, in a story published later that day, RT.com quoted Rostov region governor Vasily Golubev as saying, “The plane was descending and then suddenly dived down. Experts say this was an air pocket that dragged the plane to the left of the runway center. And the plane debris were scattered to the left as well.” Obviously, there is no such thing as an “air pocket.” But it makes intuitive sense that a plane attempting to land in high, gusty winds might succumb to shear at low altitude and low airspeed as it nears touchdown. But this, it appears, is not what happened at all. Frequent contributor Victor Iannello has created a graphic based on ADS-B data transmitted by the plane during its final moments. What it shows is that the plane had descended to land, then aborted the landing and climbed, accelerating as it went. It had already gained 3000 feet altitude and reached a speed of 200 knots when it suddenly plummeted from the sky. Here’s the data in graph form:
This security-camera footage offers a visual sense of what happened:
Mar19 #Rostov, #footage of FZ981 crash, view from Aleksandrovka micro-ds @Eisenhoden pic.twitter.com/TqIdvOmMys
— English Lugansk (@loogunda) March 21, 2016
What happened? Authorities on the scene have found the black boxes and hopefully will have answers soon. For the time being, some have speculated that the plane encountered severe windshear or a microburst, causing it to stall and plummet. But the plane’s descent was nose-down at high speed, so the pilot should have been able to at least attempt to pull up. Personally, I’m reminded of AA587, which crashed in 2001 on takeoff from Long Island after the pilot flying applied to much rudder after encountering wake turbulence from the plane ahead of him on climbout, causing the vertical stabilizer to rip off; the plane dived nearly vertically into the ground. If something similar happened here, parts of the tail should be found at some distance from the main wreckage. Another case that may offer parallels was Kenya Airways Flight 507, which crashed in 2007 while on climbout in bad weather. The pilot lost situational awareness while the autopilot was only partially engaged, the plane entered into an increasingly steep bank, and plunged into the ground. What’s different in the present case is that the plane impacted right on the runway it had been trying to climb away from, implying that it stayed on the same heading the whole time. (That is to say, it hadn’t gone into a roll.) Another unusual aspect of the case was the fact that the pilots had been holding for two hours before making a second landing attempt. I asked Phil Derner, an aircraft dispatcher and aviation expert, for his take. He replied:
For me, as a dispatcher, 1 hour is my max to let an aircraft of mine hold. It’s just a waste of gas; might as well divert and wait for conditions to improve. Shit, even fitting an additional 2 hours of holding fuel to a flight is tough as it is, and then to burn it away in a hold? Also, I only let my flights sit in a holding pattern if I think they WILL get it. If conditions don’t look to be improving right away, I won’t even have them hold…I divert and would rather have them wait it out on the ground. It saves gas, and is safer on the ground. But then again, I don’t know all of the conditions they were facing, what conditions were at their alternate airports, etc. There are so many variables and we just don’t have a lot of info, so it’s tough to determine or judge. But 2 hours….damn.
Meanwhile, on an unrelated topic, I might as well put up a picture of the latest piece of aircraft debris, this one found on a beach in South Africa. Not many details forthcoming yet, but it’s worth noting that MH370 was equipped with two Rolls-Royce Trent 800 engines. 
A quick glance at there not-very-high-res images suggests that the piece is roughly similar in appearance to the two pieces recently found in Mozambique, though perhaps somewhat more discolored/weathered. Apparently the piece is on its way to Malaysia.
UPDATE: Here’s another picture that @Susie provided a link to in the comments section:
@Jeff
Your numbered points leave out a few things worth bringing up. Again, following the numbers.
1> What would be the motive for hijacking the plane at all? It is worth noting that an airliner has never been hijacked for any cargo or passengers in the history of commercial aviation (I went back to 1970).
2> There are many places on the 7th arc that can satisfy the ISAT data that have not been searched at all – mostly to the North of the current search zone. The debris locations also infer a more Northern terminus. If am referring specifically to Brock’s paper showing the forward and reverse CSIRO drift models overlayed on the same graphic. Also the Geomar reverse drift study strongly favors a more Northern location.
3> I think it is premature to say anything about the new debris relative to the absence of biofouling. Frankly, I am surprised that experts would say anything on the basis of pictures. In any case we will be getting a lot more info on that in the days ahead.
A point I am struggling with which you did not raise is why, if the pilot did not intend suicide or mass murder, no attempt was made to communicate a location when he was aware that running out of fuel was imminent (so that the PAX would have a better chance of survival)? That is a real show stopper for me.
”The absence of wreckage on the seabed in the southern Indian Ocean indicates that if the data was not spoofed, the pilot must have been suicidal.”
Nope, what if the hijacker’s demands were to fly to Perth,Australia.
Oleksandr – 18+ years old is not a kid. He’s every bit a young man and they are more disposed to picking up girls than flotsam. If for the purpose of handing it over fine, but as a curiosity?? All this stuff would sit on a back seat, is covered in writing, and not traceable? You couldn’t get anything more emblematic than Rolls Royce – but that’s what we got. Two of these guys are Mh370 geeks, the young fellow apparently not. Nothing yet divulged about the flaperon but I’m hoping this little burst of transparency might help it along.
Sharkcaver – Absolutely. Victor has been as diplomatic as he has been diligent and I would not fear him for a moment if I found something today. Not a good look. If the next bit says MAS I’m calling in Brock to do a probability.
@airlandseaman: Figure 6 of the ATSB’s Dec/15 report strongly suggests its right engine flameout time estimate lies within a (roughly) five-minute window centred on 00:04:00 UTC. This window is between 11 and 16 minutes before their 00:17:30 estimate for left engine flameout, and between 13 and 18 minutes prior to the generation of Arc 7.
I have an e-mail seeking clarification of their distribution (mean, std dev / confidence interval) of this estimated flameout time. I will forward what I learn to you, and to this group.
If MH370’s right engine flamed out this early, the distance between Arcs 6 and 7 – per the speed function in your paper, which was calibrated to your flight simulator results – simply cannot be covered.
ESPECIALLY if the left engine didn’t flameout until 00:17:30, thus denying for an additional 100 seconds the left turn you say is required to cover the distance even at HIGHER ground speeds.
I raised this issue months ago – you said then that the ATSB fuel analysis must have been wrong.
It would be interesting if the distance turned out to be coverable under the assumption that MH370 crossed Arc 6 at a more perpendicular angle. But this forces us into “low and slow” scenarios – with all their concomitant coincidences. Plus a brand new, “bonus” coincidence: bad initial fuel analysis – combined with bad initial login sequence timing estimate – by fluke obscured for two years the logical impossibility of the officially endorsed end-game flight dynamics.
@Matty: @littlefoot’s analysis is excellent: though I am on record as skeptical of, well, everything I read about MH370, it has never struck me as particularly weird that everything is “photogenic”, because of selection bias.
But my statistical “spider sense” certainly began tingling on the “Rolls Royce” item. How many bits that size – most with the exposed honeycomb many beachcombers would by now recognize – would a high-speed impact have delivered to shorelines by now? Yet we get the logo piece first?…
Yes, I get that most beachgoers aren’t following the news as closely as we are, and are less likely to recognize honeycomb. But I expect those who ARE following the MH370 case will be covering far more terrain than will the casual comber. So yes, I do agree with you: as this running streak of debris readily identifiable from photographs as “likely MH370” keeps extending, it does start to look less and less like random finds, and more and more like a campaign to move minds.
@Matty@Brock
Brock, I am not trying to “steal your thunder here”, but I came across the following information.
For a painted 777-200 the weight of the base paint plus customer marking paint is 215 KG (typical).
For a polished skin the weight of the customer markings alone is 50KG (typical).
Found the info in a discussion of the costs associated with painting an aircraft versus polishing an aircraft.
http://www.boeing.com/commercial/aeromagazine/aero_05/textonly/fo01txt.html
I would infer from this information that customer markings would be found on roughly 25% of recovered debris.
@Brock McEwen
Brock, ALSM is a very decent guy. It’s just that he was unlucky enough to back the wrong horse at the off, and now it’s to late to change.
This is where the smart money is: distance between 6th and 7th arcs (35,000ft) on an azimuth of 187.4deg = 61.5Nm, in 8.5 mins = groundspeed 434Kts, plus headwind 33Kts = Mach 0.804 (for the temp) which is exactly same airspeed as the 5th to 6th arc leg (for the temp) This shows that the plane never slowed down approaching the 7th arc, which in turn means the engines did not flame out, which means there must have been someone in the left hand seat cycling the SDU on and off to make it appear otherwise.
Everything then makes sense – No wreckage within 40Nm of the 7th arc. Debris pointing to controlled ditching. Pilot looses plane in the deep. Mission accomplished. Malaysian government go into shock. Recover composure & begin damage limitation, assisted by Aussies (who really had no choice in the matter, when you think about it)
@Matty@Brock
Note that one could argue that painting the whole plane takes ~150KG of paint, and the customer marking takes 50KG therefore inferring 33% of the recovered debris would contain customer marking. Take your pick.
This latter interpretation assumes the entire plane is painted before the customer marking is added as an overlay. The first interpretation assumes the customer marking paint is applied first, and masked to apply the rest of the paint.
@Dennis, right now we’re just talking about written letters and some numbers which were in case of the new debris very helpful for identifying and locating the debris. The color for lettering will only be a small percentage of what is needed for applying colors and designs.
That all three debris pieces sported some crucial letters and numbers is extraordinary. But then again the sample size so far is rather small for spotting a distinctive trend. And I conceded that pieces which are marked by letters and numbers have probably a higher chance to get reported.
Dennis – I don’t think it’s a kilogram issue. It’s more visual.
@Matty
Just using KG’s to guess at a ratio of bland debris vs debris with any distinctive making.
@Littlefoot
Yes, I understand what you are saying. The markings on the newly found debris would certainly constitute a subset of carrier markings. Just bookmarking the broader ratio for use as more debris is located.
I’m only half serious atm about the writing but, it’s a trend that needs to reverse for this to be credible. Stencils and decals represent a small part of the plane. Four in a row and I’ll be shaking the cot.
@all might seem off topic, but I dont think so (doing some far less mysterious things than MH370 here too, to maintain mental health, Dareks work is great since nineties when I found him; and he is right, imho)
“If web sites don’t take responsibility over the accuracy of their content, the Internet will end up as nothing more than a collection of useless web sites that exist merely to host banner ads and collect Google click-through revenue. Google wins, consumers lose. Is this why kids today get into engineering and computers? It is certainly not what attracted me to computers in 1980.” (whole article matters and relates to media in war, these days)
http://www.emulators.com/docs/nx09_leopard.htm
Brock,
“It would be interesting if the distance turned out to be coverable under the assumption that MH370 crossed Arc 6 at a more perpendicular angle.”
Yes, it is coverable. Even in excess. See Fig5 of my TN-CTS-Rev1.1 note. I am working on the replacement of CTS assumptions with V/S or FPA mode, and my preliminary observation is that a new trajectory will roughly be the same as depicted in Fig5, slightly worse residuals.
Folks. I know that there are many on here who are far more technically-skilled & knowledgeable than I am and no doubt some of you have looked at fuel modelling in some detail.
We are told that fuel exhaustion at ~0020 is compatible with expectations and that it fits with a power-cycling explanation for the unscheduled logon at 0019.
My question is this:
1. How many hours and minutes was fuel expected to last from presumed fuel remaining at 1707 (climb completed, fuel remaining known with reasonable confidence that fits with fuel load taken on)?
2. For this “duration” of fuel endurance, what are the implicit assumptions on speed/altitude? How much difference would an alteration in those assumptions make?
3. If we assume flight at constant Mach 0.82 (ie same as reported immediately prior to disappearance) at FL350 (on whatever direction/route we choose to believe), would endurance (expressed in range or time) be any different from 0020?
Thanks in advance…
@Jeff
Thanks for clarifying your view.
The age of the items (i.e. time in water) sampled by Miriam Goldstein was not determined by her. Since the lifetime of plastic debris in the oceans (i.e. before it disintegrates or is beached) may be very long, I don’t know what can be deduced from her data on the rate of biofouling accumulation. The debris from the Japanese tsunami is much more useful for these purposes as its time in the water is known, at least the maximum.
The current search area is based on many more assumptions/models than the simple flew south/crashed/debris washes ashore on IO coast. The failure to find the aircraft in that area does not support an alternative and vastly more complex solution, IMO.
The main problem I have with these complex solutions (e.g. Russian conspiracy/data spoofing/debris planting) is that they are unfalsifiable, that is any data (current or future) can be dealt with by suitable adjustments. That gives them huge powers of explanation.
In science they would be described by some (like Popper) as metaphysics. The potential for endless adjustments in the light of new data is a problem for all theories, of course.
@Richard Cole, Yes, the debris from the Japanese tsunami is very valuable in determining what stuff should look like after 2 years at sea, and contrary to what Warren says, it is very, very different from what washed ashore.
I’m not sure what you mean when you say “the current search is based on many more assumptions/models than the simple flew south/crashed/debris washes ashore on IO coast.” The current search area isn’t even based on a theory at all; it’s based on a “ghost ship” assumption married to the DSTG analysis of route probability based on BTO data (which I think is excellent btw). The ATSB has more than once explicitly stated that they are agnostic as to how or why the plane would arrive at the “ghost ship” state. The only plausible explanation anyone has been able to come up with is the “pilot suicide” scenario, which has its own flaws, as I’ve discussed here at length.
Finally, it is silly to invoke Popper in criticism of the spoof theory. “Spoof” is actually more easily falsifiable than the suicide theory, which can be stretched to account for any inexplicable series of pre-crash actions, like flying six hours and then stretching out a glide to prolong the fatal moment, which (as previously discussed on this blog) psychologist say is not something they would expect a suicidal person to do. The issue, however, is not the structure of the theory but the extent to which one is willing to accept data that negates it. When I presented the spoof theory in 2014 I acknowledged that it could be proven wrong at any moment, simply by the plane’s wreckage being located. At the time it seemed to me that, if a spoof had been perpetrated, it would be reasonable to expect the perp to salt some pieces of debris here and there.
Now, if these pieces of debris had washed up with an amount of marine encrustation consistent with two years at sea, and carrying organisms whose species mix and oxygen and barium isotopes suggested a drift from the southeastern to the western Indian Ocean, I would consider that powerful evidence against a spoof scenario. Such evidence might even exist at this very moment, in a hangar in Toulouse! It could be released today, and spoof will have taken a bullet through the heart!
Meanwhile, the IG and ATSB’s confident assurances that MH370’s wreckage would be found in the course of the seabed search have been proven wrong. Yet there are some among them who assert that is not the case, that everything is fine and their interpretations are still on track. Not surprisingly, there is a strong Venn diagram overlap between people who hold this view and those who say that it’s perfectly reasonable that a piece of corrugated debris can spend two years at sea and still look like it just came out of a box.
In short, both suicide and spoof theories are falsifiable, the issue is the interpretation of the data.
‘The current search area isn’t even based on a theory at all; it’s based on a “ghost ship” assumption married to the DSTG analysis of route probability based on BTO data (which I think is excellent btw)’
‘“Spoof” is actually more easily falsifiable than the suicide theory, which can be stretched to account for any inexplicable series of pre-crash actions, like flying six hours and then stretching out a glide to prolong the fatal moment,’
‘ if these pieces of debris had washed up with an amount of marine encrustation consistent with two years at sea, and carrying organisms whose species mix and oxygen and barium isotopes suggested a drift from the southeastern to the western Indian Ocean, I would consider that powerful evidence against a spoof scenario. Such evidence might even exist at this very moment, in a hangar in Toulouse! It could be released today, and spoof will have taken a bullet through the heart!’
Couldn’t agree more with you on these topics!
Jeff – a quick, big thanks to you and to all the contributors for such a great blog!
Hi all,
Under a hypothetical scenario of the debris being planted by those involved in the disappearance of MH370, I am wondering what are some people’s theories on the motivation of these perpetrators?
Given the requisite sophistication of those involved in this hypothetical scenario (spoof) does anyone else think the distribution of clean* debris would look like an attempt to embarrass Malaysia and/or those parties responsible for the search? Or even intended as a message or ‘warning shot across the bow’?
Seemingly perfectly timed after the 2 year anniversary and release of the 2nd Interim Statement…
“Dear Malaysian Government: How are you going to explain this contrary evidence to the general public and maintain your current search narrative?
Yours in readiness to inflict more damage and embarrassment, Signed Perps.
P.S. You might want to consider what we could do next…”
A hypothetical scenario, just to be clear.
*debris designed to have an appearance of a drifting time an order of magnitude (or more) different than the time to disappearance.
Been reading despite not commenting. Interesting takes, conjectures, assumptions , sophisticated maths etc galore but still not a shred of convincing evidence as to where is MH 370’s tombstone in SIO, that is if it got even a whiff of the IO in the first place. And almost everyone is unconcerned with neatly stencilled finds of the same livery and scrubbed clean flotsam that even barnacles are leery of.
Me thinks Fat Controller and his minders ? pals? in both a pentagonal building and an oval office must be having a mighty chuckle over the kerfuffle whilst sipping Don Perignons . And a poor kiwi who saw something tangible at exactly the witching hour 2 or so Suns back wallows in forgotten anonymity. Life is strange, isn’t it folks?
Mmm ….. Reunion check, Mozambique check, South Africa check bet the next one on FC’s menu would be Madagascar or the Comorros or even Seychelles
@Richard Cole said, “The main problem I have with these complex solutions (e.g. Russian conspiracy/data spoofing/debris planting) is that they are unfalsifiable, that is any data (current or future) can be dealt with by suitable adjustments. That gives them huge powers of explanation.”
I agree that theories based on altering evidence are problematic because it is difficult to definitively prove or disprove a theory, barring a whistle-blower, additional witnesses, or evidence of tampering.
However, do you have an objective metric for falsifying the theory that the plane crashed in the SIO? The locus of possible crash sites gets unmanageably large to search if we consider the possibility of pilot inputs along the route, before and after fuel exhaustion. That combined with the possibility that the crash site was searched and missed makes it hard to falsify the theory.
@Matty
you said:
“I’m only half serious atm about the writing but, it’s a trend that needs to reverse for this to be credible. Stencils and decals represent a small part of the plane. Four in a row and I’ll be shaking the cot.”
A far more interesting observation (to me) relative to the debris is the lack any debris from inside the aircraft. This would seem to suggest that the plane sank relatively intact. I put the whole biofouling issue in the “tempest in a teapot” category until we get some feedback from people actually “breathing” on the debris.
@Victor
I think an objective person looking at the debris drift analysis would rule out the current search area.
By this I mean the reverse drift analysis which shows the contrast in particle density between the NIO and SIO. It is truly dramatic. Looking at forward drift analysis suffers from two inherent flaws – first it assumes a terminus (which you do not know), and secondly there will always be a finite number of particles that will arrive at the debris location. Making statements that the drift analysis supports the SIO, as the ATSB and IG have done, is truly preposterous.
@Paul Smithson
Paul re. endurance figures; airlandseaman may help you there. I’m at risk of seeming a cheeky whatsit volunteering someone, but I believe the IG have done some work on consumption rates etc. It’s way out of my league.
@dennisW
Unless you believe the story that seats and suitcases were found on La Réunion before the flaperon and simply burned ofcourse.
@DennisW: I have not studied drift models in depth the way others here and elsewhere have. My general impression is that the science is sufficiently imprecise such that the drift models can’t be used to definitively prove or disprove that the current search area is correct.
That said, I will make the casual observation that the location and timing of the debris discovered by Liam Lotter in December would suggest the plane crashed further north along the arc than the current search area.
Maybe this is the most simple explanation of all for the strange path the airplane took : ‘Researcher hacks aircraft controls with Android smartphone ‘
http://www.theregister.co.uk/2013/04/11/hacking_aircraft_with_android_handset/
http://conference.hitb.org/hitbsecconf2013ams/materials/D1T1%20-%20Hugo%20Teso%20-%20Aircraft%20Hacking%20-%20Practical%20Aero%20Series.pdf
‘Also vulnerable is the Aircraft Communications Addressing and Reporting System (ACARS), the communication relay used between pilots and ground controllers. Using a Samsung Galaxy handset, he demonstrated how to use ACARS to redirect an aircraft’s navigation systems to different map coordinates.
“ACARS has no security at all. The airplane has no means to know if the messages it receives are valid or not,” he said. “So they accept them and you can use them to upload data to the airplane that triggers these vulnerabilities. And then it’s game over.”
@Carla
A link to your claims is always helpful.
@Victor
My remarks are not directed at you so much as at the ATSB and other IG members. One can say that the drift models do not rule out the current search area, but to say they support it is simply not consistent with the observables as presented by both CSIRO and Geomar. I would be perfectly OK with the “not disprove”, I am a long way from OK with “supports”.
It is simply bad science to start with your postulated terminus, and conclude it is possible for the debris to have drifted where it was found. It is good science to start from where you know the debris was found and ask where it may have come from.
@dennisW
http://www.dailymail.co.uk/news/article-3182648/MH370-seats-suitcases-spotted-three-months-ago.html
@Jeff
My remarks were in the current context of no discovered sea-floor wreckage and hence decisions being required on what to do next. Apologies for not being clearer. All theories collapse to a solution if wreckage is found.
I note the debris analysis results that you would accept as strong evidence against the spoof idea (though it seems a rather precise definition, beyond my ability to know if it is feasible to achieve that accuracy). I note that you have also said “Therefore, the more reasonable interpretation of the Inmarsat data [in the absence of detected sea floor wreckage] is that it was spoofed”. This is independent of the debris analysis and on the face of it cannot be contradicted without detection of that wreckage.
At the moment the theory under test is ATSB’s, i.e.
1. the aircraft flew in a manner after 18.22 that can be represented by the DSTG trajectory model, as validated on the validation flights
2. After loss of power, the aircraft terminated within a distance from the 7th arc given by BEA’s loss of control crash model (simplified explanation).
Suicide is not part of the theory. Hypoxia has been mentioned by ATSB, but seems to be a later add-on/justification, not required by the original theory. I certainly have not the faintest idea of what actually happened on board. I presume that a different, suicide based theory would imply a search far from the 7th arc, but I am not proposing that, since I am suspicious of motive-based theories.
I agree that the search is close to falsifying the ATSB theory as stated above. I would also agree that stretching the borders of the current search area (changing the inputs from points 1 and 2) could represent one of ‘adjustments’ I deprecated. The lack of commitment of the search to the last section inside the 7th arc makes me wonder if they are holding back awaiting analysis of the debris which might redefine/constrain the search in some new way. If (if) the debris analysis shows a very high speed (high energy) impact, that might suggest an impact close to the arc, as you have said the aircraft cannot go far _and_ fast.
That”AA587, which crashed in 2001″ you were reminded of crashed due to rudder reversal/stick… not because the pilot applied too much…
Dennis,
Re: “A far more interesting observation (to me) relative to the debris is the lack any debris from inside the aircraft.”
How do you know that? The towelette? The laggage on Reunion? It is getting ridiculous: when debris is coming with ‘MH370 label’, people says it is planted. When debris is coming without label, people say it is not from MH370. You want from inside and labelled “MH370”?
Re: “By this I mean the reverse drift analysis which shows the contrast in particle density between the NIO and SIO. It is truly dramatic.”
Throw away reverse drift analysis into garbage bin. They do not work in this case.
@Carla
Thx. Not exactly what I would call compelling, but I am not going to ignore it either. I would wonder why someone would burn a suitcase washed up on a beach without looking inside. I am not a poor person, but I would certainly open something like a suitcase before tossing it in the fire.
Maybe the trash collector did not want to raise the possibility that someone might claim the contents? It would certainly be worthwhile to talk to this person with the offer of immunity from any claims. The content of the suitcases would be a huge +/-.
@Oleksandr
“They do not work in this case.”
Not sure I know what you mean by that statement. What is unique about “this case” unless you mean “don’t work at all”.
The oceanographers at Geomar and CSIRO thought it was worthwhile to publish reverse drift data. Are you now claiming to be an expert in oceanography?
Dennis,
Re: “It is good science to start from where you know the debris was found and ask where it may have come from.”
Not only where, but also when found. But this has to be done by forward drift studies, not backward.
@Richard
“since I am suspicious of motive-based theories.”
What a strange thing to say. I almost always to things for a reason. I thought most people behaved in a similar manner.
Also there is a huge difference between “basing” a theory on motive, and demanding that a plausible motive simply needs to exist. Therein lies the untenable position held by analysts who claim that a motive is completely unnecessary.
Dennis,
Yes.
@Oleksandr
Time is a reversible variable in this case. The Geomar reverse model includes time.
I would agree that the “arrow of time” is irreversible in some physical processes i.e. no one has seen a pile of rust reorganizing itself into an automobile. However, in the case of drift I have no issue with running the clock backwards.
@Oleksandr @Dennis
Dennis,
Re: “A far more interesting observation (to me) relative to the debris is the lack any debris from inside the aircraft.”
This is a valid point. Nearly all the inside panels, the overhead bins, the panels of the lavatories and the panels dividing the cabin are made of floatable lightweight material. the Debris field of AF447 could serve as an example how much lightweight debris there was and how big those pieces could be. Even cabin trolleys were still drifting on the water after days.
The controlled ditching leage will say, that nothing escaped the fuselage as its integrity was not damaged. But that contradicts the assumed consecutive out of fuel failure of the engines and lacks the explanation how the RollsRoyce piece got smashed like it looks like.
Another coincidence of the parts: All the pieces found so far could be removed from the aircraft without damaging considerably. Nothing structural showed up yet. Not saying though that it is what I said.
Re: falsification (pedantry trigger warning)
Since it has got so bad that we must discuss philosophers of science, allow me to point out that there is no such thing as falsification strictly speaking.
One can maintain the truth of a theory “come what may”, as Quine put it, if one is willing to make enough adjustments elsewhere in one’s belief framework system.
The way falsification is supposed to work is as follows:
H —> O
~O
———
~H
(If hypothesis H is true, then observation O will be observed; we do no observe O, therefore H is false.) Modus ponens.
In actual practice, however, hypotheses are not tested in isolation: there are always an unknown number of auxiliary assumptions. So it’s more like:
H & (A1 & A2 & … & An) —> O
~O
——————————
~H or ~A1 or ~A2 or … or ~An
If the predicted observation is not found, the hypothesis MIGHT be false, but it still could be true–all that is required is that at least one of the auxiliary assumptions is false. Take your pick.
Thus none of the three broad families of hypotheses (mechanical failure; controlled flight inputs; sophisticated hijack) will be falsified by debris objects, no matter how few or how many are found, and no matter how much or how little discernable macrofauna are found. At the end of the day it boils down to Ockham’s razor, and even this is only a Bayesian probability.
@DennisW
I am of the opinion that these extreme outlier cases are rarely what was expected (if and when they are later resolved of course). Therefore, using detailed models of what might have happened, and particularly what some unidentified person or persons might have done, are unlikely to yield useful results. I understand you approach it differently.
On the drift modelling, the structure of eddies looks pretty chaotic. If there is any attempt to model those stochastically I doubt it can be reversed and get the same result. Perhaps not reversing a pile of rust to a car, but a swirl of paints in water back to a few blobs of single colours. The models at adrift.org.au look like simple probability division for each block into nearby blocks, for each unit time. Such a model could be trained against the drifter data running time forward and back. Still not sure that would give the same result forward and back as probability would have to be used as a filter, and that doesn’t sound linear.
@Paul Smithson – There is no one answer to your questions since burn rate (in kg/hr) is dependent on not only the speed but also the altitude and Outside Air Temperature as well as the gross weight of the a/c. There are manuals (FCOM, etc.) available that give the LRC speed at various altitudes that we have been using to wrap our arms around your questions.
The FI report says the Captain ordered 49,100 kg fuel for the flight (since there was still fuel in the tanks, I believe this should have said the Captain ordered enough fuel so that the total was 49,100 kg) The FI then said that this was enough for 7h31m while the actual flying time to Beijing was 5h34m giving them about 2 hours of reserves.
The 49100/7h31 gives an average burn rate of around 6,532 kg/hr. The FI report gives a flow rate of 46,835 lb/hr (21,230 kg/hr) at takeoff and 31,369 lb/hr (14,220 kg/hr) at climb some 10 minutes later. This is significantly higher than the average fuel burn rate as it is assumed that the extra fuel burned during the climb is regained during the descent. While this cannot be precise it is close enough for fuel loading after the 2-hour reserves are added. Note, this extremely high burn rate is for a relatively short period (here about 20 minutes of a 5 ½ hour flight.
You will see that many commenters use a speed called LRC. One reason to use this is most of the fuel burn tables in the various publications are based on LRC. The speed is defined the speed that costs 1% more than the maximum range speed. Typically this means a 3% – 4% speed increase at a cost penalty of only 1%.
Another speed that is discussed as the ECON speed and uses a Cost Index that is the cost of time divided by the cost of fuel. The CI for MH370 was 52. Generally, the cost index speed lies somewhere between the maximum range speed and the LRC. Dr. Ulich’s paper shows a graph of cost index versus range.
Note, there are two burn rates that apply to a/c flight. One is the endurance or how many hours can be flown using all of the fuel on board. The second is range or how far can you fly with the fuel on board. These are usually two different speeds.
To complicate matters, besides speed, the fuel burn rate changes at different altitudes (but highest is not always the best) AND is dependent on the gross weight of the a/c. For example, on the way to IGARI when the a/c weighed 220,000 Kg the LRC would have been M0.825 at FL330 with a burn rate of 6826 kg/hr. At FL350, LRC would be M0.84 with a burn rate of 6724 kg/hr. Once weight go down to 180,000 kg (somewhere between the 5th and 6th arcs) at FL330 the burn rate would have dropped to 5544 kg/hr at M0.768. Instead if the FL was 410, the burn rate would be the same as at FL330 but the speed would be M0.840
To help simplify matters, the FCOM includes a “checkpoint” table where you can check that at your current altitude, to go this many miles you need this much fuel and it will take x time. A problem with this table for us is it includes the descent in the total distance where many of us believe zero fuel was reached while at cruising altitude.
One thing you can find by using all of these tables is there are not many combinations of speed and altitude that give MH370’s endurance (about 5.9 hours after 18:22). Even fewer meet the BTO arcs.
@Richard: Good points. Attempts to find “the most likely” whatever are bound to be forlorn. It is easy to define the “most likely” average Canadian. But grab any individual Canadian, it almost certainly will not be average. Still less if he is Rob Ford. We are looking for Rob Ford.
Jeff wrote: “I refer you to the work of Miriam Goldstein, who collected 242 objects floating in the Pacific and found that all of them greater than one square inch in size had identifiable marine life growing on them.
@ Jeff: “identifiable” is the operative phrase. What is identifiable to an expert biologist—armed with microscope and DNA sequencer who has the item in question right in front of her–will at least half the time not be identifiable to laymen looking only at photos posted on the internet.
@DennisW: The problem with reverse drift models is how to interpret the results. I believe the forward drift is calculated as a Markov chain, where there is a transition probability assigned with moving from one cell (square area) to another at each time step. This transition matrix is time dependent so that seasonal effects are included. Once this time dependent transition matrix is known, you can theoretically invert it at each time to calculate backwards in time.
However, it is not straightforward to interpret the backward drift results. Let’s suppose that there are four distinct starting points, and all the points after some time interval converge to single end point which could be caused, for instance, by joining streams.
So let’s do this same calculation in reverse. We start at the single end point, march backwards in time, and find that the probability is distributed between those four points. Let’s suppose the probabilities are 70%, 15%, 10%, and 5%. Can we really say there is 14 times higher probability that the debris started at the 70% position as compared to the 5% position?
@Victor: we could, but it would be a bewitchment of the illusion of technique.
“The locus of possible crash sites gets unmanageably large to search if we consider the possibility of pilot inputs along the route, before and after fuel exhaustion.”
Great. So the entire search so far has been based on the “Streetlight effect”. From the Wikipedia article of the same name:
A policeman sees a drunk man searching for something under a streetlight and asks what the drunk has lost. He says he lost his keys and they both look under the streetlight together. After a few minutes the policeman asks if he is sure he lost them here, and the drunk replies, no, and that he lost them in the park. The policeman asks why he is searching here, and the drunk replies, “this is where the light is.”
@Warren: The ATSB has admitted as much. The assumptions were chosen partly to keep the search area to a manageable size. From the ATSB report released in June 2014:
****
Note: Given the imprecise nature of the SATCOM data, it was necessary to make some assumptions regarding pilot control inputs in order to define a search area of a practical size. These assumptions were only made for the purposes of defining a search area and there is no suggestion that the investigation authority will make similar assumptions.
****
Also in the document:
***
Also allowing for the fact that a maximum glide distance of 100+ NM would result in an impractically large search area, the search team considered that it was reasonable to assume that there were no control inputs following the flame-out of the second engine. Accordingly the aircraft would descend and, as there would be some asymmetry due to uneven engine thrust/drag or external forces e.g. wind, the descent would develop into a spiral.
****
@Jeff, @Richard: excellent discussion. One quibble: anyone with a motive to plant electronic and shoreline evidence is very likely to have a strong motive to plant deep sea evidence.
I’m not arguing one way or another; I’m just pointing out that the discovery of deep sea wreckage doesn’t rule out that it was planted. If the physical evidence was planted by anyone with the means to do so, deep sea wreckage becomes, in fact, an expected future development. What would be the point of spending millions cultivating a false theory, without plotting a false payoff?
So even if deep-sea wreckage is found, we would still have two tracks – trusting, and skeptical. Each camp would have to submit their list of items that leads them to their stance.
For example, I could now list roughly 200 events, decisions, statements, and claims which draw me into the “skeptical” camp. Deep sea wreckage would only shift me if it was accompanied by disclosure of all the radar, signal data, acoustic ping, fuel model, and drift study analysis which led to such a deplorably dysfunctional search. In detail sufficient to resolve the deep paradoxes the discovery of deep sea wreckage – in ANY condition, at ANY location – would produce.