Survey conducted by @Jay (Joel Kaye) via the comments section of “Guest Post: Northern Routes and Burst Frequency Offset for MH370.”
635 thoughts on “Where Do We Think MH370 Went?”
SIO
The chart reflects the opinions of people with widely varying knowledge and expertise. It is an interesting chart, but it is not indicative of the true probabilities. Assuming that the Inmarsat data is authentic, it is 100% certain that MH370 is in the SIO somewhere close to the 7th arc. Even assuming that a pilot was alive at the end and flew the aircraft away from the arc like a glider, at the best L/D (~17) speed (~225 kts) after fuel exhaustion to maximize the distance, it could not be more than 100NM from the 7th arc. Taking into account the 0019 Doppler data (indicating high descent rate) and B777 simulations (also indicating high descent rate), it is far more likely (I would say >90%) that the plane is within 20-30 NM of the arc.
For all the other theories, such as northern paths, one must assume the Inmarsat data is not authentic (hacked, etc.), or it is not accurate (measurement errors). Given the scrutiny this question has received over the last 15 months by many experts inside and outside ATSB, I think the probability that the Inmarsat data is not authentic, or it is not accurate on the whole, is <1%.
In summary:
• Probability MH370 is in the SIO= 99%
• Probability MH370 is NOT in the SIO= 1%
• Probability MH370 is within 30NM of the 7th arc= 90%
• Probability MH370 is within 100NM of the 7th arc= 98%
The more important question is, what is the probability of finding MH370 given that it is known to be in some defined area? The ATSB search experts are very confident they will detect MH370 if they pass over the spot within ~1/2 NM. But the ship tracking data indicates it is possible they could miss the spot. It is a very challenging environment to work in. We don’t have enough quantitative information to estimate this probability, but intuitively, it would seem to be somewhere in the 90-99% range, not 100%. Thus, even if we know for sure MH370 is in the search area, there is no certainty it will be found. Still, the odds are fairly good, and we should continue to support the ATSB effort. Lacking any substantive new revelations, it remains the best bet.
Mike
Hi Mike
Thanks for the post.
Could you please go to the “MH370: Anatomy of the SIO Search” post and pinpoint on one of the figures which you think best suits your location of the a/c.
Thanks,
Chris
Chris: I believe the most likely area is IVO S37.71, E88.75 (IG point as of Sept 26). Draw a long skinny curved “box” along the 7th arc about 20NM each side and 200-300 NM along the arc. This is just an estimate for the highest probability area. Depending on true average airspeed post 1941, it could be little further up or down the arc, but that is still the highest priority box to check first. If the plane was flown manually after fuel exhaustion, it could be further from the 7th arc, but the likelihood is quite low, so positions past 7th arc+20NM should not be the priority at this time.
@airlandseaman, Perhaps I’m confused — isn’t that the area that’s already been searched?
Jeff:
Yes, that is in the general area where the search has been focused, but not all of the recommended priority box has been searched. They have not search “behind the arc as much as past the arc. And as I noted above, they could have been very close to to MH370 and still miss it.
@Jay, airlandseaman and all
Thanks Jay for your interesting and useful informal survey. It seems that out of those who have responded, several felt quite strongly that it was in the SIO (i.e. assigned >50%) and quite a few (maybe more?) are open to various possibilities or felt strongly about another option.
…(start of nit pick) with a background in maths and stats I am quite uncomfortable with the words probability, odds, 100% certain, likelihood etc. being used when these are really only educated guesses, “gut feelings” or mean results (end of nit pick)…
Assuming that the Inmarsat data are authentic and accurate, is it however possible that it is not 9m mro that is in the SIO but another plane, drone etc. which has been engineered to have its identity? Someone please expand or correct me if this is wrong.
I do think, BTW, that the continued silence from the US is the major elephant in the room.
ALSM – Would the seabed debris field be confined within 2nm? The energy of the crash, the 4 mile drop, variable sink rates, dispersal etc?
If the probability of flawed data is around the 1% level then you have dissolved the debris issue as have a lot of others, and that seems to be what the debate hinges on these days. One can either get their head around it(lack of) or they can’t? One camp says the data is the data, another will say every day without debris challenges the data. It’s now pings vs debris.
I happen to be in full agreement with Airlandseaman in the sense that even if it were a given that MH370 is somewhere along or near the 7th arc of the SIO, there truly is no guarantee that the technology searching for it will indeed find it. I believe that this notion is one that many people fail to consider when assigning low probabilities to the plane being the SIO. Airlandseaman assigned a probability of somewhere between 90-99% that the plane would be seen if it were there. My hunch tells me that it’s around 90%.
@PM- re: your comment about the US silence. I would agree that in the case of a story with international attention and interest such as this, it’s typically odd for the US to be overwhelmingly silent. On the other hand, the US did provide its input early on in this case. Furthermore, at least from my standpoint, this case does not seem to have much relevance or importance to the US, and thus I would expect a country to show little interest in something that has no impact on them.
Jay – an empty 777 is pushing 200 tonnes without diving into it, the cargo including passengers is also very substantial, it amounts to hundreds of tonnes of stuff out there that ostensibly disappeared. The seabed debris will be quite dispersed also – it’s not like they are looking for a thimble. The searchers sound supremely confident they won’t miss it. It can’t all be in a crevasse in a neat pile?
US silence – I’ve been beating this drum for over a year. Zero interest in this one when a missing 777 would normally rate. Even if the disappearance was entirely intrinsic to Malaysia as Rand would put it, they would need to have detailed info on exactly what happened to it to pull out the deck chair like this. Did that info come from Malaysia or somewhere else? “It’s in the SIO – everyone go home.” That was Obama’s only real statement and it was very early on. They are showing zero interest in the boxes, no presence in the search area at all but China do. I tend to think it’s either a domestic Malaysian episode with terrible consequences that does not rate in importance or it is indeed the one that got away and they know damn well there are no boxes down there.
How about two new sub categories:
SIO – Current Search Area
SIO – South West of Current Search Area
Why south west of Current Search Area?
1. Investigation fuel load at IGARI is incorrect.
At 1706:43 fuel load was around 43,800 kg (fuel could still be moving in tanks after level off)
Cruise fuel burn rate should be around 6,700 kg/hr for given conditions.
At IGARI (1720:31), fuel would be around 42,250 kg
Investigation estimated fuel load at IGARI as 41,500 kg (from Factual Information Report).
That equates to a burn rate of 10,000 kg/hr for cruise!
At that rate, it would need to refuel to get to Beijing.
Where did the 750 kg of fuel go?
750 kg of fuel equals another 50 nautical miles of cruise.
2. Satellite calibrations are incorrect.
BFO calibrations were conducted on the 1707 ACARS report
(Report Defining the Underwater Search Areas).
The active High Gain Antenna (HGA) at that time was the Left antenna.
If the aircraft flies South, then the active antenna is the Right.
Different antenna, different phase shift controller, leads to BFO errors!
3. Aircraft turned at NILAM, then SANOB – Banda Aceh causing a log on at 1825:27
Tracking from MEKAR to SANOB (North west), Left HGA would be assigned tracking, since right HGA is shielded by fuselage.
If Left HGA circuit breaker was popped (from an electrical fault), a log on would not occur until the right HGA could see the satellite.
If the aircraft turns at NILAM to SANOB (South west and assuming Mach 0.84), the right HGA would see the satellite around 1825:25
An automatic log on was initiated at 1825:27!
The aircraft would be turning at SANOB around 1827:50
4. Reaching Banda Aceh it maintained heading 192 degrees TRUE and crosses all arcs on time
Normally upon reaching the last waypoint, the aircraft will maintain the current heading in degrees Magnetic, but in Alternate Navigation mode (FMC failure), it will be in degrees TRUE.
A heading of 192 degrees TRUE crosses all the ping arcs on time.
In Alternate Navigation mode, speed can only be a fixed Mach Number. ECON, LRC are not available.
If it maintains Mach 0.84 until fuel exhaustion at 0017, then the aircraft must be at 40000 feet. As the aircraft cruised over Banda Aceh at 1835:45, it would have around 33,600 kg of fuel on board. Burn rate would be around 6,500 kg/hr, reducing to 5,400 kg/hr by 0017.
The extra fuel, turn at NILAM, puts the aircraft further south west than the current search area.
Into the area defined by the Constrained Autopilot Dynamics and suggested by the NTSB.
Oxy.
Jay – if the US did have all the details and it was a domestic Malaysian affair, so would a few other countries and it’s hard(for me) to imagine them all allowing Malaysia to turn this into the biggest mystery ever out of diplomatic sensitivity. They would be pressured into disclosure.
@Matty
Trying not to stray from the subject, but…”they know damn well there are no boxes down there”
What do you mean?
As for American silence. We did send an FBI team. It dominated the news for as much time that the networks could wring out of it until nothing further could be reached as far as causation. Y’all are on an American website discussing it with some of the best American aviation minds this nation could muster. As for the Gov’t. Are they to send a dignitary to hang out with the NOK?
Regarding:
2. Satellite calibrations are incorrect. This statement is incorrect. The antenna system has nothing to do with the frequency transmitted. Changing sides has no effect whatsoever on the frequency, no matter how much difference there might be in phase delays.
@ALSM
The probability of scanning the area containing the debris and not detecting it is a vital parameter to estimate. It could easily be, as was the case for detecting submarines in WWII and the wreckage of AF447, high enough to make searching areas already searched a higher priority than extending the search area. Do you know if a Bayesian search algorithm is being used?
@airlandseaman: so the “true” probability the pdf offered up to us (eventually) as the ISAT data log is ACTUALLY MH370’s data log is >99%, is it?
In arriving at this “true” number, to what degree have you weighted in zero surface debris? Zero radar? Zero acoustic detection? Zero seafloor debris? Baffling dysfunction in the search itself? Media manipulation?
The “wrong” probabilities you lament are merely alternative assessments of the ISAT data’s provenance, GIVEN other emerging data. Your attempt to imply these alternative assessments are less “knowledgable”, or less “expert” I find offensive, frankly, given the sharp disconnect between this pdf you trust and EVERY other source of evidence we’ve ever canvassed.
@DennisW
>Bayesian search algorithm
They are just searching the 60000sq.km. (now 120000sq.km.) area by trudging up and down by sectors, each scan parallel to the last scan, with an emphasis on clearing the Western area first, presumably because they judge it more difficult due to distance from Fremantle. There is no sign of any subtle algorithm.
The Fugro lecture I attended earlier this year emphasised that covering the ground with no gaps is the main aim.
@Oxy – Your fuel calculations support those posted a couple of weeks ago. Gysbreght and I both calculated remaining fuel of 35,600 kg at the time (18:22 UTC) and location (10 nm NW of MEKAR) of the last radar. GW was 210 metric tons. We believe the endurance from this point is just under 5.9h with range dependent on altitude. (The Oct. 8, ATSB report shows the MRC at 25,000ft, 30,000ft, 35,000ft, and 40,000 depending on the time of the FMT. Note, Brock has challenged the ATSB because, for unexplained reasons, they do not show the MRC for 40,000 ft coupled with an FMT of 18:28.) This gives an average fuel burn rate of 6.03 metric tons/hour, which is pretty close the middle of your 6500 kg/hr to 5,400 kg/hr range.
…indicates the average burn rate over the cruise period from 1706 to 0011 was 13,570 lbs/hr (6157kg/hr). The right engine burn rate was about 0.8% higher than the left over the cruise period, but it was 2.0% higher in the initial climb to 35,000 Feet. Fuel imbalance at first engine flame out was 630 lbs.
Chris – A missing 777 in this day and age is an issue that would normally implicate western intel services right away but if it’s down there they don’t seem terribly interested as to how or why. The 9/11 investigation was more open than this one and I zeroed in on the US(bloggers aside)because they normally have a more boisterous role than this. If Malaysia and Australia were prepared to walk away and it fell to China to keep the search for the boxes alive then the US is not that interested, along with UK, France, etc. So what do they know?
Chris – either not interested in the boxes or not interested in the SIO. People see the US as a sort of barometer. If it’s the latter then they know it got away.
@airlandseaman
You are correct.
I have since found that the source of my information to be totally wrong! Please rescind point 2 of my previous post.
If the aircraft is turning at NILAM and SANOB, as the aircraft rolls into the turn (right wing up), the right HGA would have a slight vertical component towards the satellite. Likewise, as it rolls out of the turn back to wings level, the right HGA would have a slight vertical component away from the satellite.
Would this cause a BFO error?
Would turbulence cause a BFO error?
And if the cabin was depressurised, and the cabin temperature was falling towards minus 56 degrees C, would this cause any errors?
Thanks
Oxy
@Richard
Thanks. That was my assumption relative to the search.
Matty-
Your perceptions of an American response, mixed with a 007 mentality fall short of reality.
While the IG has been a fascinating tool in the on going investigation the 007, Diego Garcia, shoot down theories only serve to hurt the NOK & fuel discontent regarding the U.S. with input such as this. If the U.S. was leading the charge, it would be meddling, if not leading the charge it’s either disinterested or hiding something. We’re damned if we do, or damned if we don’t. The Aussies have been doing an outstanding job on their side of the globe. Why get in their way?
@Chris Butler
At least every country having citizens onboard should do the most they can. Because the US has such big capabilities there are many people, including me, who feel they could do more. That having said, I’m also disappointed in for example the French, and my own (Dutch) goverment for apparent disinterest.
Niels.
@Matty
I’m not surprised by the response of the international community in general or the US in particular. Once the terrorist angle is taken off the table, and I think it is safe to do that, there is no self-interest reason to get involved.
Finding the plane is not likely to produce any new useful information. The voice recorder will most likely be blank, and the data recorder will support the fact that the plane is where it was found.
I’m sure the gurus at the NSA and CIA looked at the incident with some care (they have a lot of people to do just that), and collectively decided it did not pose a threat to US interests.
@ALSM
Mike,I am wondering how accurate you estimate the location of the 7th arc. In particular I have two related questions:
1. How does the unknown altitude of the a/c when it was pinging at 00:19:29 translate in the 7th arc radius at sea level?
2. The aircraft terminal induced fixed contribution to the BTO: What exactly in the system is causing this part of the BTO, and how can you be sure it is constant under all circumstances?
Thanks,
Niels.
Niels:
1. The “radius” difference between 35K and 0K is 5NM, so not a big factor in the overall uncertanty. 0 ft alt is closer to sub-sat point.
2. The delay inside the AES (mostly SDU) is due to the hard coded FPGA logic. It is very deterministic by design, so not subject to much jitter. This delay is well understood from experiments conducted by Thales on multiple MSC6000 terminals.
Dennis – if they(the US)took the terrorist angle off the board then they sort of know what happened? That’s my whole point.
Chris – you misinterpret what I’m saying. I don’t feel that the US owe anything at all to the search. I just think it’s indicative that they are not in amongst it. It looks as if it fell to China to keep the thing going.
Chen’s response provides some better understanding of where his team was focused and the assumptions used. But the bottom line is: The existing analysis is not consistent with various facts in evidence and the physics of the real problem. It is more of a purely theoretical analysis based on assumptions not realistic for the case of a B777 or a near vertical descent. I urge Chen et.al. to do their homework on the B777 physics, MH370 specifics and apply their math skills to the real problem at hand. That could be helpful. In particular, a water entry speed of ~130 kts might be realistic for a controlled near horizontal landing on water, but it is utterly impossible for near vertical entry. For near vertical entry, the assumed speed should be in the 400-700 kts range where catastrophic failure is certain without any new math.
Dennis – forgot to add, finding the plane is the only way to confirm it’s location? The gurus at the NSA/CIA would be well aware of the possibility for electronic shenanigans or a set up involving more than one plane?
And yet, Prof Chen re-iterates his findings and postulates
G. Chen:
“On the other hand, the lack of a large floating debris field rules out the possibility of a high velocity water entry.”
If I interprete that statement correctly, he is saying that the plane would have had to come to near rest at a very limited height, then assume a nose down attitude, before falling near vertically to the ocean.
Pigs fly?
Or maybe it was a ditch attempt gone wrong? Slow landing/ditching speed near surface, nose touches the water and “bounces” upwards, plane starts climbing, converting speed to altitude, stalls, backflips, comes falling down nose first.
Still sounds like pigs flying.
Cheers
Will
@ALSM. Thanks for your comments on the paper by Chen et. al. indicating that this paper is not useful as it stands. Is it true that papers published in Notices of the American Society of Mathematics are not peer-reviewed (while papers in their journals are)?
Becoming more cynical as time goes on in the search for MH370… why would a Maths Professor and colleagues publish something so blatantly incorrect? did someone lean on them? …won’t do their academic reputation any good.
correction – AMS American Mathematical Society.
There’s a lot wrong with the paper, and now Chen understands that in spades. But he is trying to dig out of the hole they dug for themselves and they need a hand. With the mainframe resources and math knowledge they have, they could be helpful, but engineers and scientists need to help shape the problem to be modelled. They get it now.
1. the only way to avoid a floating debris field is a vertical dive AT LOW SPEED (Chen, CLARIFIED)
2. the only way MH370 could possibly have entered the water is at HIGH speed (undisputed)
3. If MH370 hit the water, it would have produced a debris field (1 + 2)
4. If a debris field was produced, air/sat/ship/shore searches would have detected SOME of it by now* (Ebbesmeyer, drift models, West-Oz locals, common sense…and now, apparently, Chen)
5. MH370 did not hit the water (3 + 4)
Sir Tim is absolutely right when he states that “somebody knows more about this than they are prepared to say”. From the get-go, the official theory has been severely logic-challenged; lack of surface debris merely tops a very long list.
* SIO true believers devise creative ways to rationalize zero debris: drifted up-current (AMSA), pulverized, sank intact (finally euthanized by Chen paper, one hopes), caught up in gyres**. None make any sense to me. If the towfish search is anywhere near accurately located, then SOMETHING should have hit WA shores long before now.)
** you can’t use studies suggesting SOME debris might be caught in gyres to justify a belief that ALL debris became thus entangled. How many rubber duckies would get stuck in eddies, if we tossed 10,000 into a large river? Some: my word, YES; ALL: my word, NO.
@ALSM (on the arc radius accuracies)
Thanks,Mike. A follow-up on 2: How large is the contribution by AES (mainly SDU) to the fixed part of BTO? The fpga will be clocked; how would it react to large temperature shifts?
Niels.
Matty-
You’ve always been a fine contributor up & until your speculations regarding the U.S. Not to mention the hooka pipe & contact highs you cast espousing such ideas, that the U.S. either doesn’t care or in some way is involved with a cover up. I’m just as frustrated with Jeff & his Northern route theories at times to Russia & the like. It serves no purpose & gives false hopes to the NOK. Jeff actually posed (paraphrasing) “Dont’t be sure that he committed suicide” regarding Germanwings. THAT is JEFF. Proof positive & I admire this site for that. He wants proof positive results, who doesn’t?
BUT…when speculation travels into the heart of Xanadu, in regards to 370, I kinda melt down. My mind won’t permit the kind of BS that 370 was shot down, Diego Garcia, nor Jeff’s Russian theory.
Folks just don’t want to admit that it was a suicide. THAT is their opinion. That is why we’re here & why this site & Jeff leave us with our own opinions.
@airlandseaman, in your fuel calculation spreadsheet, you have in the column titled “left rate lb/hr” a value of 9223 lbs/hr and “right rate lb/hr” 9296 lb/h. This gives a total fuel rate of 18519 lb/hr (or 8400 kg/hr) for both engines. The fuel onboard at 17:06 UTC was 96562 lbs (43800 kg)
If fuel consumption is constant throughout the flight, then by your calculations, this would give an endurance of only 5 hours 13 minutes from the 17:06 UTC position and flame-out would then occur at 22:20 UTC. Any comments?
@Chris Butler
Well said. It’s the NOK that suffer. There is not one scintilla of evidence (circumstantial or otherwise) to support anything other than an aircraft flown and commandeered by a t7 pilot, destination SIO. Period.
@Jeff
This 18:25 ‘reboot’ is just a red herring. The plane was likely hand flow up until that point. The left bus was likely nipped.
It is then logical that Z would want to restore power to the bus to regain a fully functioning aircraft. The pax and crew were likely (IMO surely) deceased at this point. All threats had bee neutralized and he was out of radar/interception range…about to head south.
People keep bringing this up as though it requires some foolproof explanation. It doesn’t.
@Brock
What is your definition of “HIGH SPEED”? Unless you consider a ditching at the lowest speed possible a high speed event, your #2 is flatly wrong.
@Spencer
Can we “Completely” discount technology. I mean SH+T happens?
Can we “Completely” discount technology. I mean SH+T happens?
And by that I mean the SDU. We put up with tech shit all of the time .
@Chris Butler
No SH+T.
They (Jeff, Gerry Soajetman etc) harp on this to no end. It’s always the same retort, that it just CANNOT be explained.
The left bus explanation works perfectly well, and no one has said that this DOES NOT offer a technical explanation. The schematics on the t7 vis a vis EVERYTHING the left bus affects have been hard to come by.
Regardless of whether this is correct, what I do know is that the SDU issue was caused and ‘resolved’ by Z. As you say, people just don;t find the palatable. 🙁
Chris – my speculations regarding the US doesn’t go any further than to point out that they appear largely disengaged from the search when no-one expected that. Is there anything to it? But seems as we’ve put the manners aside the idea that it went down in one bit; and then got missed by the sonar as well, is equally problematic as anything else I’ve seen to be honest.
Chris – it’s not just the US though is it. It’s nearly everyone on the block when China have to revive the search and Australia/Malaysia want to walk off. Remember also the only way to pencil this thing in as “crashed” is to produce just one piece of it. Never been this hard though? MH370 will stay unaccounted until they do that, unless they know something? Why the mass disinterest in the search? And there is no need to cover anything up when it’s not your problem, though intel services are in the business of covering almost everything up. No-one is obliged to share military intel – that’s not a cover up.
Brock, your number 4) is the key one. Some (like you) are utterly convinced that if a surface debris field had been generated, parts of it would have been located by now. Other (like me) accept there is a good chance that this would not be the case. Not sure that this is resolvable until parts of the aircraft are discovered somewhere indicating its location and condition.
Matty, my take is that nobody knows anything that would materially affect the search for wreckage. Maybe, just maybe, the ‘official’ story is correct and the U.S. best info is that aircraft is in SIO. They may or may not know a little more about how it got there (eg ruling in or out action by particular groups or individuals based on cyber chatter or absence thereof).
The main thing that still sticks out to me as suspicious is the failure of the Malaysian authorities to release the unabridged Inmarsat signal logs. They had the perfect opportunity to do this with the MOT factual information report. Already a year ago Don Thompson noted inconsistencies between the timings of events (eg sat phone calls) in the signal logs and those reported by the authorities in their Actions Taken. Their failure to release may be an attempt to avoid further similar scrutiny, but one is left wondering whether additional unreported communication with the aircraft occurred. I hope they understand that the best way to squash such unfounded speculation is to get on with this hugely overdue data release.
Niels: The BTO delay in the AES/SDU is limited by system design constraints to <300 µsec (out of ~500,000 µsec). This spec applies to all AES regardless of manufacturer or model. The average probably varies over the life of the equipment by <1 µsec due to aging, temperature, etc. However, there is jitter due to the algorithm and that jitter was tested and characterized by ATSB at the request of the IG. The results are here:
BTW…Given the vintage of the MCS6000, the logic hardware inside the SDU may not be implemented in an FPGA (Field Programmable Gate Array). It may have been implemented in smaller discrete logic ICs or an ASIC. Regardless of the physical implementation, the delay must be <300 µsec.
@Matty
The reports associated with the AF447 search (and there are many) regarding the probability of a false negative (failure to detect) relative to the sonar search is around 0.1 (10%). If the probability of the plane NOT being in the SIO is 0.01 as some people insist, it would take at least three sonar passes for the probability of a false negative to be less than the probability of the plane not being there (assuming the false negative is stochastic). If the false negative is systematic, the plane might never be found no matter how prolonged the search.
If the 99% confidence of the SIO terminus is correct, it probably makes little sense to search for the plane. Just declare it to be there and move on. Especially since the recording devices will most likely yield nothing new.
@ALSM
The 1 us drift should be small enough, and makes sense in relation to the 300 us. Most Thales/Honeywell satcom units are specified down to -55C operating temperature. Is it also the case for the MSC6000?
Niels.
SIO
The chart reflects the opinions of people with widely varying knowledge and expertise. It is an interesting chart, but it is not indicative of the true probabilities. Assuming that the Inmarsat data is authentic, it is 100% certain that MH370 is in the SIO somewhere close to the 7th arc. Even assuming that a pilot was alive at the end and flew the aircraft away from the arc like a glider, at the best L/D (~17) speed (~225 kts) after fuel exhaustion to maximize the distance, it could not be more than 100NM from the 7th arc. Taking into account the 0019 Doppler data (indicating high descent rate) and B777 simulations (also indicating high descent rate), it is far more likely (I would say >90%) that the plane is within 20-30 NM of the arc.
For all the other theories, such as northern paths, one must assume the Inmarsat data is not authentic (hacked, etc.), or it is not accurate (measurement errors). Given the scrutiny this question has received over the last 15 months by many experts inside and outside ATSB, I think the probability that the Inmarsat data is not authentic, or it is not accurate on the whole, is <1%.
In summary:
• Probability MH370 is in the SIO= 99%
• Probability MH370 is NOT in the SIO= 1%
• Probability MH370 is within 30NM of the 7th arc= 90%
• Probability MH370 is within 100NM of the 7th arc= 98%
The more important question is, what is the probability of finding MH370 given that it is known to be in some defined area? The ATSB search experts are very confident they will detect MH370 if they pass over the spot within ~1/2 NM. But the ship tracking data indicates it is possible they could miss the spot. It is a very challenging environment to work in. We don’t have enough quantitative information to estimate this probability, but intuitively, it would seem to be somewhere in the 90-99% range, not 100%. Thus, even if we know for sure MH370 is in the search area, there is no certainty it will be found. Still, the odds are fairly good, and we should continue to support the ATSB effort. Lacking any substantive new revelations, it remains the best bet.
Mike
Hi Mike
Thanks for the post.
Could you please go to the “MH370: Anatomy of the SIO Search” post and pinpoint on one of the figures which you think best suits your location of the a/c.
Thanks,
Chris
Chris: I believe the most likely area is IVO S37.71, E88.75 (IG point as of Sept 26). Draw a long skinny curved “box” along the 7th arc about 20NM each side and 200-300 NM along the arc. This is just an estimate for the highest probability area. Depending on true average airspeed post 1941, it could be little further up or down the arc, but that is still the highest priority box to check first. If the plane was flown manually after fuel exhaustion, it could be further from the 7th arc, but the likelihood is quite low, so positions past 7th arc+20NM should not be the priority at this time.
@airlandseaman, Perhaps I’m confused — isn’t that the area that’s already been searched?
Jeff:
Yes, that is in the general area where the search has been focused, but not all of the recommended priority box has been searched. They have not search “behind the arc as much as past the arc. And as I noted above, they could have been very close to to MH370 and still miss it.
@Jay, airlandseaman and all
Thanks Jay for your interesting and useful informal survey. It seems that out of those who have responded, several felt quite strongly that it was in the SIO (i.e. assigned >50%) and quite a few (maybe more?) are open to various possibilities or felt strongly about another option.
…(start of nit pick) with a background in maths and stats I am quite uncomfortable with the words probability, odds, 100% certain, likelihood etc. being used when these are really only educated guesses, “gut feelings” or mean results (end of nit pick)…
Assuming that the Inmarsat data are authentic and accurate, is it however possible that it is not 9m mro that is in the SIO but another plane, drone etc. which has been engineered to have its identity? Someone please expand or correct me if this is wrong.
I do think, BTW, that the continued silence from the US is the major elephant in the room.
ALSM – Would the seabed debris field be confined within 2nm? The energy of the crash, the 4 mile drop, variable sink rates, dispersal etc?
If the probability of flawed data is around the 1% level then you have dissolved the debris issue as have a lot of others, and that seems to be what the debate hinges on these days. One can either get their head around it(lack of) or they can’t? One camp says the data is the data, another will say every day without debris challenges the data. It’s now pings vs debris.
I happen to be in full agreement with Airlandseaman in the sense that even if it were a given that MH370 is somewhere along or near the 7th arc of the SIO, there truly is no guarantee that the technology searching for it will indeed find it. I believe that this notion is one that many people fail to consider when assigning low probabilities to the plane being the SIO. Airlandseaman assigned a probability of somewhere between 90-99% that the plane would be seen if it were there. My hunch tells me that it’s around 90%.
@PM- re: your comment about the US silence. I would agree that in the case of a story with international attention and interest such as this, it’s typically odd for the US to be overwhelmingly silent. On the other hand, the US did provide its input early on in this case. Furthermore, at least from my standpoint, this case does not seem to have much relevance or importance to the US, and thus I would expect a country to show little interest in something that has no impact on them.
Jay – an empty 777 is pushing 200 tonnes without diving into it, the cargo including passengers is also very substantial, it amounts to hundreds of tonnes of stuff out there that ostensibly disappeared. The seabed debris will be quite dispersed also – it’s not like they are looking for a thimble. The searchers sound supremely confident they won’t miss it. It can’t all be in a crevasse in a neat pile?
US silence – I’ve been beating this drum for over a year. Zero interest in this one when a missing 777 would normally rate. Even if the disappearance was entirely intrinsic to Malaysia as Rand would put it, they would need to have detailed info on exactly what happened to it to pull out the deck chair like this. Did that info come from Malaysia or somewhere else? “It’s in the SIO – everyone go home.” That was Obama’s only real statement and it was very early on. They are showing zero interest in the boxes, no presence in the search area at all but China do. I tend to think it’s either a domestic Malaysian episode with terrible consequences that does not rate in importance or it is indeed the one that got away and they know damn well there are no boxes down there.
How about two new sub categories:
SIO – Current Search Area
SIO – South West of Current Search Area
Why south west of Current Search Area?
1. Investigation fuel load at IGARI is incorrect.
At 1706:43 fuel load was around 43,800 kg (fuel could still be moving in tanks after level off)
Cruise fuel burn rate should be around 6,700 kg/hr for given conditions.
At IGARI (1720:31), fuel would be around 42,250 kg
Investigation estimated fuel load at IGARI as 41,500 kg (from Factual Information Report).
That equates to a burn rate of 10,000 kg/hr for cruise!
At that rate, it would need to refuel to get to Beijing.
Where did the 750 kg of fuel go?
750 kg of fuel equals another 50 nautical miles of cruise.
2. Satellite calibrations are incorrect.
BFO calibrations were conducted on the 1707 ACARS report
(Report Defining the Underwater Search Areas).
The active High Gain Antenna (HGA) at that time was the Left antenna.
If the aircraft flies South, then the active antenna is the Right.
Different antenna, different phase shift controller, leads to BFO errors!
3. Aircraft turned at NILAM, then SANOB – Banda Aceh causing a log on at 1825:27
Tracking from MEKAR to SANOB (North west), Left HGA would be assigned tracking, since right HGA is shielded by fuselage.
If Left HGA circuit breaker was popped (from an electrical fault), a log on would not occur until the right HGA could see the satellite.
If the aircraft turns at NILAM to SANOB (South west and assuming Mach 0.84), the right HGA would see the satellite around 1825:25
An automatic log on was initiated at 1825:27!
The aircraft would be turning at SANOB around 1827:50
4. Reaching Banda Aceh it maintained heading 192 degrees TRUE and crosses all arcs on time
Normally upon reaching the last waypoint, the aircraft will maintain the current heading in degrees Magnetic, but in Alternate Navigation mode (FMC failure), it will be in degrees TRUE.
A heading of 192 degrees TRUE crosses all the ping arcs on time.
In Alternate Navigation mode, speed can only be a fixed Mach Number. ECON, LRC are not available.
If it maintains Mach 0.84 until fuel exhaustion at 0017, then the aircraft must be at 40000 feet. As the aircraft cruised over Banda Aceh at 1835:45, it would have around 33,600 kg of fuel on board. Burn rate would be around 6,500 kg/hr, reducing to 5,400 kg/hr by 0017.
The extra fuel, turn at NILAM, puts the aircraft further south west than the current search area.
Into the area defined by the Constrained Autopilot Dynamics and suggested by the NTSB.
Oxy.
Jay – if the US did have all the details and it was a domestic Malaysian affair, so would a few other countries and it’s hard(for me) to imagine them all allowing Malaysia to turn this into the biggest mystery ever out of diplomatic sensitivity. They would be pressured into disclosure.
@Matty
Trying not to stray from the subject, but…”they know damn well there are no boxes down there”
What do you mean?
As for American silence. We did send an FBI team. It dominated the news for as much time that the networks could wring out of it until nothing further could be reached as far as causation. Y’all are on an American website discussing it with some of the best American aviation minds this nation could muster. As for the Gov’t. Are they to send a dignitary to hang out with the NOK?
Regarding:
2. Satellite calibrations are incorrect. This statement is incorrect. The antenna system has nothing to do with the frequency transmitted. Changing sides has no effect whatsoever on the frequency, no matter how much difference there might be in phase delays.
@ALSM
The probability of scanning the area containing the debris and not detecting it is a vital parameter to estimate. It could easily be, as was the case for detecting submarines in WWII and the wreckage of AF447, high enough to make searching areas already searched a higher priority than extending the search area. Do you know if a Bayesian search algorithm is being used?
@airlandseaman: so the “true” probability the pdf offered up to us (eventually) as the ISAT data log is ACTUALLY MH370’s data log is >99%, is it?
In arriving at this “true” number, to what degree have you weighted in zero surface debris? Zero radar? Zero acoustic detection? Zero seafloor debris? Baffling dysfunction in the search itself? Media manipulation?
The “wrong” probabilities you lament are merely alternative assessments of the ISAT data’s provenance, GIVEN other emerging data. Your attempt to imply these alternative assessments are less “knowledgable”, or less “expert” I find offensive, frankly, given the sharp disconnect between this pdf you trust and EVERY other source of evidence we’ve ever canvassed.
@DennisW
>Bayesian search algorithm
They are just searching the 60000sq.km. (now 120000sq.km.) area by trudging up and down by sectors, each scan parallel to the last scan, with an emphasis on clearing the Western area first, presumably because they judge it more difficult due to distance from Fremantle. There is no sign of any subtle algorithm.
The Fugro lecture I attended earlier this year emphasised that covering the ground with no gaps is the main aim.
@Oxy – Your fuel calculations support those posted a couple of weeks ago. Gysbreght and I both calculated remaining fuel of 35,600 kg at the time (18:22 UTC) and location (10 nm NW of MEKAR) of the last radar. GW was 210 metric tons. We believe the endurance from this point is just under 5.9h with range dependent on altitude. (The Oct. 8, ATSB report shows the MRC at 25,000ft, 30,000ft, 35,000ft, and 40,000 depending on the time of the FMT. Note, Brock has challenged the ATSB because, for unexplained reasons, they do not show the MRC for 40,000 ft coupled with an FMT of 18:28.) This gives an average fuel burn rate of 6.03 metric tons/hour, which is pretty close the middle of your 6500 kg/hr to 5,400 kg/hr range.
Fuel analysis here:
https://www.dropbox.com/s/050k1p8uwdstng2/Fuel_Exner_2015-04-25.xlsx?dl=0
…indicates the average burn rate over the cruise period from 1706 to 0011 was 13,570 lbs/hr (6157kg/hr). The right engine burn rate was about 0.8% higher than the left over the cruise period, but it was 2.0% higher in the initial climb to 35,000 Feet. Fuel imbalance at first engine flame out was 630 lbs.
Chris – A missing 777 in this day and age is an issue that would normally implicate western intel services right away but if it’s down there they don’t seem terribly interested as to how or why. The 9/11 investigation was more open than this one and I zeroed in on the US(bloggers aside)because they normally have a more boisterous role than this. If Malaysia and Australia were prepared to walk away and it fell to China to keep the search for the boxes alive then the US is not that interested, along with UK, France, etc. So what do they know?
Chris – either not interested in the boxes or not interested in the SIO. People see the US as a sort of barometer. If it’s the latter then they know it got away.
@airlandseaman
You are correct.
I have since found that the source of my information to be totally wrong! Please rescind point 2 of my previous post.
If the aircraft is turning at NILAM and SANOB, as the aircraft rolls into the turn (right wing up), the right HGA would have a slight vertical component towards the satellite. Likewise, as it rolls out of the turn back to wings level, the right HGA would have a slight vertical component away from the satellite.
Would this cause a BFO error?
Would turbulence cause a BFO error?
And if the cabin was depressurised, and the cabin temperature was falling towards minus 56 degrees C, would this cause any errors?
Thanks
Oxy
@Richard
Thanks. That was my assumption relative to the search.
Matty-
Your perceptions of an American response, mixed with a 007 mentality fall short of reality.
While the IG has been a fascinating tool in the on going investigation the 007, Diego Garcia, shoot down theories only serve to hurt the NOK & fuel discontent regarding the U.S. with input such as this. If the U.S. was leading the charge, it would be meddling, if not leading the charge it’s either disinterested or hiding something. We’re damned if we do, or damned if we don’t. The Aussies have been doing an outstanding job on their side of the globe. Why get in their way?
@Chris Butler
At least every country having citizens onboard should do the most they can. Because the US has such big capabilities there are many people, including me, who feel they could do more. That having said, I’m also disappointed in for example the French, and my own (Dutch) goverment for apparent disinterest.
Niels.
@Matty
I’m not surprised by the response of the international community in general or the US in particular. Once the terrorist angle is taken off the table, and I think it is safe to do that, there is no self-interest reason to get involved.
Finding the plane is not likely to produce any new useful information. The voice recorder will most likely be blank, and the data recorder will support the fact that the plane is where it was found.
I’m sure the gurus at the NSA and CIA looked at the incident with some care (they have a lot of people to do just that), and collectively decided it did not pose a threat to US interests.
@ALSM
Mike,I am wondering how accurate you estimate the location of the 7th arc. In particular I have two related questions:
1. How does the unknown altitude of the a/c when it was pinging at 00:19:29 translate in the 7th arc radius at sea level?
2. The aircraft terminal induced fixed contribution to the BTO: What exactly in the system is causing this part of the BTO, and how can you be sure it is constant under all circumstances?
Thanks,
Niels.
Niels:
1. The “radius” difference between 35K and 0K is 5NM, so not a big factor in the overall uncertanty. 0 ft alt is closer to sub-sat point.
2. The delay inside the AES (mostly SDU) is due to the hard coded FPGA logic. It is very deterministic by design, so not subject to much jitter. This delay is well understood from experiments conducted by Thales on multiple MSC6000 terminals.
Dennis – if they(the US)took the terrorist angle off the board then they sort of know what happened? That’s my whole point.
Chris – you misinterpret what I’m saying. I don’t feel that the US owe anything at all to the search. I just think it’s indicative that they are not in amongst it. It looks as if it fell to China to keep the thing going.
Chen has responded to critics:
http://www.math.tamu.edu/~gchen/mh370.pdf
Chen’s response provides some better understanding of where his team was focused and the assumptions used. But the bottom line is: The existing analysis is not consistent with various facts in evidence and the physics of the real problem. It is more of a purely theoretical analysis based on assumptions not realistic for the case of a B777 or a near vertical descent. I urge Chen et.al. to do their homework on the B777 physics, MH370 specifics and apply their math skills to the real problem at hand. That could be helpful. In particular, a water entry speed of ~130 kts might be realistic for a controlled near horizontal landing on water, but it is utterly impossible for near vertical entry. For near vertical entry, the assumed speed should be in the 400-700 kts range where catastrophic failure is certain without any new math.
Dennis – forgot to add, finding the plane is the only way to confirm it’s location? The gurus at the NSA/CIA would be well aware of the possibility for electronic shenanigans or a set up involving more than one plane?
And yet, Prof Chen re-iterates his findings and postulates
G. Chen:
“On the other hand, the lack of a large floating debris field rules out the possibility of a high velocity water entry.”
If I interprete that statement correctly, he is saying that the plane would have had to come to near rest at a very limited height, then assume a nose down attitude, before falling near vertically to the ocean.
Pigs fly?
Or maybe it was a ditch attempt gone wrong? Slow landing/ditching speed near surface, nose touches the water and “bounces” upwards, plane starts climbing, converting speed to altitude, stalls, backflips, comes falling down nose first.
Still sounds like pigs flying.
Cheers
Will
@ALSM. Thanks for your comments on the paper by Chen et. al. indicating that this paper is not useful as it stands. Is it true that papers published in Notices of the American Society of Mathematics are not peer-reviewed (while papers in their journals are)?
Becoming more cynical as time goes on in the search for MH370… why would a Maths Professor and colleagues publish something so blatantly incorrect? did someone lean on them? …won’t do their academic reputation any good.
correction – AMS American Mathematical Society.
There’s a lot wrong with the paper, and now Chen understands that in spades. But he is trying to dig out of the hole they dug for themselves and they need a hand. With the mainframe resources and math knowledge they have, they could be helpful, but engineers and scientists need to help shape the problem to be modelled. They get it now.
1. the only way to avoid a floating debris field is a vertical dive AT LOW SPEED (Chen, CLARIFIED)
2. the only way MH370 could possibly have entered the water is at HIGH speed (undisputed)
3. If MH370 hit the water, it would have produced a debris field (1 + 2)
4. If a debris field was produced, air/sat/ship/shore searches would have detected SOME of it by now* (Ebbesmeyer, drift models, West-Oz locals, common sense…and now, apparently, Chen)
5. MH370 did not hit the water (3 + 4)
Sir Tim is absolutely right when he states that “somebody knows more about this than they are prepared to say”. From the get-go, the official theory has been severely logic-challenged; lack of surface debris merely tops a very long list.
* SIO true believers devise creative ways to rationalize zero debris: drifted up-current (AMSA), pulverized, sank intact (finally euthanized by Chen paper, one hopes), caught up in gyres**. None make any sense to me. If the towfish search is anywhere near accurately located, then SOMETHING should have hit WA shores long before now.)
** you can’t use studies suggesting SOME debris might be caught in gyres to justify a belief that ALL debris became thus entangled. How many rubber duckies would get stuck in eddies, if we tossed 10,000 into a large river? Some: my word, YES; ALL: my word, NO.
@ALSM (on the arc radius accuracies)
Thanks,Mike. A follow-up on 2: How large is the contribution by AES (mainly SDU) to the fixed part of BTO? The fpga will be clocked; how would it react to large temperature shifts?
Niels.
Matty-
You’ve always been a fine contributor up & until your speculations regarding the U.S. Not to mention the hooka pipe & contact highs you cast espousing such ideas, that the U.S. either doesn’t care or in some way is involved with a cover up. I’m just as frustrated with Jeff & his Northern route theories at times to Russia & the like. It serves no purpose & gives false hopes to the NOK. Jeff actually posed (paraphrasing) “Dont’t be sure that he committed suicide” regarding Germanwings. THAT is JEFF. Proof positive & I admire this site for that. He wants proof positive results, who doesn’t?
BUT…when speculation travels into the heart of Xanadu, in regards to 370, I kinda melt down. My mind won’t permit the kind of BS that 370 was shot down, Diego Garcia, nor Jeff’s Russian theory.
Folks just don’t want to admit that it was a suicide. THAT is their opinion. That is why we’re here & why this site & Jeff leave us with our own opinions.
@airlandseaman, in your fuel calculation spreadsheet, you have in the column titled “left rate lb/hr” a value of 9223 lbs/hr and “right rate lb/hr” 9296 lb/h. This gives a total fuel rate of 18519 lb/hr (or 8400 kg/hr) for both engines. The fuel onboard at 17:06 UTC was 96562 lbs (43800 kg)
If fuel consumption is constant throughout the flight, then by your calculations, this would give an endurance of only 5 hours 13 minutes from the 17:06 UTC position and flame-out would then occur at 22:20 UTC. Any comments?
@Chris Butler
Well said. It’s the NOK that suffer. There is not one scintilla of evidence (circumstantial or otherwise) to support anything other than an aircraft flown and commandeered by a t7 pilot, destination SIO. Period.
@Jeff
This 18:25 ‘reboot’ is just a red herring. The plane was likely hand flow up until that point. The left bus was likely nipped.
It is then logical that Z would want to restore power to the bus to regain a fully functioning aircraft. The pax and crew were likely (IMO surely) deceased at this point. All threats had bee neutralized and he was out of radar/interception range…about to head south.
People keep bringing this up as though it requires some foolproof explanation. It doesn’t.
@Brock
What is your definition of “HIGH SPEED”? Unless you consider a ditching at the lowest speed possible a high speed event, your #2 is flatly wrong.
@Spencer
Can we “Completely” discount technology. I mean SH+T happens?
Can we “Completely” discount technology. I mean SH+T happens?
And by that I mean the SDU. We put up with tech shit all of the time .
@Chris Butler
No SH+T.
They (Jeff, Gerry Soajetman etc) harp on this to no end. It’s always the same retort, that it just CANNOT be explained.
The left bus explanation works perfectly well, and no one has said that this DOES NOT offer a technical explanation. The schematics on the t7 vis a vis EVERYTHING the left bus affects have been hard to come by.
Regardless of whether this is correct, what I do know is that the SDU issue was caused and ‘resolved’ by Z. As you say, people just don;t find the palatable. 🙁
Chris – my speculations regarding the US doesn’t go any further than to point out that they appear largely disengaged from the search when no-one expected that. Is there anything to it? But seems as we’ve put the manners aside the idea that it went down in one bit; and then got missed by the sonar as well, is equally problematic as anything else I’ve seen to be honest.
Chris – it’s not just the US though is it. It’s nearly everyone on the block when China have to revive the search and Australia/Malaysia want to walk off. Remember also the only way to pencil this thing in as “crashed” is to produce just one piece of it. Never been this hard though? MH370 will stay unaccounted until they do that, unless they know something? Why the mass disinterest in the search? And there is no need to cover anything up when it’s not your problem, though intel services are in the business of covering almost everything up. No-one is obliged to share military intel – that’s not a cover up.
Brock, your number 4) is the key one. Some (like you) are utterly convinced that if a surface debris field had been generated, parts of it would have been located by now. Other (like me) accept there is a good chance that this would not be the case. Not sure that this is resolvable until parts of the aircraft are discovered somewhere indicating its location and condition.
Matty, my take is that nobody knows anything that would materially affect the search for wreckage. Maybe, just maybe, the ‘official’ story is correct and the U.S. best info is that aircraft is in SIO. They may or may not know a little more about how it got there (eg ruling in or out action by particular groups or individuals based on cyber chatter or absence thereof).
The main thing that still sticks out to me as suspicious is the failure of the Malaysian authorities to release the unabridged Inmarsat signal logs. They had the perfect opportunity to do this with the MOT factual information report. Already a year ago Don Thompson noted inconsistencies between the timings of events (eg sat phone calls) in the signal logs and those reported by the authorities in their Actions Taken. Their failure to release may be an attempt to avoid further similar scrutiny, but one is left wondering whether additional unreported communication with the aircraft occurred. I hope they understand that the best way to squash such unfounded speculation is to get on with this hugely overdue data release.
Niels: The BTO delay in the AES/SDU is limited by system design constraints to <300 µsec (out of ~500,000 µsec). This spec applies to all AES regardless of manufacturer or model. The average probably varies over the life of the equipment by <1 µsec due to aging, temperature, etc. However, there is jitter due to the algorithm and that jitter was tested and characterized by ATSB at the request of the IG. The results are here:
http://www.atsb.gov.au/media/5187038/mh370_burst_timing_offset_dec2014.pdf
BTW…Given the vintage of the MCS6000, the logic hardware inside the SDU may not be implemented in an FPGA (Field Programmable Gate Array). It may have been implemented in smaller discrete logic ICs or an ASIC. Regardless of the physical implementation, the delay must be <300 µsec.
@Matty
The reports associated with the AF447 search (and there are many) regarding the probability of a false negative (failure to detect) relative to the sonar search is around 0.1 (10%). If the probability of the plane NOT being in the SIO is 0.01 as some people insist, it would take at least three sonar passes for the probability of a false negative to be less than the probability of the plane not being there (assuming the false negative is stochastic). If the false negative is systematic, the plane might never be found no matter how prolonged the search.
If the 99% confidence of the SIO terminus is correct, it probably makes little sense to search for the plane. Just declare it to be there and move on. Especially since the recording devices will most likely yield nothing new.
@ALSM
The 1 us drift should be small enough, and makes sense in relation to the 300 us. Most Thales/Honeywell satcom units are specified down to -55C operating temperature. Is it also the case for the MSC6000?
Niels.