The total subsea search for MH370 comprised more than 240,000 km2 of seabed in the Southern Indian Ocean (SIO) along the 7th arc, which is derived from the metadata from the last transmission from MH370’s SATCOM terminal. The search of the first 120,000 km2 was managed by the Australian Transport and Safety Bureau (ATSB), and included the areas that Australia’s Defense Science and Technology Group (DSTG) deemed most likely as the Point of Impact (POI). The ATSB’s subsea search along the 7th arc extended in latitude from 39.4 S to 32.8 S, varying in width from 130 km at the southern end of the search area to 40 km at the northern end.
An additional 120,000 square kilometers of seabed was scanned by Ocean Infinity (OI) using a fleet of autonomous underwater vehicles (AUVs). OI extended the length and width of the ATSB’s search so that a full 110 km width was scanned along the 7th arc north to a latitude 31.5 S. The search area was then narrowed to a width of 84 km and extended north along the 7th arc to a latitude of 24.8 S.
Despite this unprecedented large search in the area deemed most likely to find the debris field, the search was unsuccessful. So why wasn’t MH370’s debris field identified? There are only three realistic possibilities:
- The aircraft was manually piloted after fuel exhaustion and glided beyond the area that was previous searched. Although the final BFO values suggest an increasingly high rate of descent that would certainly have resulted in an impact within kilometers of the 7th arc if there had been no further pilot inputs, there is a possibility that the pilot arrested the steep descent and transitioned into a long, efficient glide.
- The point of impact (POI) occurred along the 7th arc further south than 39.4 S or further north than 24.8 S. For instance, although the statistical match to the satellite and drift model data is not as strong, Ed Anderson has discovered an acoustical event along the 7th arc at 8.4 S that he believes is related to MH370. Meanwhile, Paul Smithson believes an impact further south than 39.4 S is within the uncertainty limits of the fuel consumption and drift models, and should not be excluded.
- The debris field lies on the seabed within the area already searched, but was not identified due to challenging terrain, low quality data, or equipment issues.
Here we address the third possibility. In particular, we again consider whether the debris field might be located in the high probability search area previously identified, which is in proximity of the last estimated position (LEP) calculated in the UGIB 2020 study. We further consider whether parts of MH370 were detected but were never fully investigated because they were not part of a larger debris field.
In the figure below, the two inner yellow lines show the approximate limits of the area searched by the vessel GO Phoenix (under contract with the ATSB), and the outer lines show the limits of the Ocean Infinity search area. Also shown in the figure are olive-green areas which represent areas that were not scanned by GO Phoenix’s towfish due to steep terrain. The outlines of these and other areas of missing or low-quality data were made available by Geoscience Australia.
There is a steep slope to the south of the LEP, and the portion of the steep slope that was not scanned by the GO Phoenix towfish is about 60.3 km2. Of this, about half was later scanned by Ocean Infinity AUVs, leaving about 30.5 km2 of seabed surrounding S34.53° E93.84° that was never scanned. We designated this area as a “High Priority Search Area”, and it may be here that the debris field lies.
The subsea search for aircraft wreckage that many deem most similar to the search for MH370 was the search for Air France 447 (AF447), which was an Airbus A330 that crashed off the coast of Brazil in June 2009 in around 3000 m (9,800 ft) of water. Floating remnants of the aircraft were found within 2 days of the crash, but the subsea search was not successful in locating the debris field until April 2011, about 2 years after the crash. The sonar image from the debris field, which measured around 200 m x 600 m, is shown below.
AF447 is believed to have impacted the ocean surface without breaking up in flight and with a nose-up attitude. As such, the debris field that AF447 generated may be significantly different from the debris field created by the impact of MH370, as the final two BFO values suggest a high downward acceleration of 0.7g, and descent rates greater than 15,000 fpm. Without pilot intervention, MH370 possibly entered the water at a descent angle greater than 45 deg and at an airspeed approaching or exceeding Mach 1.
The debris from MH370 may more closely resemble the debris from SilkAir 185 rather than the debris from AF447. SilkAir 185 was a Boeing 737 that crashed into the Musi River near Palembang, Sumatra, Indonesia in December 1997. The aircraft experienced a rapid, nearly vertical dive that the US NTSB attributed to control inputs from the captain. During the high speed descent, parts of the control surfaces, including a large portion of the tail section, separated from the fuselage due to the high aerodynamic forces from the high speeds. The airspeed of the fuselage before impact is believed to have exceeded Mach 1.
SilkAir 185’s debris was found in two areas: the main debris field of around 60 m x 80 m at the bottom of the Musi River, which was only 8 m (26 ft) deep; and other larger debris, mainly flight control parts that separated before impact, that were widely scattered on land no closer than 700 m (2,300 ft) from the main debris field. According to the accident report, due to the high energy of the impact, the parts recovered from the river were “highly fragmented and mangled on impact” which made identification difficult.
If MH370 experienced the rapid descent suggested by the final BFO values, then it is probable that the fuselage broke apart before impact, and also probable that many large parts would be found outside of the main debris field. The flaperon recovered on Reunion Island is a good example of a flight control part that may have separated before impact. We would also expect the main debris field to be smaller in extent than for AF447, and within that debris field, the debris to be smaller and more difficult to identify. For instance, for the case of SilkAir 185, the landing gear was identified only by its subcomponents (struts, landing gear door actuators, wheels, brakes, tire pieces, etc.). This counters conventional wisdom that says that aircraft engines and landing gear should be among the easiest parts to identify by sonar on the seafloor, as it was the case for AF447.
The subsea search for MH370 was focused on finding the main debris field at the expense of identifying other parts that may have separated. For the search phase conducted by GO Phoenix, reports were written for a total of 45 “contacts” (observable features in images) that merited a further review. All the contact reports are compiled here. Of these 45 contacts, 24 contacts were within 100 km of the LEP, 10 contacts were within 50 km of the LEP, and 4 contacts were less than 25 km from the LEP. The locations of the contacts are shown in the figure below.
Of the 45 contacts, 11 (GP-002, 016, 018, 019, 021, 025, 026, 028, 030, 031, 047) were described in the reports with phrases like possibly “man-made”, “not geological”, or “not of natural origin”, and one (GP-046) was considered for further investigation with an AUV, which seems to have never been done. Of course, many of the man-made objects on the seafloor could be marine debris from sea vessels unrelated to MH370.
Andy Sherrill is an experienced ocean engineer who has conducted deep water search and salvage operations for a number of missions. He was a key member of the team that reviewed the sonar data for the subsea searches for MH370 that were conducted by the ATSB and Ocean Infinity. Andy was also part of team that identified the debris field for AF447 off the coast of Brazil as well as part of the team that found Argentina’s ARA San Juan submarine. Andy graciously offered these comments as to why many of the MH370 promising contacts were never investigated further:
“Typically, if there were small isolated objects that appeared to be man-made and marked as a target, but nothing else was of interest within several kilometers then we did not investigate further.
We certainly took into account if the debris field did not look like AF447 or any others, however there still needed to be enough debris to be at least a fair amount of the aircraft to warrant further investigation.
Sure a small part of the plane could have drifted and sunk, but we were looking for the main field. A decision was made to focus on finding the main field of debris, not just one small piece – and likely all of those “potentially man made” contacts are from passing vessels given there was no associated debris within several kms.
Having said that, there is always a chance it [a tagged contact] could be from MH370, but based on our assessment the time it took to investigate each of these small contacts was not worth taking vs searching new areas.”
Discussion
As the final BFO values, the lack of IFE log-on, and the end-of-flight simulations all suggest a high speed impact close to the 7th arc, a high priority should be to completely scan the areas closest to the 7th arc. MH370’s debris field may be smaller in area, consist of smaller parts, and be much more difficult to identify than searchers were anticipating. It’s also possible that the debris field is located in an area that was not fully searched due to challenging terrain, low quality data, or equipment issues, such as the steep slope identified above as the high priority search area due south of the LEP. As such, the investigation of many of the contacts previously identified becomes more important, as one or more of these contacts could be parts of MH370 that separated before impact. It’s also possible that one or more contacts are part of a less conspicuous debris field.
We again acknowledge that with pilot inputs, it is possible that MH370 glided after fuel exhaustion beyond the areas that were previously scanned. Therefore, searching wider along the 7th arc should also be part of the search plan if pursuing areas close to the 7th arc is unsuccessful in locating any of MH370’s wreckage.
Update on Nov 3, 2023
Andy Sherrill offered these additional comments:
“We did get rerun over GP16, and collected some higher frequency AUV SSS on that one. Looks highly likely to be geologic in my opinion.
We did not reacquire any more data over GP46, however that one looks very similar to GP16 and I would still classify it as highly likely to be geologic.”
Tags: AF447, debris field, MH370, SilkAir 185, sonar, subsea
@VictorI. Your bottom diagram indicates that there were 46 “man-made”, “not geological”, or “not of natural origin” objects detected by GO Phoenix, yet none in the adjacent OI area. In particular the sharp presence/absence at the go Phoenix north eastern boundary is striking.
This absence of ‘overlap’ between the two areas suggests either that OI has not disclosed such detections or those it detected were confirmed to be of no interest.
If the latter, that would weaken the chances that the GO Phoenix detections could prove of interest. If on the other hand any were left uninvestigated by OI on the same grounds as GO Phoenix, it would be of interest if they would disclose where these were.
Also, if in fact none such were detected by OI near the GO Phoenix area perhaps the two had different sensor sensitivities or detection criteria?
I am convinced that someone was at the control at the end of the flight as all data show that the aircraft was piloted during the 1st hour of the disappearance. Why should not be it the case at the end?
The image with all detected debris is interesting.
But I find your end point too close from the 7th arc. In my opinion, a huge wreckage as a Boeing 777 has highly reflective parts especially engine and gear parts. And more likely with a field debris. So the detection of it should have been highly probable with a not so high resolution.
For me, we have to search further south as the aircraft was glided not so far from the already searched area. That’s what I have suggested for years and this is what suggests the latest study of Patrick Blelly and Jean-Luc Marchand. And I agree with their analysis.
@David: 1) Only 11 of the 45 contacts were described as man-made. Many were described as geological, and for many, the description did not specify whether the investigators believed the objects were man-made or natural. 2) We don’t have contact reports from OI, which is why there are no contact locations in the areas searched by OI.
@Gilles Diharce: We don’t know if there was a controlled glide after fuel exhaustion, despite the strong beliefs that some have claiming with certainty that it did or did not occur. The BFO values at the final log-on suggest a progressively steep descent, and the shattered parts from the cabin interior suggest a high speed impact. If there was a glide, it would mean a dive, a glide, and a final dive. It’s not impossible that this occurred, but it is far from certain that it did. (You might believe the flaperon was damaged during a controlled ditching, but even in that scenario, there was a dive before the controlled glide.)
You believe that it is unlikely that MH370 was missed during the subsea search, despite the data holidays and man-made contacts that were identified but not investigated further. I suggest you read a previous blog post describing OI’s search for ARA San Juan, where the submarine was missed during the first pass.
Again, I am completely in favor of searching wider than what was previously searched to cover the possibility of a long glide after fuel exhaustion. However, I would advocate doing this after completing the search closer to the 7th arc, as there is much less area to cover, and there are some interesting areas and contacts to further investigate.
Andy Sherrill offered these additional comments:
“We did get rerun over GP16, and collected some higher frequency AUV SSS on that one. Looks highly likely to be geologic in my opinion.
We did not reacquire any more data over GP46, however that one looks very similar to GP16 and I would still classify it as highly likely to be geologic.”
From the Inspector General of the US Dept of Transportation:
California Pilot Indicted for Interfering With a Flight Crew
On October 18, 2023, a grand jury in the U.S. District Court for the District of Utah indicted Jonathan J. Dunn for interfering with the crew of a commercial airline flight.
The indictment alleges that, on or about August 22, 2022, Dunn, the crew’s First Officer, interfered with the performance of a crew member by using a dangerous weapon to assault and intimidate the Captain. Dunn was authorized to carry a firearm through the Transportation Security Administration’s Federal Flight Deck Officer program. After a disagreement about a potential flight diversion due to a passenger medical event, Dunn told the Captain they would be shot multiple times if the Captain diverted the flight.
DOT-OIG is conducting this investigation with the Federal Bureau of Investigation with substantial assistance from FAA.
https://www.oig.dot.gov/library-item/39680
@Victor
@Gilles
I am out-voted 99-to-1 but I do not feel an active pilot flew straight to 34-38s with little or no maneuvers just to run out of fuel at Arc7 at high altitude and manage that situation. I do believe Active Pilot: the BFO at Arc7 is very indicative of active pilot making a descent. We are not expecting a ghost flight to dive, and especially not to fly level for 2-minute period and then dive. We seem to be systematically ruling out slow down after Arc5 due to straight flight advocacy. Also everyone is ruling out fuel remaining at Arc7, and that may be wrong assumption too.
What’s coming in a podcast soon probably is Jeff Wise is going to say conspiracy theorists are correct because $300Million of searches failed to find MH370 as a ghost flight ending near Arc7. I am going to say — MH370 is still in the SIO — but the basic assumptions (straight flight to fuel exhaust at Arc7) are probably wrong.
There’s maybe a mistake in this sentence: ”Of these 45 contacts, 24 contacts were within 100 km of the LEP, 10 contacts were within 25 km of the LEP, and 4 contacts were less than 25 km from the LEP.”
Does the linked report include all of Go Phoenix’s contacts? If it does, I find it surprising that the total number is this low and includes only one(?) level 2 contact while the rest are level 3. Maybe the standards of classification were different between Phoenix and Fugro.
Relevant section from the 2017 ATSB report:
”Sonar contacts (anomalous features) identified in the sonar data were classified in three ways: level 3 contacts were marked but assessed as unlikely to be related to the aircraft, level 2 contacts were marked but assessed as only possibly being related to the aircraft, and level 1 contacts were of high interest and warranted immediate further investigation. There were 618 level 3 contacts, 41 level 2 contacts, and two level 1 contacts identified and reported. The two level 1 contacts were investigated and found to be iron and coal remains of a wooden shipwreck and the other was a scattered rock field. In total, four shipwrecks were found. Throughout the search 82 separate sonar contacts were investigated and eliminated (as being related to MH370) by the AUV, ROV, or deep tow vehicles.”
This link has all the contact reports for the ATSB subsea search:
https://dapds00.nci.org.au/thredds/catalog/iy57/documents/contact_reports/catalog.html
The link I provided in the article is a compilation of only the GO Phoenix contacts. There are many more Fugro contacts.
@Victor
I am more cautious about the last BFO. It seems showing a steep dive but this doesn’t mean that nobody was piloting the airplane at the end. It is not so far certain that this occurred too. It depends on what hypothesis you consider and I agree that we can have a different opinion on this aspect.
Of course, some interesting points may need to be investigated further. But I think it is unlikely as I am confident on professionals who worked on it to consider if it is relevant or not.
Despite our different analysis of this last moment of the MH370, we all have the same goal: find the plane to expect solving this mystery and terminate all conspiracy theories on it.
This is not a course for a winner.
We want the truth for the families.
@Gilles said: I am more cautious about the last BFO. It seems showing a steep dive but this doesn’t mean that nobody was piloting the airplane at the end.
I never said that the final BFO values prove the aircraft was unpiloted. There are two ways to achieve the downward acceleration of 0.7 g:
1) A nose-down input from a pilot
2) A banked, steep descent with no pilot inputs
However, if there were pilot inputs and the aircraft is beyond the area that was already searched, it would mean there was a dive-glide-dive sequence. It’s possible, but that’s the sequence that would have occurred.
We want the truth for the families.
Of course we do. There is not a person that contributes to this blog that is not in pursuit of the truth.
I will once again say that I fully acknowledge that it is possible there was a long glide. I am only saying that the next search should finish the job of searching near the arc before extending wider. Is that really controversial? Why is there pushback?
@victor @Don @Gilles
Thanks for the new post.
We think you are right to raise Point 1 including a glide and a recovery of the fast rate of descent. It is an essential one. We investigated and described this as Descent Scenario #2 in Appendix 1 of our March report (https://www.mh370-caption.net/wp-content/uploads/3-known-trajectory-and-recalculated -trajectory.pdf) and also explained it to Don (and others 🙂 at the RAeS conference in London a few weeks ago (https://www.youtube.com/watch?v=CjjySxoo_AQ ).
The other Descent scenario 1 also includes a glide.
From these one can deduce the shortest potential distance from Arc7 to the POI which is between ~42 Nm and ~67 Nm (from FL300). We therefore think that the strip of sea to look further than Arc 7 is wider than the few kilometers that you propose. It would be even wider if started at a higher flight level.
Videos of the respective simulation sessions for each scenario are available here: https://youtu.be/4gdZAFg7wJI and https://youtu.be/x7ezFNFca-A
Sherrill’s comment:
”but based on our assessment the time it took to investigate each of these small contacts was not worth taking vs searching new areas.”
In general, this argument is still just as valid as it was then. And it could even be argued that, as search technology has improved (lower cost per km²), the cost of checking a contact (in terms of wasted km²) may have gone up. However, if certain latitudes can be prioritized based on probability, it may make more sense to check the contacts in those areas.
So when you say ”a high priority should be to completely scan the areas closest to the 7th arc”, do you mean the whole previously searched 7th arc or some more prioritized latitudes (e.g. S33-36)? And by ”completely scan”, do you mean checking contacts and previously missed areas, not researching already searched areas?
@Victor
The more important thing is to obtain the continuation of the search.
If the official accepts searching on both areas (yours and Blelly/Marchand’s), I think we have a good chance to solve this affair.
But we have to be aware that some points of this disappearance will still remain unexplained. We will never know what really happened in the cockpit at the time of the disappearance as it would have been erased in the CVR unfortunately.
🤞🤞🤞🤞
@Jean-Luc Marchand: We therefore think that the strip of sea to look further than Arc 7 is wider than the few kilometers that you propose.
Where did I say the search should be constrained to a few kilometers from the 7th arc? With a long glide from a high altitude, the glide could have been 140 NM! The search area we proposed in the last post included the possibility of a long glide (Zone 3).
I will yet once again say that I fully acknowledge that it is possible there was a long glide. I am only saying that the next search should finish the job of searching near the arc before extending wider. Is that really controversial? Why is there pushback?
@eukaryote234: I wouldn’t suggest researching everything already searched. As I said in the discussion section of the post:
It’s also possible that the debris field is located in an area that was not fully searched due to challenging terrain, low quality data, or equipment issues, such as the steep slope identified above as the high priority search area due south of the LEP. As such, the investigation of many of the contacts previously identified becomes more important, as one or more of these contacts could be parts of MH370 that separated before impact. It’s also possible that one or more contacts are part of a less conspicuous debris field.
So basically, search data holidays (including challenging terrain not searched) and promising contacts. I would prioritize the GO Phoenix portion of the search because that’s where the statistical match to the data is best (i.e., UGIB 2020).
Then search wide, which is a much larger area and will take a lot more time.
@Gilles: Doesn’t the area we recommended to search in the last post include the area that Marchand/Blelly are now recommending to search?
One thing that Ocean Infinity learned from the ARA San Juan search was that you ignore canyons at your peril. It seems empirically obvious – as undersea currents drag things, they’ll get caught somewhere, most likely within a geological feature. The GO Phoenix and Fugro searches of the original area scanned the high-priority area in the same way that Ocean Infinity scanned the San Juan high-priority area, not pausing to investigate areas where the debris field might well have got stuck.
We have lots of evidence that suggests the wreckage is quite likely pretty close to the arc and pretty close to the LEP. Surely what makes the most sense is to search the areas in that zone we know we haven’t looked? It’s not just that cliff, there are other smaller areas like the subsea canyons that haven’t been carefully investigated.
@Rob Moss: I agree 100%. I don’t understand the pushback to completing the search close to the 7th arc before expanding it wide.
FWIW I also agree (with Rob Moss and Victor last 2 posts).
Regarding Rob’s comment:
« One thing that Ocean Infinity learned from the ARA San Juan search was that you ignore canyons at your peril. It seems empirically obvious – as undersea currents drag things, they’ll get caught somewhere, most likely within a geological feature. »
I was just about to put this very same questions to the subsea experts here:
Are there typically currents around canyons which would pull sinking objects (such as MH370 debris) into canyons so that there is a HIGHER probability of the debris ending up in a canyon (than around it)? Or are these canyons rather stationary zones without much water movement from/into them?
In the former case, Rob would be right that the chance of finding the debris in data holidays due to difficult terrain would be higher than usual.
@VictorI:
Thanks for including my 370Location.org candidate site into your recent paper. I continue to research the MH370 acoustics every day, and have had some recent progress in detecting weaker events using spectral whitening and phase alignment methods applied to beamforming.
No pushback from me on searching targets near the 7th Arc. I suggest resuming the search wherever a narrow track can be followed, consistent with the 7th Arc. That includes unpiloted tangents that some have bet their houses on, and also untriangulated hydrophone bearings from Cape Leeuwin that could be mapped when returning to Perth from a sortie.
My concern about the current paper is that a focus on isolated small sonar targets is pegged to the theory that these pieces of debris would have separated before impact due to “flutter” from a high speed dive, and thus not part of the expected debris field.
I believe the flutter theory was first proposed within a day of the Flaperon being found, to explain the trailing edge damage (and to be consistent with BFO).
Additional found debris adjacent to the Flaperon indicates that the flaps were retracted, which for some has confirmed an unpiloted high speed dive.
The paper compares to AF447 as a slow impact to emphasize how MH370 debris might be highly fragmented due to a supersonic impact.
Let’s look a bit closer. AF447 was descending from high altitude in a 110% full thrust nose-up stall at 108 knots ground velocity and 108 knots downward velocity for vector impact of 152 knots.
The BFO approximation of vertical speed near the last ping was 15,000 fpm or 148 knots. We don’t know the forward velocity. You mention that glide-dive-glide is a possibility. A steep dive has been ruled out when my analysis showed that the last sat pings were at normal signal strength, showing that the attitude of MH370 was not outside the ability of the SATCOM unit to optimally steer the phased array antenna as in normal operation. (No vertical dive, steep bank, or inverted flight).
Trailing edge damage appears very similar on all the debris items found – flaperon, ailerons, horizontal stabilizer. The supersonic flutter theory requires that all these pieces oscillated so violently that they lost their trailing edges in flight and then further detached in unison from the airframe so that they were *not* damaged on impact.
I have not been able to find any examples of violent trailing edge flutter in large aircraft. What I’ve found is flapping and violent twisting of the wings along the chord (wingspan), which didn’t cause all the trailing edge airfoil parts to spontaneously eject.
So, let’s consider the possibility that MH370 was nose-up at lower speed when it hit the water to account for the flaps-up trailing edge damage. Rupture of the fuselage forward of the wings is not inconsistent with the interior debris that has been found.
@Peter Norton:
I don’t expect that there will be significant currents near the seafloor, which includes canyons. Surface currents are driven by winds, and deeper currents by thermal differences. At the abyssal seabed, it is considered quiet – although researchers have documented “storms” where currents approach 0.5 m/sec. Compare that to a max 2 m/s surface current around the 7th Arc.
Now consider the descent rate of debris. It is dependent on the density of an object, including entrained air and water. From the surface, a large fuselage section could sink with captured air pockets. As it sinks, any air would be compressed to an insignificant amount, and the water contained within the structure would add to its mass. I believe this is why the acoustic recordings from a dozen hydrophones and some 45 seismometers were able to pick up the seabed impact of MH370.
If we assume a 4 m/s average velocity for sinking debris, that’s about 15 min to descend to 3400m, the depth at the epicenter of the recordings. It also dwarfs any subsurface currents during the descent.
I expect the seabed debris pattern will be very similar to that of Air France 447. I don’t anticipate isolated fragments will be found far from the crash site.
@370Location
I think you have been rather selective in the data that you use to dismiss a steep dive and flutter.
There is ample evidence of the aircraft speed immediately before, and about the time of the second engine flameout. That might be extrapolated to provide an estimate of forward velocity. Then at that speed the descent angle can be estimated to fit with the 15,000fpm descent.
I was one of the early proponents of the possibility of flutter causing the flaperon damage.
There are different types of flutter, and the type of twisting and oscillating that you refer to is not the same as the aerodynamic flutter that the flaperon may have experienced. Aerodynamic flutter does not require supersonic speeds. It can occur at very moderate speeds in all types of aircraft. The Boeing manual describes the possibility of flaperon flutter occurring during engine testing with the aircraft parked on the ground.
The TE damage on the flaperon shows separation along the line of skin fasteners, i.e. along the weakest part of the TE skin structure. Exactly what one would expect with violent aerodynamic flutter, where the magnitude and frequency of the flutter rises to a peak within a couple of seconds.
@Brian Anderson:
Thanks for the info. It’s inspired some fascinating reading about historic aeroelastic stability.
The case of parked flaperon flutter is interesting. I found that turbofan exhaust velocity on the ground can exceed 325 kt, and is likely very turbulent. It also can cause flutter on new B777 horizontal stabilizers:
https://youtu.be/EKeCZ2nDkDU?si=qWO6TFHWtb61YPkK&t=134
We do have a recent example from the China Eastern MU5735 crash and debris photos. Some granular data is here:
https://www.flightradar24.com/blog/china-eastern-airlines-flight-5735-crashes-en-route-to-guangzhou/
The B737 was cruising 455 kt at 29,100 ft when it went into a sudden vertical dive. I don’t know if it was a powered dive. It initially lost airspeed at the beginning of the dive. It reached a peak airspeed of 590 kt as it was pulling up to level out at 8000 ft. Vertical speed also peaked at -30,976 fpm, but that same value is oddly at three different times (it may be a capped reading so the actual vertical speed could have been greater). Debris photos are sparse, but we can see at least a piece of the rudder and wingtip. Neither appear to have any trailing edge damage.
I’m not entirely dismissing flutter and detachment of most control surfaces, just noting that the re-search recommendation in this paper depends on it.
@370Location: China Air 006 (B777) and SilkAir 185 (B737) both loss parts of the tail control surfaces due to flutter at high aerodynamic speeds. In the case of Silk Air 185, the plane broke apart, with parts that separated during the descent found scattered some distance from the main debris field. For China Air 006, the pilots were able to recover and eventually land the plane.
Link to image showing damage to China Air 006:
https://upload.wikimedia.org/wikipedia/commons/8/83/Damaged_empennage_of_China_Airlines_Flight_006-N4522V.JPG
@370Location:
thank you for your reply concerning underwater currents.
@”We did not reacquire any more data over GP46, however that one looks very similar to GP16 and I would still classify it as highly likely to be geologic.”
To my eyes, GP46 also doesn’t resemble any part of a plane.
@Victor Iannello: “a dive-glide-dive sequence”
Why would a pilot interrupt a dive only to plunge into a dive again ?
Or did you mean a “glide-dive-glide” sequence ?
It would seem more plausible to me that a pilot terrified by the dive and shying away from the suicide would pull out of the dive and continue gliding (and therefore avoiding dying) as long as possible.
On the other hand, a dive-glide-dive sequence could be the result of a pilot aborting the suicide dive, gathering up his courage (gliding) and then carrying out the suicide dive on the second go.
So I guess, both scenarios are imaginable and can’t be excluded.
The 7-hours-long suicide run never made much sense to me until I recently listened again to William Langewiesche‘s interview by Megyn Kelly:
https://youtu.be/v_y0A2cheo4
Although I take issue with a lot of things he says, I think he offers a plausible explanation for hours-long suicide flight:
He couldn’t bring himself to do it, but couldn’t turn back after having killed everyone on board. So he essentially had nowhere to go.
https://youtu.be/v_y0A2cheo4?t=1231
But why then does the flight path match the simulator data weeks earlier ?
That’s left unexplained and doesn’t quite match Langewiesche’s theory.
Your thoughts ?
@VictorI:
Thanks for the followup. One crash due to a suicidal pilot, and the other a crew losing control in whiteout after an engine failure, but eventually recovering.
It’s impressive that the CAL006 B747 could still fly level after the tips of its horizontal stabilizer broke off due to flutter. It appears to have been the flapping/twisting type, as there is no sign of damage along the trailing edge besides the missing sections.
I couldn’t find any photos of intact SilkAir185 trailing edges, but it’s clear that much of the tail broke off. No dispute that aeroelastic flutter is a very real effect.
@Peter Norton:
I try to be objective, but my reply was a bit focused on my own work. Seabed currents could elongate the debris field, but the AF447 debris field orientation did align with the track at impact.
@Peter Norton asked: Why would a pilot interrupt a dive only to plunge into a dive again ? Or did you mean a “glide-dive-glide” sequence ?
That sequence was suggested without regard to pilot intentions. The first dive is based on the final BFO values. The glide is based on a potential impact site beyond what was searched. The final dive is based on the shattered parts of MH370 that were recovered, which suggest a high-energy impact. If that sequence actually occurred, your guess is a good as mine as to why.
@Peter Norton
Peter, you asked, “But why then does the flight path match the simulator data weeks earlier ?”
The short answer is that the flight path does not match the simulator data (unless you have an extraordinarily loose definition for the word “match”).
Suffice to say that the sim data is easily one of the most misrepresented and consequently misunderstood items relating to the disappearance. Leaving aside the blazingly obvious discrepancies such as the simulator flight turning left towards the north-west after take-off and MH370 turning north-east, the only real similarity between the sim flight and MH370 is that both ended up in the Southern Indian Ocean.
In the recovered sim session, the Captain was almost certainly simulating the take-off and departure of his then upcoming flight to Jeddah on MH150. This was a route that he hadn’t flown for some three months. The data shows that after what appears to have been a fairly standard instrument departure, the Captain appears to have simulated some sort of in-flight upset that was then followed by his turning the aircraft around, either heading back to his departure airport, Kuala Lumpur, or possibly to the then nearest suitable airport, Medan, Indonesia. This diversion occurs before the aircraft reaches waypoint VAMPI. That first “phase” of the sim session ends there.
After that, he relocated the simulation aircraft to eastern edge of the Bay of Bengal, roughly equidistant from Port Blair and Car Nicobar. He had flown through that region on the MH17 return flight from Amsterdam to Kuala Lumpur just two weeks prior. From the Bay of Bengal location the aircraft is almost certainly not flown south as many portray it, rather the aircraft was almost certainly flown south-east back towards Indonesia/Malaysia.
During this phase the simulator aircraft was almost certainly jettisoning fuel, such that it reached fuel exhaustion somewhere over the south-western extent of the Andaman Sea, near Banda Aceh, Indonesia. The aircraft began a brief unpowered descent after fuel exhaustion had occurred. The second “phase” of the sim session ended at this point, with the aircraft having reached fuel exhaustion somewhere near north-western Sumatra.
After that, the simulation aircraft, already having experienced fuel exhaustion, was relocated to the Southern Indian Ocean about 940 nm south-west of Perth, Australia. Shortly after that relocation, after another brief period of unpowered descent, the aircraft’s altitude was manually adjusted down from its then current 37,650 feet to 4,000 feet.
It is worth noting that the simulation aircraft was a Boeing 777-200LR, not a -200ER as flown by Malaysia Airlines. The LR has a slightly different wing to, and very much more powerful engines than the -200ER so while the simulator offers reasonable fidelity procedurally, it is offers very limited utility for say, fuel planning or flight performance.
@Peter Norton
My 2 cents, I would advise Langewiesche that the pilot may have had every intent to hide the crash site. The home sim data seems to represent an outline the plan (for MH150). After Arc5 there may have been maneuvers (not straight flight). The reason dive-glide-dive scenario seems awkward is that our assumption that an active pilot agreed with us the he should not touch any controls until he ran out of fuel at FL350+ is questionable.
@TBill
Do you sense increased “push-back” recently ?
My two bob’s worth.
The view (long held in some circles with a fervent religious conviction) that the 7th Arc 8-sec BFO’s are an unarguable slam-dunk indication of a rapid and accelerating unrecoverable descent, leading to imminent and certain unscheduled instantaneous disassembly, is (to my eyes) not supported by the published descent profile of Egyptair 990 in this NTSB graphic.
https://upload.wikimedia.org/wikipedia/commons/3/34/Msr990-ntsb-f1.jpeg
It appears 990’s ROD was comparable to, and indeed greater than, 370’s, and note, it recovered from the dive, and actually climbed, before (according to the radar returns) finally crashing.
I feel that those in some circles, although apparently conceding the remote possibility of an active pilot, still can’t really accept any serious consideration of any active pilot scenario, and are only paying lip service to it, because they have far too much capital invested in the history of the unresponsive pilot scenario.
I am far more inclined to accept the French Captain’s recently presented (at the Royal Aero Society in London) descent profile(s). They are credible, and fit what an active pilot could do, and they allow glides far, far away, from the 7th Arc.
Specifically, they credibly allow for a deliberate intent by an active pilot, to carefully and skillfully deposit his aircraft in either of our two chosen “crevices” (or elsewhere).
I think both crevices should be searched.
@Victor Iannello:
Thank you for the clarification. You are probably right that it might be better to disregard psychological aspects and just focus on the data at hand.
@370Location:
Thanks for the follow-up. According to these 2 sources,
“Submarine canyons are hotspots for litter accumulation”:
https://archive.is/VRFWE#selection-3527.0-3527.54
https://www.mbari.org/news/mbari-research-shows-where-trash-accumulates-in-the-deep-sea/
The question is, how much airplane debris would would be affected.
These studies seem to somewhat validate Rob’s comment:
« One thing that Ocean Infinity learned from the ARA San Juan search was that you ignore canyons at your peril. It seems empirically obvious – as undersea currents drag things, they’ll get caught somewhere, most likely within a geological feature. »
My hunch is that the probability of finding the debris in a canyon could be higher.
@Mick Gilbert:
Thanks for your detailed recount of the simulator data, which apparently may be less incriminating than generally believed.
@Ventus
The recent report by Capt Blelly/Jean Luc Marchand (and formerly Captio) have always been most welcome from the point of view of giving thought to the active pilot scenario. However, I do not currently feel they have yet nailed the correct location and/or scenario/plan.
“Do you sense increased “push-back” recently?” since about Feb_2019, but don’t get me started.
@ventus45 said: The view (long held in some circles with a fervent religious conviction) that the 7th Arc 8-sec BFO’s are an unarguable slam-dunk indication of a rapid and accelerating unrecoverable descent, leading to imminent and certain unscheduled instantaneous disassembly, is (to my eyes) not supported by the published descent profile of Egyptair 990 in this NTSB graphic.
I’m not sure who you are referring to. My recommendation is to finish the search close to the 7th arc before searching wide. The wide search is based on a 140 NM controlled glide.
Of course a high speed descent can be arrested. I estimate that around 2,000 feet would be lost in the nose-down descent to 15,000 fpm followed by a recovery. Who thinks otherwise?
@All,
1. 9M-MRO flew to the 7th Arc and ran out of fuel shortly before that event circa 00:17:30. This combination of events is extremely difficult to achieve in a B777-200ER, knowing what we know about the aircraft’s location and speed up to 18:22:12 (the last radar contact).
2. The difficulty is in the fuel consumption. That aircraft simply cannot cruise that far in that elapsed time unless significantly abnormal fuel savings occurred. The required fuel savings compared to the normal cruise configuration amounts to the equivalent of 1,068 kg at 19:41 (out of 27,113 kg necessary in the normal configuration with air packs ON). See Case 2 in Table D-1 in UGIB (2020), which is summarized as follows:
a. The fuel needed at 19:41 for MEFE at 00:17:30 (with the wing tank cross-feed valves always closed) is 27,113 kg.
b. The 19:41 fuel needed to match the MEFE time is reduced by about 450 kg to 26,660 kg if the air packs are OFF after 19:41.
c. Using the FMT Route from UGIB, the available fuel at 19:41 is nominally 26,045 kg, for a shortfall of at least 615 kg.
d. Turning the air packs OFF from 17:26 to 19:41 saves 245 kg of fuel, but the shortfall is still 370 kg.
e. Turning the right engine OFF when at FL100 saves another 350 kg of fuel. Now the shortfall is only 20 kg, which is within the prediction noise to match MEFE.
f. A small 5% reduction in electrical load after 17:26 would save another 20 kg of fuel to produce an exact MEFE match. Larger electrical load reductions may be possible.
3. If the BEDAX 180 degree route is even roughly correct, which is confirmed by the drift model analyses in Ulich and Iannello (2023),then it is probable that both the air packs were OFF after 17:26 and the right engine was shut down for about an hour when at FL100.
4. Alternatively, even if there was no descent in the FMT route, it is still necessary for the air packs to be OFF after 17:26.
5. Even if the electrical loads were cut in half, that would save only about 200 kg of fuel, and that is insufficient to exclude the air packs being OFF after 19:41.
6. If the air packs were off after 17:26, or even after 19:41, no humans were alive in the aircraft at the 7th Arc. In my opinion all but the pilot were deceased by the last radar contact at 18:22. Even the pilot would be highly unlikely to survive past 21:00, and I suspect his death occurred shortly after 19:41. Once the course and cruising speed and altitude were set circa 19:41, no further human intervention was needed to fly until fuel exhaustion. My conjecture is that the pilot committed suicide then, his mission being assured of accomplishment.
7. A piloted glide scenario near Arc 7 is extremely unlikely. For this to have occurred, either (a) the pilot achieved a cruise fuel flow savings of about 3% without turning the air packs OFF or (b) the pilot somehow survived for more than four hours in an unpressurized aircraft near 40,000 feet. The presence of ample supplemental oxygen and a pressure-demand mask, which were available in the cockpit, still does not allow pilot survivability because the cabin pressure was very low for many hours, and the effort required to exhale cannot be maintained for such lengthy periods.
@DrB: What’s your best explanation today for why the debris field was not found?
@All,
Re fuel endurance.
I think that a hypoxic, unresponsive crew scenario can also fit with fuel flow endurance calculations.
If the flight controls have been degraded to ‘secondary’ at IGARI due to damage of the pitot/static system, the autopilot/autothrottle would have failed. Therefore the throttles would remain at their set position until the end of flight. At IGARI the fuel burn was 6400kg/hr, as the fuel weight reduced, by the end of the flight and at around 40000ft with the throttles in the same physical position the fuel burn would be around 5600kgs/hr.
So, no need for complicated scenarios as to how the aircraft would have the endurance to to make it to 01:19z
@Tim: The military radar data shows MH370 rounded Penang Island, then flew direct waypoint VAMPI and joined airway N571. That is not consistent with secondary flight mode with a hypoxic crew.
Sorry, that should be MEFE at around 00:19z not 01:19z.
So from IGARI, 42000kgs of fuel used over 7 hrs. An average of 6000kgs/hr. I was just trying to show with the fuel flow reducing from 6400 to 5600kgs/hr, the average is 6000kgs/hr.
@Victor
I still feel the military radar trace after 18:01z has never been backed up by sufficient evidence
@Tim: If you want to demonstrate there is enough fuel with constant throttle lever position, you have to show that for a given power setting and altitude, there is a trajectory such that the time-varying weight, wind, temperature produce an airspeed (and hence a groundspeed) has sufficient fuel AND the satellite data (BTO and BFO) are satisfied. If it not enough to say there is sufficient endurance for the fuel to last until 00:19z.
As for rejecting all the military data after 18:01z, that would imply the Malaysians are lying in the SIR when they claim that MH370 joined N571 and followed waypoints. That’s not impossible, but it would be odd that they chose to do so.
@Victor
If you say so, but 2000-ft drop sounds like bare minimum. I was thinking around 10000 ft from flight sims by hand. Of course, you gain some speed so the momentum trade off is there. For my scenario it is a lower altitude event.
@DrB
Right now, I am not going to take a crack at where I differ on that list. But I agree somewhat that the long glide scenario from high altitude, passive, straight flight is problematic. Of course Blelly and Marchand take a crack at explaining it.
@Victor,
You asked: “What’s your best explanation today for why the debris field was not found?”
Your current post on the debris field provides a partial answer. The high-speed impact shredded the aircraft (except for the flight control surfaces which departed prior to impact). That includes the main landing gear, which may have separated into smaller components. In this case, the largest debris was likely the engine cores, and these were somewhat smaller than would be typical for lower-speed crashes. So, one factor in the non-detection is that the debris on the seabed is smaller than is typically the case. A second effect of the high-speed impact is that the smaller debris are dispersed over a larger area on the seabed (like confetti falling in the air), especially at the significant depth in this case. So, I expect the MH370 debris to be smaller and more widely dispersed, making detection and especially classification more difficult.
Another factor is the seabed terrain. The so-called “difficult terrain”, including ridges and canyons, prevented complete coverage. The debris field may be located in one of these areas which was not searched or was inadequately searched, as you discussed in your previous post.
I think both factors are contributing to the failures of the seabed searches to date.
@DrB
The amount of fuel saved by having the air packs OFF after 19:41 (450 kg) is roughly equivalent to the amount of fuel saved by moving the 7th arc crossing point about 0.5 degrees north and thereby shortening the route.
According to the PDF in UGIB (2020), there’s only about 53% probability of S34.23 +/- 0.5 degrees, while latitudes north of S33.7 have about 24% probability. S33.7 is also well consistent with the more recent June 2023 drift study.
Therefore, I don’t understand the conclusion ”A piloted glide scenario near Arc 7 is extremely unlikely”, when the whole argument about air packs is dependent on S34.23 being very accurately true (in addition to the fuel model itself being highly accurate with high confidence), and clearly there’s no high confidence in this level of latitude accuracy even in UGIB (2020) or U&I (2023).
@TBill said: If you say so, but 2000-ft drop sounds like bare minimum.
That’s based on analysis of losses as well as FSX simulations. Here’s a back of the envelope calculation: The BFO values suggest a downward acceleration of 0.7g. At this acceleration, it would take 11.1s to go from level flight to a descent rate of 15,000 fpm (250 ft/s) assuming uniform acceleration. (In a simulation, you would have to aggressively lower the nose.) In that time, the plane would lose 1389 ft of altitude. If there were no losses, all of this altitude loss would be recoverable, but because of the high airspeeds, the frictional losses are high, and the plane could not recover to the initial speed and altitude. In any event, I don’t see how 10,000 ft loss of energy could occur over such a short time (11.1 s).
@Victor,
Unfortunately, I’m not able to do the maths to make the all BFO/BTOs fit a trajectory. But what I have observed over the years is there are many flight paths that can be made to fit the data.
If this flight was a ghost flight with no autopilot or pilot, then I fear there are just too many variables to narrow down exactly where on the seventh arc the flight ended.
@Victor
I am not going to argue the physics, but do we know when the pilot ended the dive? He could have kept descending for a bit, once going that steepness. I am thinking we are lucky to capture this descent in the data, almost as lucky as catching the cell phone connect in Penang.
@TBill: If you are going to argue there was a controlled descent after the initial steep descent, then you have to assume the maximum possible glide distance to determine the limits to the search area. The maximum glide distance occurs when there was only a small loss in energy (altitude) during a short descent/recovery transient.
@eukaryote234,
Near MEFE the fuel flow is about 90 kg/min and the aircraft speed is about 8 NM/min. Moving the flight path intersection with Arc 7 from 34.2 S to 33.7 S reduces the range from the best-fit 19:41 location by about 36 NM. This represents about 4.5 minutes of range, or 400 kg of fuel savings. As you pointed out, this is not too different from the 450 kg fuel savings predicted with the air packs OFF from 19:41 to MEFE. Thus, from a fuel perspective, you are correct – both flight paths are possible. That’s why the fuel probabilities of these two cases are close to the same value.
The route and fuel probabilities for the best-fitting case at each latitude along 7 were presented in UGIB (2020). The updated version (from earlier this year) is available here:
https://drive.google.com/file/d/1r90vOQDpNGzFFS5E3ZoUdWCWWSTAVix1/view?usp=sharing
To summarize:
1. As you said, the need for air packs OFF after 19:41 could be offset by flying at a reduced speed (perhaps MRC) after 19:41 to a point to the NW along Arc 7 from the UGIB (2020) LEP, following a shorter path.
2. This route is significantly less probable in the route probability.
3. The fuel probability is essentially unchanged.
4. The combined route and fuel probability is about 1.8X higher at 34.2 S than at 33.7 S.
5. Therefore, a lower-speed route to circa 33.7 S could be flown with the same fuel as the higher-speed route to 34.2 S, but this slower route has a 56% lower probability of having occurred. I agree that my “extremely unlikely” descriptor for the air packs being OFF after 19:41 is overstated for your counter-example. Calling it “unlikely” is more accurate.
In addition, Figure 16.1-1 in UI (2023) shows the boundary of the predicted 00:21:07 impact zone reaches to 33.5S, so an impact circa 33.7 S is certainly possible, even for the best-fit route, although an impact in that case at 33.7 S is also less likely than an impact circa 34.2 S.
@DrB. While your Nov 14 8:56 PM pack-off/no-pilot-at-the-end scenario is well supported, nevertheless it is based on APU auto-start at MEFE prompting the final SDU reboot.
Also feasible though is a piloted glide following a manual engine shut down sometime after the 6th arc and before MEFE, that being to conserve fuel for a later APU start.
One possible purpose would be to attain full control in a planned final plunge.
Neither the pilot nor others like him would know of the residual fuel accessible to the APU, this having arisen subsequent to the accident.
However APU auto-start after engine shut down with engine fuel remaining would not be unexpected and in this scenario he would inhibit that to save that fuel for later.
LOR would be two minutes after he manually selected that later APU start, as per the auto-start.
All the same, were his glide continued in the preceding direction, ie following the UGIB track it would need to be at high speed to meet final log-on BTO criteria, too high probably for that to be realistic.
Still, turning north would reduce that transit length, as @eukaryote234 has observed.
Though you do indicate his 33.7degS final LOR latitude to be less likely overall than going straight on to 34.2, this above manually-controlled APU scenario would provide an explanation as to how, piloted, the final LOR came to be rather than he awaiting MEFE passively.
More importantly though, introducing such a glide at normal speed would obviate the reduction of powered speed otherwise needed to time the LOR after shortening the transit from the current 6th arc UGIB position.
That otherwise goes unraised.
And were this final glide short then there might be a solution with the LOR south of 33.7.
Another aspect of the above glide scenario is, to me, that it reduces the likelihood of a pull out from the plunge for a controlled ditching. (Besides, he would be aware that that could leave him alive…)
Summarising, the final transmissions might have been initiated by the pilot selecting APU start using remaining fuel while in a glide. A final plunge would follow a little before two minutes later, the SDU coming on line during that.
So this scenario would support a 6th to 7th arc transit ending further to the north than if unmanned.
For completeness, a high speed glide as above to a 34.2 LOR, while unlikely, would result in getting there at a lower altitude, making it even more likely that there would be no glide subsequent to that plunge.
@DrB. Adding to your Nov 15th, 10:21 AM, a couple of comments:
About wreckage non-detection you wrote, “the largest debris was likely the engine cores, and these were somewhat smaller than would be typical for lower-speed crashes.” A thought here is that it is unlikely that the MH370 engines were other than windmilling, that low speed tending to reduce damage compared with, for example, the full engine speed in the AF447 impact. This and possibly the engine in-line force, as against the AF447’s ‘perpendicularity’, might offset the effect of high impact airspeed, to some extent.
Also, were the undercarriage fractured, to me that doesn’t mean necessarily that a larger number of smaller but still solid bits would be less reflective or as a field than if unbroken.
Further, adding to your, “The so-called “difficult terrain”, including ridges and canyons…..which was not searched or was inadequately searched…” Mud too might well accumulate in pockets on the slopes and at their bottoms, part-burying wreckage that has tumbled, thereby reducing its sonar reflectivity.
All in all, hard to say.
@DrB
In the original UGIB fuel model, there’s a sharp drop in probability that starts around S34.2. This is consistent with the points made in the earlier comment (Nov 14, 8:56 pm) regarding the shortage of fuel at S34.23.
However, according to the updated fuel model, there’s still about 97% probability for S35.8. How is this possible, if at the same time there’s barely enough fuel for S34.23? Based on the increase in distance alone, the S35.8 route should require at least 1,100 kg more fuel:
Starting point (19:41): [2.930, 93.788], Fuel 26,716 kg
Endpoint 1: [-34.234, 93.788], dist. 4,113.6 km
Endpoint 2: [-35.80, 91.73], dist. 4,292.7 km (+179.1 km, +4.354%)
Additional fuel: 0,04354*26,716 kg = 1,160 kg
It’s explained in the paper that the new shift toward S36 is based on identifying new routes, but what is it about these new routes that compensates for the increase in distance?
You have to work out what was the criteria required for the creation of the FMT. What was it designed to do, or avoid? Radar avoidance when heading south, after being seen heading Northwest is one such major consideration.
I have worked out such an FMT, which is followed by a route to the 7th arc on a 180T track. It takes 150NM of the track created by the FMT apparent in the simulator data. It crosses the 7th Arc at ~ S36.15.
@Mike Glynn: What you are assuming about the FMT is possible, but you also have to acknowledge it is a guess.
@Mike Glynn
A route ending in S36.15 would require something like 1,350 kg more fuel for the post-19:41 portion of the flight compared to S34.23. So instead of having to save 615 kg (which already required significant measures), the FMT portion of the S36.15 route would have to save 1,965 kg compared to the standard UGIB FMT route. Is this really possible?
Victor, it is based on the maximum range of the Sabang PSR and the probability that Zaharie intentionally flew inside that range while heading NW, as he did with the other three TNI PSRs on Sumatra, and avoided flying inside that range after the turn due south. Add to that the imperative to burn as little fuel as possible during this manoeuvre. A mate of mine, a former RAAF Fighter Combat Instructor, for whom this was his bread and butter, confirmed the range details for me.
So it’s not a guess, there Is reasoning behind it.
Zaharie could not know that Sabang was not operating and had to assume it was. The initial simulator coordinates had a turn based on avoiding Sabang, and tracking towards NZPG. Calculating the range of Sabang shows that Zaharie plotted a track that gave him ~ a 12 nm buffer outside Sabangs max range.
Using the same logic for a 180 degree track south, and the same buffer distance from the radar of 12nm, leads to an intersection with the 7th arc at the point I mentioned. As I said, the distance flown is 150 nm less. I haven’t yet done the fuel calculations however.
The initial ruse worked as evidenced by the search activity in the Andaman sea.that went on for a week or so before the INMARSAT data was analysed
@eukaryote234,
You said: “In the original UGIB fuel model, there’s a sharp drop in probability that starts around S34.2. This is consistent with the points made in the earlier comment (Nov 14, 8:56 pm) regarding the shortage of fuel at S34.23. However, according to the updated fuel model, there’s still about 97% probability for S35.8. How is this possible, if at the same time there’s barely enough fuel for S34.23? Based on the increase in distance alone, the S35.8 route should require at least 1,100 kg more fuel. . . . It’s explained in the paper that the new shift toward S36 is based on identifying new routes, but what is it about these new routes that compensates for the increase in distance?”
My updated curve indicates a sharp drop in fuel probability circa 36.5 S instead of circa 34.5 S as in UGIB (2020). That occurs because readers of this blog (including George G) identified additional combinations of altitude, speed, and track which end circa 36 S at Arc 7 and for which sufficient fuel would be available if all the proposed fuel-savings configurations were also applied. Since for this 36 S route the distance to the intersection at Arc 7 from N571 is greater than the distance to the UGIB LEP, how is this possible? The answer to that question is that the 36 S route can be flown at a more fuel-efficient speed (lower Mach, such as MRC or M0.81) if one allows the 19:41:03 location to be shifted SSW of the UGIB best-fit location for the 34.2 S route. Slightly lower flight levels are also necessary, circa FL360. Thus, it is possible to fly a B777 to 36 S and have MEFE at 00:17:30 with these constraints (#3, 5 and 6 below answer your question about what compensates for the increased distance):
1. The air packs are OFF after 17:26 until MEFE.
2. A lengthy HOLD at a reduced altitude must also occur after the turn from N571.
3. The speed setting is MRC and the flight level must be near FL360.
4. The BTO/BFO fit is not quite as good, slightly reducing the route probability compared to the 34.2 S route.
5. There can’t be a flyover of Kate Tee’s sailboat.
6. The “FMT Route” is not along FIR boundaries, and BEDAX is not overflown. The HOLD route must be SSW or SW.
7. It is uninvestigated if such a route could have been conveniently entered into the FMS using existing waypoints or airways.
Using MRC speed for the 36 S route, instead of LRC, reduces the speed by about 2.2% and reduces the fuel flow by about 3.2% at the same flight level. Thus, the fuel mileage is improved by 1% for MRC compared to LRC (that’s actually how LRC is defined). Further reductions in speed below MRC worsen the fuel mileage, since MRC (= Maximum Range Cruise) has by definition the highest fuel mileage possible.
To summarize, for Arc 7 intersections at latitudes near 33.7 S, the route probability is reduced compared to 34.2 S, the fuel probability remains high even with the air packs ON after 19:41, and the drift probability is high.
At 36 S the route probability is only slightly reduced, the fuel probability can be high if one relaxes all constraints on the FMT route track (such as following FIR boundaries and overflying Kate Tee), but the drift probability is strongly reduced.
@Mike Glynn: It might be an informed guess, but it’s a guess. Various informed investigators over the years have proposed a wide range of POIs.
In UGIB 2020, we tried very hard to not inject our biases, and let the physics and statistics point us to a hot spot with the assumption of automated flight with no pilot inputs after 19:41. Whether we succeeded or not is debatable, and we can’t even be 100% sure there were no pilot inputs, but that was our goal. Of course, if there were pilot inputs, it becomes practically impossible to define a hot spot.
I’m not saying your assumptions were wrong. I am simply saying that they reflect your biases about the pilot’s intentions.
@eukaryote234, @DrB: I was about to respond that @eukaryote234’s assumption that the position at 19:41 for the two paths is not correct. In general, for LNAV paths (and also for constant true track paths), the position at 19:41, the position at 00:19, and the track at 19:41 are all related–you can’t change one without changing the other two while also satisfying the BTO data.
@David,
You said: “Also, were the undercarriage fractured, to me that doesn’t mean necessarily that a larger number of smaller but still solid bits would be less reflective or as a field than if unbroken.”
Smaller pieces of wreckage spread over a larger area will make the debris field more difficult to detect and especially more difficult to classify as being man-made.
@Victor, all
Interesting article with details about the GO Phoenix search effort. I haven’t been following all discussions lately, so perhaps it has been addressed already: Is there an explanation for the rather uneven distribution of these 45 contact points, in particular why are there no contacts south of S35.25? If I remember well the GO Phoenix search extended further south to about S36.0 (?)
By the way, at the time of the OI search I was much disappointed that they went quite far north and initially went “around” already scanned areas. For many years, I have been in favor of focusing the search close to the 7th arc under the assumption that you would probably need to scan the same area more than once given the techniques that were used at the time. I was discussing the reversed drift modelling that David Griffin had applied to the “pleiades” observations with him at that time, and thought it should be applied to the area closer to the arc, so in the area already scanned. I still think this option should be followed, as part of a possible next search.
@Niels: I don’t know why the distribution of contacts is what it is.
My estimation is there isn’t much appetite to re-scan areas already scanned.
@DrB, @Victor Iannello:
To be clear, the S36.0 route has about 1,000 kg worth of fuel savings prior to crossing the latitude N2.930 compared to UGIB S34.23 (by not following FIR boundaries, no flyover of Kate Tee’s sailboat etc.)?
The 1% reduction in the SIO (using MRC instead of LRC) provides about 250 kg, but this is only a small part of the original deficiency (1,300 kg) caused by the difference in post-N2.930 distance (200 km) between S36.0 and S34.23 (the actual longitude at which the S36.0 route crosses N2.930 makes almost no difference in terms of the distance). The 19:41 location of the S36.0 route shouldn’t matter much in terms of fuel, other than facilitating the use of MRC in the SIO in this case (almost all of the 1,000 kg has to be explained between 18:22 and N2.930).
Adding to the previous comment, that regardless of the 19:41 location of the S36.0 route, there can’t be any significant fuel savings between N2.930 and 19:41 if S34.23 already has 99% of the maximum fuel efficiency per km in that part of the route (LRC).
@eukaryote
How are you calculating so much extra fuel required to get to 36s? Great circle distance-wise 3N to 34.2s is about the same as 1N to 36s. Off the top of my head, in general, proposed nominal 180s flight paths have ranged from 0N to 5N on Arc2 at 1941.
@eukaryote,
You said: “To be clear, the S36.0 route has about 1,000 kg worth of fuel savings prior to crossing the latitude N2.930 compared to UGIB S34.23 (by not following FIR boundaries, no flyover of Kate Tee’s sailboat etc.)?”
That is incorrect. The fuel savings only occur AFTER 19:41, because the fuel flows before then are the same or higher. The fuel required at 19:41 is about 3.3% (870 kg) LESS for the 36 S MRC Route compared to the 34.2 S LRC Route. That’s what makes the MEFE occur at the same time (00:17:30 UTC) because of the reduced fuel flow from 19:41 to MEFE. At the optimum altitude using MRC the speed is 2.2% less than LRC but the fuel flow is 3.3% less, making the fuel mileage 1.0% higher.
As I shall demonstrate below, there is enough fuel to fly to Arc 7 at 36 S using MRC. There is also (barely) enough time if a turn to 180 S with no descent is made immediately prior to the 18:40 phone call.
First, I address the fuel. The nominal UGIB 19:41 location is 2.9136 N, 93.78 E. The range from there to the Arc 7 intersection is 2,566 NM to 34.23 S and 2,693 NM to 36.00 S. Thus, the post-19:41 36 S Route is 4.7% (127 NM) LONGER. At a 2.2% lower ground speed, covering the longer distance at the slower speed means the starting point (the 19:41 position) must be 7.0% closer to Arc 7 than the UGIB 2.91 N latitude. This 7% is 180 NM, or 3.0 degrees of latitude. Thus, the 36 S Route would be virtually crossing the Equator (~ 0.0 N) at 19:41. It takes 23 minutes to fly that 180 NM at the 469 knot ground speed using MRC, so the aircraft would pass the UGIB 19:41 location near 2.91 N approximately 23 minutes prior to 19:41, or circa 19:18.
It’s not obvious one can gain a 180 NM “head start” compared to the UGIB FMT Route, while only burning an excess of 870 kg or less after the turn off N571. One can gain about 50 NM by taking the most direct path after 18:40. Therefore, one must increase the speed well before 19:18 (compared to the nominal UGIB FMT Route) to gain the remaining 130 NM of the required “head start”. You have to cruise at MRC for about 45 minutes to gain than much compared to low-altitude Holding speed. That means you can’t do a low-altitude descent at INOP Holding – you have to cruise at MRC for the entire time from N571 departure to 19:41, and during this time you are burning fuel at a HIGHER rate (roughly 5,900 kg/hr at MRC versus 5,160 kg/hr at INOP Holding at FL100). So, it costs you an additional 740 kg/hr to fly the FMT Route at MRC, and over that 45 minutes you burn an extra 550 kg compared to the UGIB FMT Route. This is less than the 870 kg reduction you can accept and still meet MEFE at the correct time. So, If you don’t descend at all after departing N571, you can reach the 36 S location with the available fuel.
Next, I address the speed. From the nominal 18:41 UGIB location at 7.5 N, the distance to the Equator is 450 NM. That must be covered in the 60 minutes from 18:41 to 19:41, so one needs a ground speed of 7.5 X 60 = 450 knots. This is doable at MRC (469 knots). So, in principle one could fly the distance from the phone call location to the Equator, but what about the 18:40 BFOs during the phone call? In UGIB these BFOs are matched during a descent, and if you make a descent then you can’t reach the Equator on time. We also know that a track of 180 degrees at MRC can match the 18:40 BFOs with no descent, so here I assume a turn to the south occurred immediately PRIOR to the phone call, and no descent was made. Then all the BFOs are matched and the Equator is reached on time.
I conclude the 36 S MRC Route is flyable both in terms of available fuel and in matching the BTOs and BFOs. This is consistent with the (updated) UGIB fuel and route probabilities.
@DrB
You said: ”That is incorrect.” in response to the statement ”the S36.0 route has about 1,000 kg worth of fuel savings prior to crossing the latitude N2.930 compared to UGIB S34.23 (by not following FIR boundaries, no flyover of Kate Tee’s sailboat etc.)”.
In your estimation, what is the amount of fuel on board the S36.0 plane at the time it passes N2.930 around 19:18? And how does this compare to the amount of fuel on board UGIB S34.23 case 7B1 at the time it passes N2.930 at 19:41 (26,565 kg)? I’d need to look at the details in your post more carefully before a more proper response, but I can already say that you’ve probably misinterpreted what my above statement was intended to mean.
I have not seen what the proposed S36.0 route looks like, but I assume it’s significantly shorter in the period between 18:22 and N2.930 compared to UGIB, not taking the long detour north for the 18:39 call. And for this and other reasons, it reaches N2.930 earlier (19:18) and with significantly more fuel than 26,565 kg, making it possible for it to reach S36.0. After writing the earlier comment, I’ve since learned/noticed that UGIB has a much more ”fuel heavy” FMT compared to some other proposals, so I’m definitely no longer questioning whether the S36.0 route is possible in light of the UGIB fuel model (it can be if the FMT is replaced).
@TBill
I used a fixed latitude point, not timing. If the route length is 200 km longer after N2.930, it represents extra fuel that has to be compensated. I just don’t see how the timing is so relevant. Let’s say that the 2nd arc was at N1.16 for the S36.0 route. Then the post-19:41 length would be the same as for S34.23. Fine, but then you need extra fuel to get to that N1.16 compared to N2.93. In the case of the proposed S36.0, the major form of compensation is almost certainly a significantly shorter route between 18:22 and N2.930.
@eukaryote
Because we are all just guessing at flight path between Arc1 and Arc2, you can just change the path Arc1-Arc2 to get further south with the same fuel to Arc2. I would critique the UGIB path because it needs a little bit extra loiter to hit Arc2 a bit further north, which is needed for the match of LRC speed hitting Arc7 on LNAV to South Pole at the correct time.
@TBill:
My understanding of this topic now (someone can correct if wrong):
1. In addition to the FMT route used in UGIB, there are other possible FMTs that use much less fuel before N2.93.
2. The original fuel PDF was derived using only that one specific FMT for all latitudes (except adjusting for the 19:41 point), causing the cutoff around S34.3.
3. It’s only by allowing the use of other FMTs that S35-36 becomes possible in the updated PDF.
4. I do think it’s somewhat misleading to say that this shift toward S36 was based on ”identifying new routes”, when there was no attempt to find/use other FMT routes when creating the original PDF. It’s basically just a matter of gaining more search results by removing a search restriction (specific FMT) that was previously put in place.
5. You could also use a different, less fuel consuming FMT for a S34.2 route, thereby removing the need for air packs being OFF.
6. The probability of a particular FMT is a matter opinion and can’t be objectively assessed (unlike the post-19:41 route probabilities), as long as some basic criteria are met like the 18:40 BFOs.
(Small correction to earlier comment: the specific fuel amount I used for 7B1 26,565 kg, while for 19:41, is not exactly for N2.930 but right after that. So the exact figure would have been 20-40 kg higher.)
@eukaryote,
You are generally correct. The 36 S route actually has TWO fuel conservation benefits. One occurs AFTER 19:41, when the highest fuel mileage speed mode (MRC) is used instead of LRC. The other one occurs BEFORE 19:41 because a more fuel-efficient speed mode is used after 18:29 AND the FMT path is more direct. These two benefits combine to allow 36 S on Arc 7 to be reached on time and with the available fuel, if a new and optimized FMT Route is substituted.
More specifically, you asked: “In your estimation, what is the amount of fuel on board the S36.0 plane at the time it passes N2.930 around 19:18? And how does this compare to the amount of fuel on board UGIB S34.23 case 7B1 at the time it passes N2.930 at 19:41 (26,565 kg)? “
Regarding the 36 S Route, the only way I know of to match the 18:25-18:28 BTOs and BFOs is a lateral offset maneuver off N571. Then a turn to the south after 18:28 could have been made which will match the upcoming 18:40 phone call BFOs. For the 36 S Route, that turn must have occurred between the end of the lateral offset maneuver and the beginning of the phone call. The exact location of the south turn when departing N571 is not critical in terms of distance to Arc 7 because N571 is nearly perpendicular to the line to the 36 S end point, but it is important in terms of time and fuel. The distance to the UGIB 19:41 position at 2.91 N from the turning point circa N571 varies from 324 NM at 18:29 to 320 NM at 18:39. Similarly, the distance to the Arc 7 end point varies from 2,997 NM at 18:29 to 3,012 NM at 18:39. So, within a few miles, the distance to be flown to Arc 7 is nearly constant, no matter the turn time off N571, but the fuel is continuing to be consumed at LRC flows over the 9 minutes it takes to fly the same distance parallel to N571 as shown in UGIB Figure E-1 before the south turn. Thus, from 18:29 to roughly 18:38, with no descent occurring, the aircraft would cover about 72 NM putting it near 94.4 E and consuming an extra 900 kg of fuel. Thus, turning at 18:29 saves about 900 kg of fuel compared to turning just before the phone call, and the distance to the 36 S end point is virtually the same.
The distance from the UGIB 19:41 position at 2.91 N is 2,566 NM to the 34.2 S end point. Thus, we can compare the two “final” flight legs:
(a) for the 34.2 S Route, the starting point is 2.91 N, 93.8 E at 19:41, the fuel on board is 26,680 kg, the distance is 2,566 NM, and the speed setting is LRC, and
(b) for the 36 S Route, the starting point is circa 7.2 N, 95.6 E at 18:29, the distance is 2,997 NM, and the speed setting is MRC.
From UGIB Table C-3, the fuel required at 19:41 for the 34.2 S Route is 26,680 kg (which is Case 9B1), with bleed air OFF. During this same period, from 19:41 to MEFE at 00:17, the MRC fuel flows would be 3.3% less, and therefore one needs only 25,830 kg, for a reduction of about 850 kg for the 19:41 – 00:17 fuel used in the 36 S Route. From 18:29 to 19:41, the average MRC fuel flow would be about 3,000 kg/hour/engine at 205 tonnes average total aircraft weight. Thus, the fuel needed at 18:29 is 7,200 kg more than 25,830 kg, or 33,030 kg.
The 36 S MRC Route would reach 2.91 N (at roughly 95.0 E) about 38 minutes after the turn, or circa 19:07 (using the 18:29 turn). At the MRC speed, the aircraft could fly about 2,510 NM (i.e., 2.2% less than at LRC) from 19:41 to Arc 7. This range from the 36 S end point puts the aircraft at 0.2 N at 19:41, near 94.5 E. Working back in time from there, the 36 S MRC Route crosses 2.91 N latitude about 21 minutes prior, circa 19:20. That 21 minutes of flight uses 2,100 kg of fuel, so the fuel on board at 19:20 (and at 2.91 N) is about 25,830 + 2,100 = 27,930 kg.
When the UGIB 34.2 S LRC Route is at 2.91 N at 19:41, the fuel on board is 26,680 kg (for case 9B1; we did not show 7B1).
When the 36 S MRC route is at 2.91 N at 19:20, the fuel on board is about 27,930 kg, which is about 1,250 kg higher. This difference is nearly the same for Case 7 (1,365 kg).
As you supposed, this “extra” fuel at 2.9 N is part of what allows the more-distant 36 S to be reached. The other part is the more fuel-efficient speed AFTER 19:41.
The “extra” fuel at 2.9 N occurs because a more fuel-efficient speed is used between the turn off N571 and 2.9 N, AND the FMT route is more directly south and therefore shorter.
@All,
As I said last spring, the route to 36 S can be flown with the available fuel and with a good, but not the best, match to BTOs and BFOs (which occurs for the UGIB Route). The UGIB FMT Route is driven by minimizing FMT fuel flow while matching the best-fit 19:41 location, track, and speed. This requires either a HOLD or a somewhat meandering route. Otherwise, the aircraft reaches the best-fit 19:41 location too soon.
Other considerations for the 36 S Route are:
1. Because of the more fuel-efficient MRC speed and the more direct FMT Route, it is unnecessary for the Air Packs to be OFF after 19:41 to reach Arc 7.
2. The route seems completely arbitrary after 18:28. So far, there are no identified waypoints along the path or beyond the splash point. That does not mean it was difficult to command, just that the course might have to be set in an unconventional way. One could set a true track mode, but that is an abnormal procedure. One could also set a destination using latitude and longitude which is beyond Arc 7.
3. There is no attempt to be stealthy when passing by Sumatra or the islands offshore. In fact, the 18:29 turn path actually overflies Banda Aceh at the NW tip of Sumatra. Granted, it was the middle of the night, but that is not a normal airway, and ground observers might have noticed the plane.
4. No 36 S Route which is otherwise acceptable can be consistent with Kate Tee’s sighting of an airliner at an unusually low altitude (unless she saw a different, and as yet unidentified aircraft).
5. There is no attempt to fly along FIR boundaries, which might have been a motivation for the overflight path across Malaysia (see Section 2.1 in UGIB).
None of these considerations provide iron-clad confirmation or denial of the 36 S Route.
@DrB
You said: for the 36 S Route, starting at 18:29 at position circa 7.2 N, 95.6E, the distance is 2,997 NM, and the speed setting is MRC.
Q1. what is your altitude and fuel on board in this case ?
Q2. could the range and/or endurance be extended with a ‘drift climb’ from that point onwards ?
Q3. If so, how much further could it go (in Nm), and how much additional endurance (in minutes) could be gained before MEFE ?
Q4. how much further SW on the 7th Arc could it theoretically reach with these changes ?
I cannot see any reason why the PF would use ANY waypoints for navigation after the FMT, except for a distant end point, like the south pole. Why would anyone use any intermediate WPs if you are headed for MEFE somewhere in the SIO?
@Ventus45,
You said: “You said: for the 36 S Route, starting at 18:29 at position circa 7.2 N, 95.6E, the distance is 2,997 NM, and the speed setting is MRC.
Q1. what is your altitude and fuel on board in this case ?”
A1: From my previous comment on 11/20 at 12:00 p.m.: ”Thus, the fuel needed at 18:29 is 7,200 kg more than 25,830 kg, or 33,030 kg.” At the 206 tonnes aircraft weight then, the optimum flight level at MRC is FL360 – FL380. That’s slightly below the estimated flight level during the N571 offset maneuver of FL385 at LRC. Thus, to slow down from LRC to MRC, one could descend slightly to maintain optimum fuel mileage, but the fuel mileage penalty incurred by remaining circa FL385 is very small, and a descent then is not necessary from fuel considerations.
You also said: “Q2. could the range and/or endurance be extended with a ‘drift climb’ from that point onwards ?
Q3. If so, how much further could it go (in Nm), and how much additional endurance (in minutes) could be gained before MEFE ?
Q4. how much further SW on the 7th Arc could it theoretically reach with these changes ?”
A2: The fuel flow penalty caused by being slightly off the optimum altitude is less than 0.6% in the worst case in the stratosphere at MRC. Yes, a slow rise in flight level as the fuel weight drops can buy you a small amount of fuel. Remaining at a constant FL385 for the rest of the flight limits the penalty in fuel flow to 0.4% or less at MRC, which is 132 kg. This is “lost in the noise” in the fuel flow modeling.
A3: Having 132 kg of “extra fuel” extends the range by about 1.5 minutes and the range by about 12 NM.
A4: The extra range allows reaching Arc 7 circa 36.15 S.
From my previous comments, the fuel cost of the 36 S MRC Route prior to 19:41 is about 550 kg compared to the UGIB FMT Route. The fuel savings after 19:41 is about 850 kg. Therefore, the net fuel required for 36 S at MRC is about 300 kg less than 34.2 S at LRC, with identical aircraft configurations. Turning the air packs ON after 19:41 costs about 450 kg of fuel. Thus, the 36 S MRC Route is only about 150 kg shy of allowing the air packs to be ON after 19:41. That is within the fuel model uncertainty (or can be obtained by a slow climb after 19:41 from FL370 to FL400), so I conclude it is possible for the air packs to be ON after 19:41 for the 36 S MRC Route.
@All,
There are several considerations which, in my opinion, favor the 34.2 S Route:
1. The best-fit 19:41–00:11 route passes over a waypoint (BEDAX) near the end of the FMT Route, as shown in Figure 20 in UGIB (2020). I am not aware this occurs for any other great circle route, but I have not thoroughly searched for them. A few may occur by coincidence. There may or may not be a waypoint that defined the start of the SIO Route, but I think there was one past MEFE. I doubt the PF set a final course using true track.
2. The best-fit 19:41–00:11 route has a track of exactly 180.0 degrees (see Figure 19 in UGIB). That can easily be accomplished by setting the South Pole as the waypoint after BEDAX.
3. The Route Probability has a unique and very narrow peak at 180.0 degrees azimuth and a shape which is complex but symmetrical about 180 degrees (see Figure 19 in UGIB).
4. A plausible FMT Route using standard waypoints can connect to the best-fit 19:41 latitude and longitude at the correct time, speed, and track.
7. That route passes over Kate Tee in the right place and at the right time for her sighting to be 9M-MRO.
@eukaryote
Here is one way to look at Arc1 to Arc2.
According to Boeing there is 2806nm fuel remaining at max fuel efficiency at MRC and FL400 from 18:25. Until 19:41 that is 1.26-hrs of flight and about 610nm at your disposal at MRC until you hit Arc2 somewhere. Debit from distance if you envision not flying at MRC (but I personally envision fuel conservation was a priority).
@DrB
I agree with the lateral offset maneuver 1828 to 1828 however I am thinking it could have been a left offset, which has several advantages. Left offset saves fuel and time and also possibly more consistent with turn south at 1840 near ANOKO, whereas with the right offset, it is a little awkward to explain whether 1840 BFO is due to descent or turn south.
@TBill: Are you sure you can match the BTO and BFO with a left offset?
@DrB
Regarding the S34.2 route: your point 3 would be perhaps be the most objective criterium from the list
(3. The Route Probability has a unique and very narrow peak at 180.0 degrees azimuth and a shape which is complex but symmetrical about 180 degrees (see Figure 19 in UGIB).)
However, given the rather large possible errors in measured BTO and BFO I find it still hard to understand how you can extract meaningful detailed information from such “noisy” data. When I do a basic statistical analysis on my MC generated data set (20 million paths) I get typically a wide range of possible 00:19 latitudes (S30/31 – S37 or so), which I find hard to reconcile with your much more focused result.
In other words: can the “unique and very narrow peak at 180.0 degrees..” be a coincidental result?
@Niels,
You asked: “. . . can the “unique and very narrow peak at 180.0 degrees” be a coincidental result?”
There are several reasons why I believe the narrow peak is both real and informative.
The Figure of Merit being evaluated in Figure G-7 in UGIB (2020) is the Route Probability, which is the Fisher chi-squared composite probability based on nine Z-statistics for LNAV routes. For each of the nine parameters you must estimate both the mean and the standard deviation of the parameter so you can compute its Z-statistic value (as shown in Table G-2 in UGIB). The parameters include the mean and the standard deviation of the BTO residuals and the 1-hour BFO residual deviation (not the standard deviation). The other six statistics are Pearson’s correlation coefficients. I know these correlation coefficients are the primary source of the narrow peak, because if you don’t include them (and use just the three BTOR and BFOR parameters) the composite Route Probability does not show the narrow peak.
When I did the LNAV route fits, I had no idea such a narrow peak might exist. I was surprised, and somewhat amazed, to watch the Route Probability value during the route fitting process when the “additional” 5% in the narrow peak appeared and then disappeared as I stepped the value of just one parameter – the 19:41 initial bearing – through 180 degrees. I believe this narrow feature is real and informative, because you don’t see it when using any single parameter or a small number of parameters.
You will also note that the amplitude of the narrow peak is much larger than the apparent “noise” in the calculated Route Probability for nearby bearings. In addition, the observed structure in the narrow peak implies coherence, because there are multiple peaks like grating lobes. I concluded that the combination of six correlation coefficients has detected a unique, and faint, coherence (in bearing space) among them at 180.0 degrees bearing which is not present elsewhere. The SNR in one parameter is too low for detection using the limited data we have, but in a “coherent” detection process the combined (and enhanced) SNR allows optimum detection. The “detection process” is the Route Probability, and it is coherent (in bearing space) because I compute the product of the individual conditional probabilities. In other words, the Route Probability is effectively a coherent product detector. This coherence among parameters can produce multiple peaks in the product. The observed presence of this pattern implies that the source lies in multiple parameters, as one expects for a real feature. I don’t know how noise or spurious effects could produce the multiple sharp lobes, so I have concluded the feature is real and reflects actual route information embedded in the data.
@DrB
Many thanks for explaining (once more); I’ll reread some parts of UGIB (2020) with this information as back up. Two thoughts already now:
1. Did you check the (non)existence of this peak/pattern for all paths analyzed and is it therefore literally unique / for how many paths did you do this check?
2. In principle, knowing the main error characteristics for BTO and BFO, you could generate an artificial 180 deg path and then add (several times) an error signal to the BTO and BFO values, in accordance with the error statistics. This would generate several BTO/BFO data sets (for example 100 sets). Shouldn’t you then be able to (with being “blind” for the 180 bearing initially used in the datasets) to extract the 180 deg bearing by looking for your special peak pattern (for all or the majority of the 100 data sets!)? In such a way the method could perhaps be further tested.
@Niels,
1. The check I did was just for the routes near the peak at 180 degrees, as in Figure G7. I suspect the correlations also affect the Route Probability in a noticeable way at bearings well away from 180 degrees, but I have not tried that.
2. I believe what you propose as an experiment for retrieving the bearing would work, and it perhaps would increase confidence in the result. However, I am skeptical that we have a complete understanding of all the errors and noise sources in the real system. So, such a trial might still be criticized as not being a sufficiently accurate representation of the actual system to provide 100% confidence in applying the result to the MH370 data set.
3. What would be convincing, to me at least, would be to analyze a large number of actual flight data sets, rather than attempting to create them artificially. I tried but could not get enough data from other flights to perform a statistical analysis of bearing retrieval accuracy. The MH370 data set is at least rare, if not unique, because of the very long time without apparent turns or climbs.
4. I spent about 1,000 hours just to develop, test, and compute the route probability once for the entire range of bearings shown in UGIB. The fitting process is agonizingly slow, with manual operations being frequently required. I won’t do that again.
Neils,
Regarding your point 2: “In principle, knowing the main error characteristics for BTO and BFO, you could generate an artificial 180 deg path and then add (several times) an error signal to the BTO and BFO values, in accordance with the error statistics. This would generate several BTO/BFO data sets (for example 100 sets).”
That is exactly what I did in this report:
https://drive.google.com/file/d/1DTNe-4zh7FR8P0GNJhZLsW4RbTTsGzId/view
See Section V “Simulations”. I only generated 25 data sets, but they show that the most likely bearing is, indeed, significantly affected by noise in the data.
@DrB: Many thanks for further clarifications. I indeed think that the numerical experiment (point 2) could be worth exploring and apparently @sk999 already gave such numerical experiment a try (thanks for sharing!) -> I’ll have a closer look.
Regarding your point 4: I can really recommend to move form Excel based procedures to for example Python implementation. I revived my programming skills (mainly 6510 assembly based and Microsoft BASIC) in recent years through learning some Python and it is amazing how fast such higher level language can run on a modern computer.
On the side: cherry picking was a respected activity in the region where I grew up, and it was the way I earned money to buy my first computer when I was 15, 16 years old. I spent many weeks high up in the trees doing shifts of 10 -11 hours a day, to become a proud owner of a C64.
@DrB and UxIB:
I mentioned this before, but multiplying correlation results can be problematic, especially when they are from completely different origins. Say you have several different peaks from histograms. If even one is inaccurate, then multiplying it against all the good ones leaves an inaccurate result – a magnified value away from the real peak.
Summing/averaging the histograms makes more sense, like stacking in seismology – and imaging. Try multiplying a bunch of noisy image captures together instead of averaging. You just get more noise.
The coherence structure indicated by side lobes that you see is likely other peaks leaking through.
@ALSM: “Why would anyone use any intermediate WPs if you are headed for MEFE somewhere in the SIO?”
Note the “if” in your premise. It harbors a basic assumption of pilot suicide/murder.
A counter question might be: Why would a pilot quit flying between waypoints right after we lost the ability to track the plane, when it appears he had been using waypoint navigation for the entire flight up to that point?
@Niels:
I was sure your aside about cherry picking was going in a different direction. Whew! I’m from the same era, but started on an 8080. In my spare time, I wrote a disassembler in assembler. 😉 Python is much more productive.
@370Location,
Including the correlation statistics in the Fisher probability is not adding dominant noise to the UGIB Route Probability. That’s because, while the correlation coefficients are based on a small number of data points and are therefore relatively noisy, their noise contribution to the product is normalized by the large predicted standard deviation when the Z-statistic is calculated. Thus, while the correlation coefficient values are noisy, this is compensated when they are scaled to a common probability variable (first the Z-statistic and then chi squared), and their noise contribution is the same as all the other statistical parameters. In other words, the noise contribution of each parameter is equalized by using the Z-statistic (which has units of standard deviations).
Your example of multiplying by a noisier image does not represent the process I used, and your conclusions are not applicable. In my process, the probability function of each parameter (probability versus Arc 7 latitude, for instance) is normalized so they are equally noisy before they are multiplied.
Since they are each conditional, the probabilities of each parameter must be multiplied in order to accurately predict the joint probability. Then one can say the probability at latitude X is Z% of the probability at latitude Y.
Summing the individual probabilities cannot accurately represent the true route probability, and one could not estimate even the relative probabilities of two latitudes with such a method. For example, suppose at latitude Y the probabilities of two parameters are each ½ of the value at latitude X. If summing, you get a combined relative probability of (1/2 +1/2) / (1 + 1) = ½. When multiplying, you get (½ X ½) / (1 X 1) = 1/4, which is the correct result. So, the averaging method you proposed gives very large errors in the predicted relative probability. The summing method is appropriate for averaging multiple estimates of the SAME parameter, but even in this case each probability function must be scaled to normalize the noise before being added. In other words, in this case one should do a noise-weighted average, not a simple unweighted average.
@370Location: I agree that it is difficult to justify any particular path based on unknown pilot intentions.
On the other hand, you asked:
Why would a pilot quit flying between waypoints right after we lost the ability to track the plane, when it appears he had been using waypoint navigation for the entire flight up to that point?
If the intent is to hide the plane in the SIO, then flying along an airway like N571 in a direction towards the deep Indian Ocean and away from protected countries would likely not attract much military interest while within radar range. When beyond radar range, that constraint is removed, and the most efficient route, i.e., a great circle (actually, a geodesic has @sk999 pointed out) makes more sense.
@Victor
Re: Left Offset proposal
I am not sure if it could have been a left offset or alternately a jog over to path B466 to ANOKO (I fly it as the latter). I tend to view MH370 never actually joined N571 after MEKAR, trending a slight bit of N571 after MEKAR.
I have not worked out a definitive BTO/BFO proof for Left Offset, except to say, relooking the 1825-1828 BFO data, it seems that the LEFT/RIGHT offset direction is unclear. Timing to ANOKO seems to fit a turn south, which also might make logical sense as the edge of Indonesian FIR.
@DrB:
Your explanation is completely counter to my experience with beamforming techniques using multiple pairs of sensors to determine direction of arrival.
You are assuming that each of your latitude histograms has a broad peak that tails off to a high noise level, and that each histogram will agree with the others on the peak. That’s not the shape of your contributing probability histograms, but let’s take another approach.
Hypothesize that each of your probability histograms is near perfect, providing a single sharp peak at one precise latitude, with the noise floor going to nearly zero everywhere else. Yet, one or more differ from the others on where the latitude peak is. If you multiply your probability histograms, it doesn’t matter how many agree – the one peak that is off will mismatch all the others, multiplying its peak value by the near zeros of the others.
I made this same mistake early on when beamforming the hydrophone triad arrays. Each of the three pairs should agree on the Direction Of Arrival of a signal. Sadly, the coordinates of the hydrophones were inaccurate, so the DOA on each pair didn’t agree. Multiplying the three correlation peaks gave garbage results. I have since calibrated the hydrophone positions to now agree with DOA of known signal sources all around the SIO, but I still don’t multiply probabilities when deriving the beamforming peak.
(Hint: I’m actually having great success now with using kurtosis rather than simple summing in cross correlation or phase correlation. It is much better at comparing the shape of the signal vs noise distribution.)
@VictorI:
Again, the IF of an intentional flight to oblivion. We have scant evidence (a slow detent) of intent for the diversion. Many have chosen to think that the pilot was a mastermind of evasion – not only avoiding radar but flying exactly along FIR boundaries with right angle turns on the assumption that it would confuse radar operators, even hell bent on depositing his aircraft into the deepest hole in the SIO beyond Broken Ridge, eaking out the maximum range by depressuring the cabin to save a few kg of fuel to make it there.
More likely, no radar operators were aware of any anomaly. It took days for Malaysia to analyze their saved radar data and match it to MH370. Thailand says it agrees, but won’t show why. The pilot wouldn’t know that Indonesia didn’t detect his plane along its coast in real time, or that JORN wasn’t focused on him, or that the SATCOM left breadcrumbs. (Though some experts like JW think even that was anticipated and hacked to fool us all before BFO and BTO were documented.)
@All:
Sorry, but dubious statistical methods and debating about kg of fuel remaining as the current way to optimize the final latitude seems like arguing about the number of angels dancing on the head of a pin. If that’s all you have to go on, then do go on. I believe there’s additional firm evidence to consider, but it points way outside the current consensus.
@TBill said: I have not worked out a definitive BTO/BFO proof for Left Offset, except to say, relooking the 1825-1828 BFO data, it seems that the LEFT/RIGHT offset direction is unclear.
I don’t see how you can match the BTO and BFO with a left offset. To start, see if you can match just the BTO data.
@370Location: UGIB 2020 attempted to define the POI without considering pilot intent. But given that, one path between 18:28 and 19:41 was defined as a possibility, with intent included. That portion of the overall route doesn’t invalidate the POI, which was agnostic about intent. This has been a constant source of confusion.
@370Location
I worked a bit with the Z80 (P2000) which I think came after the 8080. On the C64 for example I wrote one of the most compact speed loaders (based on counting exact clock pulse to distinguish 1 and 0 coming from tape). All this low level stuff: Good training in logic and discipline. On the side a lot of fun games to play. Nice, simple pre-internet times.
@DrB: You wrote, explaining the occurrence of the “narrow peak”: “…The other six statistics are Pearson’s correlation coefficients. I know these correlation coefficients are the primary source of the narrow peak, because if you don’t include them (and use just the three BTOR and BFOR parameters) the composite Route Probability does not show the narrow peak.”
Now @sk999 in the paper he just shared wrote (fig. 9 caption):”Right: Auto- and cross-correlation statistics included in UGIB but not here. The only statistic that has significant sensitivity to 7th arc latitude is the BFO-BFO autocorrelation.”
A question to both Bobby and Steve: are you both referring to the same correlation coefficients here? And how can we explain the difference in your findings? Could it be that such “narrow peak” pattern could be found in each or most of the 25 trails that @sk999 performed, but that the indicated latitude would vary with the specific deposition of the BTO and BFO errors for each of the 25 sets?
@Niels,
You asked: “A question to both Bobby and Steve: are you both referring to the same correlation coefficients here? And how can we explain the difference in your findings? Could it be that such “narrow peak” pattern could be found in each or most of the 25 trials that @sk999 performed, but that the indicated latitude would vary with the specific deposition of the BTO and BFO errors for each of the 25 sets?”
In UGIB we used this one: “5. r for Leg Start BFOR to Leg End BFOR.” I don’t know exactly which BFO auto-correlation statistic Steve used. Maybe it is the same as UGIB’s #5.
Steve has not repeated the full calculation done in UGIB, so comparisons are not expected to exactly match. I have compared our SIO route probability results, and the agreement is generally good over the latitude range from 32S to 36S.
Trials of the same route can be done using artificial noise generators to simulate the BTO and BFO noise. I did not do that. I simply evaluated our FOM using a large number of statistics and using the actual MH370 BTOs and BFOs for the best-fit routes over the full range of crash latitudes.
I doubt you will see a very narrow peak (in probability versus latitude) for artificially generated data over a range of crash latitudes. If you use different data for each of those trial routes, I doubt you will find a correlation-driven peak, versus crash latitude, because none of the trials will have noise which is correlated over time in the same route or between different routes with different data sets.
What we did in UGIB was different from that. We used the SAME MH370 data for fitting ALL the routes. It’s possible the actual MH370 data has unrecognized low-level characteristics that won’t be present in artificially generated pseudo-data (such as BTOs and BFOs). It’s likely the MH370 BFOs will be correlated with time because of the OCXO drift. However, that in itself does not induce any latitude dependence of what is a time-dependent effect. So, it’s not the BFOs that are correlated with crash latitude. It’s the BFO residuals (i.e., the BFORs) which become correlated because of systematic route errors. Thus, the actual true route will have just the time-dependent (and slightly correlated, due to OCXO drift) BFOs, and the BFORs will show that same slight degree of correlation as the BFOs, because no systematic route errors are introduced when using the correct route. On the other hand, all incorrect routes will have systematic predicted-BFO errors caused by the route errors, in addition to the time-dependent BFO correlations caused by OCXO drift. Route errors can affect the BFO residuals on time scales of a leg duration and longer. Therefore, systematic route errors, which will be crash latitude dependent, induce additional auto-correlation in the BFO residuals. Steve also sees this effect, and I think it is simply the conversion of systematic route errors to more highly correlated BFO residuals on time scales of the leg duration (typically 1 hour).
@370Location,
You said: “Your explanation is completely counter to my experience with beamforming techniques using multiple pairs of sensors to determine direction of arrival.”
Your analogy is irrelevant. MH370 does not have multiple, time-coherent detectors.
You said: “You are assuming that each of your latitude histograms has a broad peak that tails off to a high noise level, . . .”
That is incorrect. I don’t make any assumption about “latitude histograms” or how the noise varies with latitude (it doesn’t, by the way).
In addition, I don’t have multiple “latitude histograms”. I have only one probability curve versus crash latitude since I use one figure of merit for the route probability.
It’s clear that the route selectivity is improved by using multiple statistics, each of which has equalized noise in our processing method, compared to any single statistic, which has coarser selectivity which varies from statistic to statistic.
It’s also important to understand that systematic route errors introduce correlations along single routes with time and also from route-to-route in space and time. That’s primarily what creates the “structure” in our route probability function. There is a noise component too, but for this MH370 case we don’t have multiple trials, so the measurement noise in the Inmarsat data is highly correlated (i.e., it is identical) for all our route fits. Thus, the high frequency structure seen in the route probability function is created by the completely correlated measurement noise (from route to route) and the partially correlated route errors (along each route and also from route to route).
@DrB:
Sorry if I wasn’t clear. I’m referring to the UGIB Mar 2020 paper where you take four normalized probability histogram data sets (Satcom, FE, drift, aerial search) and multiply them together for a final route pdf.
You have zero probability for the aerial search missing the debris at latitudes 30 and 32, and 10% nearby. Even though the other three datasets have strong values there, when multiplying by zero you get zero expectation of finding the plane at those latitudes, and very low values nearby.
Zero percent chance of missing the debris field seems extreme given the variables of movement by the time searches were conducted plus weather and visibility. It wouldn’t be an issue if the probabilities were summed, but you’ve excluded large tracts of the 7th Arc by taking the product of probabilities.
This caught my eye because drift probability is fairly good from near Java, it plausably fits satcom, there were no searches within hundreds of miles, and fuel range is not an issue. Yet, your results exclude it with zero probability.
@DrB
You wrote:
“I doubt you will see a very narrow peak (in probability versus latitude) for artificially generated data over a range of crash latitudes. If you use different data for each of those trial routes, I doubt you will find a correlation-driven peak, versus crash latitude, because none of the trials will have noise which is correlated over time in the same route or between different routes with different data sets”
I wrote:
“Could it be that such “narrow peak” pattern could be found in each or most of the 25 trails that @sk999 performed, but that the indicated latitude would vary with the specific deposition of the BTO and BFO errors for each of the 25 sets?”
I could better have mentioned “indicated bearing” as I was referring to your fig. G7 peak pattern. If you would use your G7 procedure to find the preferred bearing for each of the 25 trials, this would indirectly result in a range of indicated latitudes. The key question thus would be if for each of the 25 trials you would find approximately the same bearing and thus a narrow distribution in indicated latitudes. I doubt this, but the best way to find out is to apply your exact procedure to such a set of trials.
It is noted there are more measured BTO and BFO data points (I work with 12) than the number of degrees of freedom (initial bearing, initial lon, initial lat, FL, M; NB Steve mentions 6 degrees of freedom). In my analysis I work with the the arc 2 – 6 BTOs and the arc 3 – 6 BFOs + the 23:11 phone call BFO.
@Niels,
I think you are asking about a method whereby one artificially creates a set of pseudo-BTOs and BFOs while assuming one set of route parameters, applying accurate GDAS conditions along that route, and assuming the noise characteristics of the BTOs and BFOs. Then you compute the SIO Route Probability for about two dozen assumed initial bearings near the assumed value, preferably in a blind test, fitting all the other route parameters for each assumed initial bearing so that a multi-parameter Figure of Merit (such as the “Route Probability” used in UGIB Figure G-7) is maximized. Those fits will also have to incorporate the actual GDAS weather conditions along the route. The winds and temperature field have spatial and temporal correlations that produce common mode (i.e., correlated) BTO and BFO residuals for nearby tracks. This effect contributes to the structure seen in plots of the route probability. That fitting process will give you one plot similar to Figure G-7 in UGIB. Then you repeat that entire process 25 times. Now you have 25 plots, one for each set of artificially generated BTOs/BFOs, but all for the same assumed initial bearing. Then you observe the peaks in each plot and observe if any of the plots have a very narrow peak. You can also average the plots and see if that result has fine structure and a very narrow peak. Is that what you are suggesting?
An alternative method is to allow the initial bearing to be a fitted parameter and then find the value which maximizes the FOM for each of the 25 data sets. The distribution of the 25 fitted values will tell you roughly the precision of determining the initial bearing with one flight’s data. However, using the same flight model to generate the artificial data and to analyse the artificial data does not guarantee the accuracy of the “retrieval process” for estimating the true initial bearing using the actual MH370 data. For instance, errors in the flight model will not be detected. Thus, this second method basically only tells you how the data measurement noise propagates through the flight model to affect the estimated initial bearing. Both methods are incapable of detecting small systematic differences between the actual flight and the flight model.
If the first method were done (two dozen fits to each of 25 data sets), I doubt you would see a pattern similar to Figure G-7 if you use a simple statistical model, such as a gaussian probability distribution function, for the measurement noise of the BTOs and the BFOs. Figure G-7 has a complex peak structure which is created by at least three effects: (1) the weather is highly correlated among nearby routes, and that correlation decreases with distance and elapsed time, (2) the systematic route errors are correlated along the route and among nearby routes, due to the geometry of nearly straight flight paths intersecting circles (i.e., constant-BTO arcs) with time-dependent concentricity, and (3) the BFOs are correlated by uncompensated OCXO frequency drift. In UGIB we included the first two effects, which are known. We don’t know the last one. For example, look at Figure 5.5 – the BFOR PDF – in the DSTG’s “Bayesian Methods” report. The BFO residuals have significant deviations from a gaussian PDF. The same is true for the BTO residuals, as demonstrated in their Figure 5.1. I don’t know what caused these non-gaussian noise statistics, and I would be wary of trying to model them or even assuming they are intrinsically and identically present in the MH370 data.
@370Location,
UGIB (2020) uses the spatial averages near Arc 7 of the aerial-search probabilities of detection of the drifting surface debris field calculated by CSIRO. We did not identify any flaws in their method, and we did not modify their results. As I said before, summing conditional probabilities does not accurately predict the combined probability.
You said: “drift probability is fairly good from near Java”. This is incorrect. See Figure 12.1-1 in Ulich and Iannello (2023). The overall drift probability from the northern tip of Arc 7 is virtually zero, being low for typical floating debris and undetectable for the flaperon.
There is a condition on the UGIB predictions. It is that the SIO Route (19:41 to 00:19) was flown without significant turns. Without this assumption, it is impossible to make a prediction, using only route probability, with a searchable area near Arc 7. If turns were made, no route analysis is capable of identifying them with the flight data we have. That does not mean you can’t eliminate large portions of Arc 7, even if significant turns occurred. The fuel analysis eliminates south of 37S. The debris drift analysis eliminates south of 39S and north of 24S. The aerial search eliminates 30S and 32S. So, with no assumptions about the route, we are left with 25-29S, 31S, and 33-36S. Java is not allowed, even with no assumptions about the route to Arc 7.
@DrB
Right, your first paragraph describes approximately what I had in mind to better test the relevance of the “peak pattern”, except that I would not “average the plots and see if that result has fine structure and a very narrow peak”, I would be interested to see how wide the distribution of resulting bearings would be.Therefore, I would not necessarily do that “all for the same assumed initial bearing” for each of the 25 BTO/BFO sets. So that could mean a combination of the two approaches you describe would be required for this “experiment”.
I have to think a while about the implications of your further explanations/reflections in the third paragraph.
@DrB: “As I said before, summing conditional probabilities does not accurately predict the combined probability.”
Or, perhaps it doesn’t give the answer you’re looking for. The summed result would be a much broader range along the 7th Arc, with outliers. That could be a better reflection of reality.
I agree that much of the 2020 and 2023 papers don’t apply to the Java candidate. Aerial searches didn’t go N of 19S. Any turns are excluded.
We’ve discussed the drift before. Your chart shows a nil probability for a drift origin at Java, but that’s because the 2023 paper uses an even narrower window of arrival than the 2020 paper, again excluding tropical latitudes. Real drifters and modeled particles go from the Java site directly to where debris was found, arriving earlier that you allow. You now set a window of a few days for barnacled debris, which may not be realistic. There were only four pieces with barnacles, two pieces had many. You excluded one, “Roy,” entirely as an outlier. It doesn’t fit because even though it traveled the farthest of any debris, it “apparently” traveled the fastest. It was spotted in 2015 covered in barnacles, and again in 2016. It was the first major find after the flaperon.
We now know that the earliest estimates of log barnacle growth rate were far too slow. The flaperon barnacles were likely weeks old, not 15 months. News reports are that the flaperon had been spotted in the area months earlier. I’ve come to realize that each of the pieces with barnacles were found either in cold SA waters, or in the Mascarene Islands just near the coldest time of year in the Austral winter. Graphing satellite Sea Surface Temp shows that the nearby waters had dropped into the reproductive/growth range just weeks before the barnacled debris was found. All the rest of the debris were found in warmer waters with no barnacles. The ones that were analyzed found no evidence of past barnacle growth, only snails and algae.
Considering that debris continued to be found for years after 2014, it’s reasonable to expect that debris was not found within a narrow window of when it arrived in a locale, and that barnacles grew when the water temps at the locale favored the growth, including after arrival.
I see how the statistical methods applied affect your conclusions about my candidate, but I don’t accept “Java is not allowed” as gospel based on those methods.
Likewise, I doubt that the CSIRO aerial search index containing zero values near well-searched latitudes was intended as a single variable that could exclude large swaths of the consensus area still to be searched.
@370Location said: Or, perhaps it doesn’t give the answer you’re looking for. The summed result would be a much broader range along the 7th Arc, with outliers. That could be a better reflection of reality.
I think we agree that mathematically, UGIB 2020 properly treats how to combine probability distributions of independent data sets. However, I think the disagreement is whether the PDFs of some of the data sets should approach zero over the stated range.
@DrB
….you said “There is a condition on the UGIB predictions. It is that the SIO Route (19:41 to 00:19) was flown without significant turns.”
Can I expand this to say it also means: flown without significant turns and/or significant descent/slowdown (eg; before Arc6)?
@TBill,
Yes, the UGIB (2020) fits to the 19:41 – 00:11 Inmarsat data used a flight model with a route which had no turns and no major descents or slow-downs.
@370Location,
You said: “Or, perhaps it doesn’t give the answer you’re looking for.”
I’m not looking for any specific answer. We took great pains to avoid bias in our analyses, and we described exactly what we did and the answers we got. That’s how an objective science is supposed to work.
You said: “Your chart shows a nil probability for a drift origin at Java, but that’s because the 2023 paper uses an even narrower window of arrival than the 2020 paper, again excluding tropical latitudes.”
That is incorrect. The earliest arrival date analyzed was 1 day after the crash. No “early arrivals” were excluded. Some plots may not have had the abscissa origin at 1 day, but that’s because there were no predicted arrivals prior to the plotted dates. No arrivals were excluded from day 1 out to the maximum delay of the CSIRO calculations.
In both the 2020 and 2023 papers we used the full latitude range of CSIRO drift predictions – from 8S to 45S – for calculating the drift probability, and we analyzed all arrival dates over the full range of CSIRO predictions. No predicted arrivals were excluded.
@DrB: Relative to earlier arrivals, I think @370Location means northern latitudes were calculated to have a near zero probability because the predicted first wave of debris was well before the reported recovery date. He has made the observation before that we can’t know for sure when the debris actually beached, the presence of barnacles notwithstanding.
@VictorI, @DrB:
Right. We can’t know when debris actually arrived in an area based on when it was later found beached.
I think we’ll have to agree to disagree on the joint probability methods. I suggest summing these multimodal distributions for overall maximum likelihood, and exponentiate if you want a sharper peak.
I’ve taken another look at your 2023 drift report. I have to again conclude that it doesn’t apply to my candidate site. Page 11 of the report notes that “Roy” was excluded, presumably because of too few drifter hits. It appears that 12 locations with 17 pieces were excluded, and 17 analyzed.
All validation tests were only applied to latitudes 27-40S. Most of the focus of the paper was on sorting two peaks at 34S and 38S. It gets unclear which assumptions for abandoned Methods I and II were utilized in the final Method III.
Appendix A pg 67 discusses how multiplying by zero probability for drifter predictions is avoided by adding an additional bin count of one. Pg 69 talks about how there must be a minimum of two drifter location+time matches at each and every debris find location or an origin is excluded. (Again the problem of multiplying by zero when taking the joint probability of 27 flat PDFs).
The time window to include a match is that the predicted day must not be earlier than a median of 53 days from when it was estimated to be found. That is just not realistic. For example, the Rodrigues Island find was 11 months after the flaperon find at nearby Reunion Island. Such disparity can’t be explained by waves of predicted late arrivals from different eddy releases, which are considered matches when they are found very late.
Example graphs on pp 22 and 74 show that predicted drifters are at least 5 times more likely to arrive very early in a large group, and from latitudes 8-25S. Page 81 gives a PDF/histogram without the time constraint, showing that arrivals at 11S are just about as likely as those at 30S. It’s another multimodal histogram which is not selective for latitude, but it may indicate that drift really isn’t very selective for latitude.
Again, applying the late aerial searched probabilities to exclude search areas is questionable. Had the more recent and accurate underwater searched areas been included in the report and multiplied the same way, your result would be *all* zeros for the joint PDF result. The PDF on the last page 89 showing the combined probability without the aerial exclusion may be the most realistic for the consensus search area range. It still uses the dubious time selection exclusions for drift, and assumes no turns giving zeros for the other three PDF components, which make it inapplicable to my candidate near Java.
@370Location,
You said: “The PDF on the last page 89 showing the combined probability without the aerial exclusion may be the most realistic for the consensus search area range. It still uses the dubious time selection exclusions for drift, and assumes no turns giving zeros for the other three PDF components, which make it inapplicable to my candidate near Java.”
1. There is no page 89. Perhaps you meant page 83. Regarding aerial search, since none was done near Java, we used an aerial search probability of 100% [see Figure 14.5-1 in Ulich and Iannello (2023)]. That does not exclude or penalize Java.
2. Regarding the drift probability, Figure 12.1-1 in UI (2023) shows the probability at Java latitudes is essentially zero, based on the CSIRO drift simulations. There is no time exclusion since all latitudes are processed identically and over the entire calculation window. Very low probability events such as “Roy” cannot be processed to estimate crash latitude (as explained in that paper), even with the large number of predicted drift tracks available from CSIRO. Not including any single debris does not induce a bias in the crash latitude prediction. It only reduces the precision of the latitude prediction.
3. It is impossible to predict the probability of MH370 SIO routes by matching the Inmarsat data unless one assumes the condition there are no major turns or speed changes along the route. We don’t know yet whether this condition is true (and we won’t unless the FDR is found and the data are successfully recovered). However, UGIB and others demonstrated there is no necessity of turns or speed changes to fully explain the Inmarsat data. Since no simple route to Java is consistent with the Inmarsat data, our conditional probability for a crash there is essentially zero. That prediction does not prove that did not happen, because it is possible (although I think it is unlikely) the assumed condition was violated.
@Victor Iannello,
@370Location,
Victor said: “@DrB: Relative to earlier arrivals, I think @370Location means northern latitudes were calculated to have a near zero probability because the predicted first wave of debris was well before the reported recovery date. He has made the observation before that we can’t know for sure when the debris actually beached, the presence of barnacles notwithstanding.”
The flaperon arrival at Reunion is unlikely to have been missed if it beached at an earlier date, because the beach cleaners who found it made daily clean-ups. It’s also highly unlikely to have been stranded on coral reefs near the shore. Photographs of the rocky beach near Saint-Andre when the flaperon was found show no signs of shallow offshore reefs which could snag a floating debris. If a debris not stranded near the shore, then the presence of many live barnacles does indicate a very recent arrival, consistent with the finding date.
From Section 4.3 (Reporting Delay) in UI (2023): “The reporting delay is bounded by a range of many months’ duration for barnacle-free debris, but it is otherwise free to vary within its bounds because we don’t know the actual arriving date, only that it was found after an unknown but possibly considerable length of time. Barnacle-free debris are typically less effective in discriminating crash latitude because the arriving date is loosely constrained. Barnacle-free debris depend primarily on their finding locations to discriminate crash latitude, rather than on their arriving date.
We allow the reporting delay (delta) to be from 10 to 150 days for debris with no barnacles attached, from 5 to 30 days for the two debris we analyzed (D9 and D23) which were found with a few barnacles attached, or zero days for the flaperon (D2), which was found with many barnacles attached. Thus, the estimated arriving date has an allowed range of values which depend on the number of barnacles on the found debris.”
In UI (2023) we allowed the reporting delay to be up to five months for each latitude bin, so no latitudes were penalized because the reporting delay was long or uncertain.
Niels,
You asked, “… are you both referring to the same correlation coefficients here?”
I can only speak for myself, That was certainly my intent. I would use somehwat different language – e.g., where UGIB wrote “r for Leg Start BFOR to Leg End BFOR” [Appendix G1, #5], I would say that it was “BFO autocorrelaton at lag 1 step” where “step = leg”. To my mind, they say the same thing.
You also asked “… how can we explain the difference in your findings?”
Once again, I can only speak for myself. The best comparison between my findings and those of UGIB is given in the linked “updated route and fuel probabilities” from the DrB’s comment on Nov 15, 2023, Figure 5. While there are differences in detail, we both find a broad range in the probability distribution of the arc 7 latitude. My sharp cutoff at -35 deg is not due to lack of fuel (I agree that there is enough fuel to reach to at least -36) but rather an increasingly poor match to the BFO data. Upon further reflection, I have probably over-weighted the BFO data, so the cutoff at -35 should probably be more gentle than what I have drawn, which would give better agreement with the UGIB curve. These distributions are also broadly consistent with the results of my 25 simulations.
As far as the “narrow peak” in UGIB Fig 19 (repeated as Fig G-7), the description of this calculation is given as follows: “… we explored the sensitivity of the route probability to varying one flight parameter at a time, while the other flight parameters were constrained.” Without knowing what the constraints are, it is difficult to comment.
On a technical note, I use a gradient descent algorithm to determine the best-fitting route parameters, and while UGIB do not state which algorithm they used (other than that they used Excel Solver), I suspect that it was the gradient descent algorithm as well. A requirement for any such algorithm is that the objective function being minimized have a “smooth” dependence on the route parameters (and specifically, that the objective function be twice continuously differentiable.) That is not the case here. The culprit is the fact that we interpolate in the GDAS table in 4 dimensions at each step along a route. Such interpolation introduces discontinuities in the derivatives as one crosses cell boundaries. The proper procedure would be to fit cubic splines to the tabular data, but doing so in 4 dimensions is rather daunting. The net result is that, while the minimizer can arrive close to a true minimum, it sometimes ends up in left field. I used the good solutions to derive inital values for certain of the route parameters that improved the likelihood of converging to the true minimum for the remaining routes. UGIB have a worse problem because their equation G-2 involves an absolute value of the Z-score, which introduces a large discontinuity in the 1st derivative at Z=0.
In case you are interested, here is a detailed explanation of the equations that I use:
https://drive.google.com/file/d/1IjLUTD-vNuVzlfXRY-ajJF9aN1NBbM2a/view
@DrB said: The flaperon arrival at Reunion is unlikely to have been missed if it beached at an earlier date, because the beach cleaners who found it made daily clean-ups.
I agree. I was only trying to better explain what Ed meant.
The Java candidate site has many strikes against it, including complexity of the path and the disagreement with drift modeling results. However, I don’t exclude it with 100% certainty because we don’t really know whether there were pilot inputs and we don’t know for sure the accuracy of the CSIRO drift model, not to mention that we simply don’t know what we don’t know. What is attractive about Ed’s site is the precision of the location of the acoustic event along the 7th arc, which means it can be searched relatively inexpensively. For that reason, I don’t completely dismiss it, even though I think there is a considerable higher probability that the impact site is elsewhere.
Within the constraints of our assumptions, the available data sets, and the accuracy of fuel and drift models, I don’t know how we could have done much better.
@VictorI, @DrB:
1. The current links to the 2023 drift paper on this site redirect to a filename ending in _old.pdf which might explain the differing page numbers. I found no newer link. BTW, in the current _old file, the TOC page numbers are off by one.
2. The CSIRO drift patterns don’t show zero probability for a Java site, as you showed in figure C.3-1 using only debris locations. It your time constraints that result in zero probability.
If low probability of location matches is the reason only 17 debris sites can be analyzed, that may indicate that the drift model isn’t accurately predicting the paths to where debris was actually found.
3. I agree that the flight path to Java requires turns, which cannot be covered by your models. It’s why I said the report does not apply to the Java site. The candidate is based on very specific new acoustic evidence. Route optimization methods are not needed to search the site.
DrB: “The flaperon arrival at Reunion is unlikely to have been missed if it beached at an earlier date, because the beach cleaners who found it made daily clean-ups.”
That’s an odd assumption, that the flaperon would show up in the exact same spot months later. Roy moved 8 km in four months. The flaperon could have been stranded earlier at any of the nearby islands before being found at Reunion. It was reported as seen months earlier, but we don’t know exactly where.
@VictorI:
Thank you for acknowledging that the Java site has a relatively precise location that makes it worth checking, even if it doesn’t match previous assumptions.
I don’t quite get the need for such elaborate constraints in the drift modeling statistics. You mention available data sets. If the CSIRO drifter tracks are available for another meta-analysis, I’d be curious to run them through python Pandas for gathering some simple histograms using all of the daily data weighted by the inverse distance from where debris was found. Grouping by transit speed might also be interesting.
@370Location: The link to the recent paper has been corrected to what should be the most recent version of the paper. (I am verifying with Bobby that this version is correct.)
The links to the CSIRO drift results can be found at the bottom of this article which discusses the UGIB 2020 paper:
https://mh370.radiantphysics.com/2020/03/09/new-report-released-for-mh370-search/
@sk999,
@Niels,
For the UGIB sensitivity studies shown in Figures G-5 to G-7, the parameters which are fixed are listed in each figure under the title. These are fixed at the optimized values shown in Table G-1. Parameters not listed were allowed to vary. For example, in Figure G-5 (the longitude sensitivity study), the fixed parameters are LNAV at 180 degrees, FL 390, and LRC. Therefore, as the longitude of the route is fitted and plotted, the 19:41 latitude (not listed) is also allowed to vary for each fit. You must do that or you are not isolating the longitude sensitivity. In Figure G-6 we see the 19:41 latitude sensitivity with all other parameters being fixed. In Figure G-7, we plot the bearing sensitivity. When you do this, you have to allow the 19:41 latitude and longitude to vary as well as the bearing, because you can’t fully optimize the bearing unless you also allow the starting point to float. In G-7 the only fixed parameters are LNAV, FL390, and LRC. Thus the 19:41 latitude and longitude are allowed to vary for each of the bearing fits. The only way to isolate the bearing sensitivity is by excluding non-optimum starting locations.
I am gratified to learn that Steve’s slight tweak on the BFO probability produces even better agreement between our predictions of overall SIO Route probability.
Steve is correct that fitting up to six variables with complex figures of merit is not for the faint of heart. Good initial guesses help a lot, as well as using both forward and backward derivatives. Sometimes I also used the “Multistart” feature in the EXCEL SOLVER, which generates multiple nearby starting points that can detect a global minimum you can’t otherwise reach with the maximum-descent algorithm.
@All,
We know that SIO routes using LNAV are flyable with the available fuel at least as far down Arc 7 as 36S.
However, as sk999 recently and correctly pointed out, for Arc 7 latitudes north of 37S, a delay is necessary to connect the post-18:28 location near N571 to the fitted SIO route. That delay is about 20 minutes for the 34.2S UGIB SIO Route. For the 36S SIO Route, a delay of about 8 minutes is necessary to connect the FMT Route to the SIO Route. Taking into account the multiple means of satisfying the 18:40 BFOs, I have concluded that three turns would be needed to transition from the N571 right-offset path to the LNAV SIO Route at MRC. It’s not possible to achieve even a marginally acceptable 36S SIO Route probability without having at least three turns or two turns plus a HOLD. I also note that the Route probability is significantly lower at MRC than it was at LRC for 36S (as plotted in UGIB). At MRC the BTO residuals increase to about 55 microseconds RMS, based on my modeling results. If anyone has achieved much lower BTO residuals at MRC, please post your route particulars. This apparently poor BTO fit substantially lowers the MRC SIO Route Probability compared to that of LRC (which has about 34 microseconds RMS BTO residuals). Thus, there is a significant penalty in the BTO fit caused by reducing the speed to MRC so sufficient fuel is available to fly until MEFE. This reduces the SIO Route probability at MRC compared to LRC. So, if my MRC route fits are reasonably optimized, the 36S route is flyable but suffers from a poor BTO fit, and it requires a fairly complex FMT Route.
@sk999
you say…”My sharp cutoff at -35 deg is not due to lack of fuel (I agree that there is enough fuel to reach to at least -36) but rather an increasingly poor match to the BFO data.”
That is interesting because for example the IG historic path to 37/38s eg by FFYap shows excellent match to BFO, except maybe Arc6 may be where that 37/38s path starts to diverge from a best BFO match. Part of the problem critiquing 37/38s path believers, is the match to BFO is so darn good at least to Arc5.
@VictorI & all,
Thanks for the link to David Griffin’s CSIRO drift model data cache.
I’ve taken a deep dive into tracking where the two datasets match up with where debris was found. One is labeled “flaperon”, and the other “non-flaperon”. Within the metadata there are further classes of rounded debris that catch windage vs honeycomb and flat materials that drift more slowly with the surface currents.
CSIRO didn’t randomize windage, instead they varied the starting points with some radius from the 7th Arc, and split the dataset into windage and flat segments.
With those 2x 86,400 tracks, we can see how close they come to actual debris finds. For each day of drift, I calculated the distance from every modeled drifter to the discovery locations. Assuming poor accuracy (to avoid weighting a few good matches), I took the inverse of the distance from each drifter to every discovery site over time, and summed the results.
I then binned the results by latitude of the starting points. This plot is for 30 lines per degree of latitude:
https://drive.google.com/file/d/1DHOKxmH94c2Aqipq1IwplRcHSqJz1Ztb
Green is for the high windage flaperon drift dataset, and red is for flatter debris that drift more slowly. When a drifter stops moving, like it’s stranded, the proximity is zeroed.
The graphic is fascinating. I expected to to see waves of arrivals from debris arriving each grounding site. Instead there are peaks as debris arrives at multiple sites. It’s clear that slower debris wanders much longer with more southern latitudes. Below 32S, most of the low windage items are beyond the 3 yr calculation window.
There is a curious quirk at a narrow band around 30S, where all types of debris arrived within 200-400 days of where debris was found. Around 22S there is a mimimum, where all debris traveled fastest to discovery sites. From 23S to 32S, there is a wide band where flaperon-like debris lands, but slower debris drifts for years in proximity to the sites. South of 32S, there is a common narrowish band of arrival for the flaperon dataset, but non-flaperon drifters never approach the discovery sites.
The plot near the Java candidate at 8.32S shows a good mix of flaperon-like arrivals at 300-400 days, with non-flaperon arrivals landing around 500-600 days.
This plot might be improved by breaking out each unique discovery site (20) by latitude. I’m looking forward to any further suggestions, or insights.
— Ed Anderson
@Ed
Interesting. Victor’s critique of Prof Chari’s model is lack of wind effects. Assuming that is true, Chari’s model does show the debris timing approx as observed.
@TBill: You are misstating my “critique” of Chari’s drift model. To be clear, I don’t know the details of Chari’s model. I am only hypothesizing that neither windage nor Stokes Drift are modeled because his transport speeds are lower than CSIRO’s. It would be better to know the details of his model, but so far we can only guess. I have no idea why this remains a big secret, but it makes it very difficult to compare and contrast results.
@370Location said: This plot might be improved by breaking out each unique discovery site (20) by latitude. I’m looking forward to any further suggestions, or insights.
This will make it much easier to understand the plot.
@DrB
You wrote: ”I also note that the Route probability is significantly lower at MRC than it was at LRC for 36S (as plotted in UGIB).”
I apologize for not being well familiar with the route fitting process, but the UGIB route PDF shows almost the same probability for some parts around S36 as the peak probability around S34.3. Does this mean that that for any endpoint latitude (e.g. S36), changing the speed changes the starting point (19:41) and probability? And that the original probability was for the most probable speed without separate consideration about the fuel demands for that particular latitude endpoint?
@370Location,
You said: “South of 32S, there is a common narrowish band of arrival for the flaperon dataset, but non-flaperon drifters never approach the discovery sites.”
Your conclusion about the BRAN2015 predictions is incorrect, as demonstrated by the several CSIRO reports on debris drift, as well as by UI (2023). All discovery sites in UI have close approaches from crash sites south of 32S for non-flaperon debris. More specifically, Figure 10.1-1 in UI shows examples of trial drifter Tracks from 34S arriving at all the debris locations analysed. Thus, your statement that “South of 32S . . . non-flaperon drifters never approach the discovery sites” is incorrect.
I suspect your error is mostly driven by your assumption of a figure of merit which ignores actual arrival dates and includes non-zero contributions for every day of every drift track. This allows a large number of days with large distance errors to contaminate your FOM. Your assumed FOM, which is the inverse of the distance from a finding location, introduces “smearing” and bias in both the crash latitudes and the arrival dates. Your FOM can’t tell the difference between having a very large number of days with very large miss distances and a smaller number of days with smaller miss distances. Common sense says that an analysis method which cannot do this is unlikely to be effective in predicting crash latitude. In UI we avoid this deficiency by allowing only one day from each trial, when the miss distance is the minimum.
To compare and contrast your FOM with UI:
1. Your FOM over-weights very small miss distances, which are smaller than the geographical location error of the drift model. That is, within the model localization error distance, the probabilities are close to being equal, because there is no basis in the model generating the drift tracks for distinguishing a difference in probability between a 1 km miss distance and a 10 km miss distance, for example. That’s why in UI we use a “flat-topped” miss distance window with a radius of 10-56 km. Within that window, we consider the probability of arrival to be equal, because we can’t know any different as a result of the drift model localization error. Section 11.1 in UI discusses the Bayesian PDF we used for the BRAN2015 localization error.
2. Your FOM over-weights trials with slower average speeds because you reward every day along a drift track at every finding location. In fact, there is only one debris per trial, so the maximum possible number of arrivals is the same for all trials – exactly 1 – independent of the average speed. In addition, your probability function biases the arrival times to later dates because it rewards slow tracks which don’t exit the calculation window.
3. Your FOM over-weights tracks with very large miss distances because your assumed FOM has too slow a cut-off with miss distance. Your FOM means that many days at large miss distances can overwhelm a smaller number of trials with small miss distances. The result is a skewed and smeared probability plot which ignores the actual finding dates of debris and which has reduced crash-latitude selectivity.
@eukaryote234,
You posed some good questions.
1. You asked: “Does this mean that that for any endpoint latitude (e.g. S36), changing the speed changes the starting point (19:41) and probability?”
Yes. Changing from LRC to MRC, for example, reduces the speed by about 2% and therefore the 19:41 location must be about 2% closer (farther south) to the end point at 36S.
2. You asked: : And that the original probability was for the most probable speed without separate consideration about the fuel demands for that particular latitude endpoint?”
Yes. The UGIB results for (maximum) SIO Route Probability assumed there was adequate fuel. See Figure 5 in UGIB. It shows the route probability was still fairly high circa 39S, for example, where the fuel probability is essentially zero. At 36S the route probability was high (at LRC), but the fuel [probability was low (also using LRC). What we later learned was that there is a slower speed (MRC) which is flyable to 36S with the available fuel. However, recent route fits show the route probability for the 36S MRC route is actually poor compared to LRC. That’s because of the higher BTORs (BTOR = BTO Residual). The expected BTOR standard deviation is 29 +/- 10 microseconds. The 36S MRC Route is about 52 microseconds, so that is about 2.2 sigmas above the expected value for the True Route. The probability that this value (and higher) is due to measurement noise alone is less than 2%, so it is much more likely to be caused by systematic route errors. That’s why I previously concluded the overall probability is low for 36S because at LRC the route probability is reasonably high e but the fuel probability is low. Using MRC does not help overall – it has a good fuel probability but its route probability is very low.
@DrB:
I should not have said “Never” about the the non-flaperon particle drifters never approaching the discovery sites. I was looking at the lower probability at far south latitudes, and noting that many of the low windage set appeared to be past the 1028 day window compared to the faster flaperon set.
Your fig 10.1-1 of course shows 34S arrivals, because you have selected the data that way, excluding any tracks that don’t reach all the discovery sites twice.
True, you can see that there is a weak contribution from distant sites. It’s noticible as a fog on the left edge of the plot, brighter at the bottom where the roaring 40’s latitudes on the are farthest west.
I think you misunderstand that there is no smearing of arrival times, or any time sensitivity at all. Any integration over time is the slight fog on the left where no arrivals are possible, which merely added to all values and easily distinct from nearer proximity contributions.
1: I’m not doing a simple inverse of distance that would indeed give infinite values on an exact match. I take the inverse of each distance plus an uncertainty estimate. I started with 100km, which is actually larger that your 10-56 km cutoffs. For kicks, I did try setting a value as low as 10km, and got very sharp spikes over time. In fact, the graph I shared used an uncertainty of 500km, because there’s little difference between 100 and 500 km. Using a broader uncertainty is compensated by raising the proximity result (FOM) to a small exponent for contrast. At 500 km, the exponent is 1.4.
2&3: There is little if any bias toward slower debris. I suspect just the opposite. For any one day, the plot is simply showing how close the unbeached particles have gotten to all of the discovery sites. Look how the faster debris in green arrives in sharper waves. The CSIRO non-flaperon dataset in red are showing a similar pattern of arrival, but generally delayed from the green and more stretched over time. They may have been more likely to wander in eddy currents. Farthest south, many probably went east past Oz. (I could try a separate proximity plot for debris going east of the coast). If theres a bias, it might be that modeled debris is more likely to beach in clusters on a broad coastline of Madagascar or Africa rather than a small island.
Before seeing your response, I was working on a slight modification of the algorithm. Instead of zeroing proximity only when a particle has beached, it is now also zeroed when a drifter is receding away from a discovery site. This brings out some detail about debris caught in eddy currents that I had suggested in 2020, where debris from the lower arc wanders until it is caught in the SEC, and the higher latitudes head straight west in the current and then get caught in eddies. Here’s the new intensity plot:
https://drive.google.com/file/d/1DKIZXy05kpnkK5taC2UzxkUDSPJbJLvJ
I’ve also computed a histogram summing the two datasets:
https://drive.google.com/file/d/1DHPKOUohi7OrqLIaYaPJUj5ML9LPv-gl
Oddly, my histograms are showing results almost opposite to yours, with a null between 32S-35S in both datasets. To emphasize my point about multiplying low probabilities, I’ve done just that for only the final two summed sets, labeled as “PDF product”. You can see that for certain latitudes (like an odd notch at 33.4S), the result goes from likelihood of 0.14 down to 0.024. I don’t think that matches reality any more than the CSIRO model accuracy, or the very large peak at 30.2S. If I used fewer histogram latitude bins, it would smooth out. The CSIRO dataset was randomized by starting longitude in a swath around the arc. If I instead either narrowed the swath or computed a more accurate starting latitude by distance from the arc, we might get a slightly different answer.
I combined debris finds that had duplicate discovery site lat/lon if they arrived at different times, but didn’t group by wider neighborhood. That seems correct, but tiny changes might shift the result.
As I mentioned in the past, the CSIRO tracks appear to split more north of Madagascar than other modeled drift studies using randomizations of windage.
A key difference between your approach and mine is that I am using the entire datasets. None of the 86,400 drifter tracks are excluded, and they are given equal weight. It seems like common sense to me that nowhere along the 7th Arc should the drift or search probability be cut off to zero. My charts don’t favor my own candidate site with the largest peak of all, but they do show that the probability is increasing again toward Java, with a peak there that is higher at 8.32S than any value north of 23S.
(BTW, the unoptimized python code is using a single CPU core to crunch the results in 96 seconds, so iterative changes are fairly quick). Making separate plots for each discovery site should show unblur the waves of arrival.
@370Location,
1. You said: “I should not have said “Never” about the the non-flaperon particle drifters never approaching the discovery sites. I was looking at the lower probability at far south latitudes, and noting that many of the low windage set appeared to be past the 1028 day window compared to the faster flaperon set.”
You appear to draw the conclusion that arrivals after the calculation time window closes somehow lower the drift probability relative to other latitudes. That is not the case. The drift probability is the PDF, as a function of assumed crash latitude, that debris drift trials are predicted to MATCH the finding locations and the ranges of arrival dates. Even if a particular latitude had predicted arrivals after 1,028 days (which we cannot know, since the tracks are not calculated) that does not affect the probability of matching the MH370 debris reports. We start with an equal number of trials at each crash latitude bin. Then, for each finding location, we count how many of those trials match both the finding location and the range of possible arrival dates. The crash latitude which has the highest number of matching trials has the highest probability. This calculation does not measure, nor does it require us to know, how many trials might have arrived after the end of the calculation window. Since the calculations end at 1,028 days, we can’t know how many might have arrived after that date. We can count those which are still adrift, not having arrived anywhere, but even that number is immaterial to our calculation of drift probability. The property we want to know is what fraction of the starting number of trials arrived at finding locations within the allowable time windows. The area-normalized variation of that likelihood with latitude is the drift probability PDF. Your conclusion that your observation (“many of the low windage set appeared to be past the 1028 day window”) somehow infers lower relative matching probability is incorrect.
2. You said: “Your fig 10.1-1 of course shows 34S arrivals, because you have selected the data that way, excluding any tracks that don’t reach all the discovery sites twice.”
I don’t know where you got the notion that figure shows tracks which reach the discovery sites twice. The UI text in Section 10.1 says: “In this figure we selected the one trial per debris site which gave the closest match in the time and distance windows.” The purpose of this figure is simply to show that CSIRO drift tracks exist, which have predicted arrivals consistent with all the debris reports we analyzed, and which start from the location of the UGIB LEP. No debris site in UI has zero predicted arrivals from 34S.
3. You said: “I’m not doing a simple inverse of distance that would indeed give infinite values on an exact match. I take the inverse of each distance plus an uncertainty estimate. I started with 100km, which is actually larger that your 10-56 km cutoffs.”
Equal distance weighting within a cut-off range made sense to us, and that’s why we did it in UI. Our distance limits were optimized for each finding location, and they ranged from 18 – 55 km. If I correctly understand what you wrote, your FOM is now equal to a constant divided by the sum of the distance and 100 km. Therefore, your FOM at 100 km is half of what it is at zero distance. That’s certainly better than using just the inverse distance, as you first described. However, the 100 km is still larger than needed for good trial statistics, and this reduces the latitude selectivity (i.e., it “smears” the crash-latitude PDF).
4. You said: “I think you misunderstand that there is no smearing of arrival times, or any time sensitivity at all.”
There is smearing and bias in your crash latitude plots because (a) you allow all possible arriving times with no temporal selectivity and (b) you combine all finding locations in a single plot with no optimization among the finding locations. You don’t even disallow predicted arrivals after the finding dates. If you allowed only trials which match the plausible ranges of arriving dates at each finding location (i.e., use a time window which is unique for each finding location), you get much improved latitude discrimination (which is less” smeared” and less biased in crash latitude). The significant advantage of employing both spatial and temporal discrimination versus only spatial discrimination is demonstrated in Section C.3 in UI: “The lack of selectivity when using debris reporting locations only is caused by the fact that most debris from Arc 7 are carried westward by the combined West Australia, South Equatorial, and East Madagascar Currents and so end up in mostly the same locations. The more important discriminator is the variable length of time required to reach the westward currents from different parts of the arc. Therefore, the arriving times add significant information which enable a precise POI-latitude determination that is not possible with only several dozen debris recovery locations and no arriving times (as demonstrated in Figure C.3-1 above).”
5. You said: “My charts don’t favor my own candidate site with the largest peak of all, but they do show that the probability is increasing again toward Java, with a peak there that is higher at 8.32S than any value north of 23S.”
The four “likelihood” plots you showed in your second link show values at 8.32S which are much LOWER than likelihoods at latitudes SOUTH of 23S. On that point we can agree.
@DrB:
You mentioned that my first graphical plot was biased toward slower debris. That wasn’t true for the intensity of the proximities, but I believe it is true for the now summed histogram. Fast tracks where debris went straight to discovery sites and beached are prominent in proximity but low valued blips on the histogram. Slow particles S of 23S end up drifting around near the debris sites after the discovery, which I agree is time sensitive and shouldn’t get positive weighting. Debris can’t be discovered before it arrives.
I applied a 60 day linear taper after debris had passed a discovery site date, to allow for uncertainty in the model. It mostly changes proximities below 23S, putting dark bands on the right side of the plot after 850 days. This only on drops those histograms by a few percent.
I also took your suggestion, narrowed the distance uncertainty to 50 km, and made the summary linear. There is a higher fog factor from distant debris mostly visible on the left, and it is now more “smeared” with fewer proximity peaks:
https://drive.google.com/file/d/1DVtjQBuPZaX3NM0TZDUJfb6nr6ud5808
The histogram is flatter with lesser peaks:
https://drive.google.com/file/d/1DUyQQroQt_NB-GS7D2VmwmVva8f7tuda
DrB: “I don’t know where you got the notion that figure shows tracks which reach the discovery sites twice.”
I’ve read all your papers, and I know you’ve put over a thousand hours into the drift studies. As I said, the data selection seems overly complicated, and it has changed over time. Perhaps I’m losing track. In your 2023 Appendix A.5 (referenced in A.7a) you say:
“a) the minimum number of trials simultaneously in both the distance and time windows is at least 2 per latitude bin”
Your 2020 coauthor wrote a 2021 drift paper that appears to detail similar methods, results, and shared content. He notes in section 7:
“”
(a) The number of trials arriving at the debris location within the distance window must be ≥ 100.
(b) The number of trials arriving within both the distance and the time window must be ≥ 20.
(c) It must be possible to identify a single mode across all crash latitudes.
(d) There must be a crash latitude bin count ≥ 4 trials in the resulting crash latitude bin.
“”
(distance window 30km for exclusion)
I may be confusing your previous methods with current methods I, II, and III.
DrB: “The four “likelihood” plots you showed in your second link show values at 8.32S which are much LOWER than likelihoods at latitudes SOUTH of 23S. On that point we can agree.”
If we are comparing specific candidate sites, I’ll note that on my histogram result the peak at 8.32S is 50% higher than the range around 33-34S.
My approach to a metadata analysis may not be as mathematically rigorous, but it is simple, uses common image and signal processing methods, makes use of the entire dataset, and does not exclude other candidates.
It is not designed to be selective for any outcome.
I try to follow the scientific method, in this case running an experiment to see if it invalidates my acoustic hypothesis. You’ve been claiming that only your site has a search probability match, and that your analysis makes my candidate and others impossible. I still don’t see that as valid, and I’ve pointed out why.
Again, because my candidate is based on new acoustic evidence, it is not dependent on being the highest peak on an optimized search probability curve from inexact data with narrow assumptions. It only needs to be plausable. I believe it is.
@370 Location,
First you said: “Your fig 10.1-1 of course shows 34S arrivals, because you have selected the data that way, excluding any tracks that don’t reach all the discovery sites twice.”
Then you said: “Perhaps I’m losing track.”
I think so. Figure 10.1-1 showed (as explained in the introductory text) just one track per debris site, to demonstrate non-zero probabilities for all MH370 debris reports from 34.2S.
Appendix A.5 in UI (2023) says: “To assure the statistical noise in a PDF (computed using one of the probability equations listed above) is not excessively high, for the non-MH370 validation test cases we applied two conditions using Method I over a 3-degree wide region of interest (ROI) which is centered at a predicted POI latitude:
a) the minimum number of trials simultaneously in both the distance and time windows is at least 2 per latitude bin, and
b) the average number is at least 5.”
This lower limit of two trials assures the noise in the Method I optimization route is not excessively high. That’s a different kettle of fish than Figure 10.1-1, but it also proves the fact that drift tracks from 3.4S provide significant probability for all debris reports, because these noise-related conditions were satisfied for 34S and all nearby latitudes.
You said: “If we are comparing specific candidate sites, I’ll note that on my histogram result the peak at 8.32S is 50% higher than the range around 33-34S.”
Your plot has significant deficiencies which contaminate and bias your histogram. Most importantly, it fails to fully consider the arrival times at various sites. As UI pointed out, time is the dominant discriminator for latitudes south of 30S. So, your histogram plot, which, as far as I can tell, only penalizes sort-of-late arrivals and does not reject any early arrivals, is strongly biased toward nearby crash sites which have predicted arrivals which are generally too early. Your histogram is therefore not useful in comparing the northern end of Arc 7 with the south end of Arc 7, which is the comparison you are trying to make, because you don’t remove those arrivals which are too early (and which, for the flaperon, is all of Arc 7 north of about 27S).
You said: “Again, because my candidate is based on new acoustic evidence, it is not dependent on being the highest peak on an optimized search probability curve from inexact data with narrow assumptions. It only needs to be plausible (sic). I believe it is.”
I don’t agree. It is implausible because it has extremely low probability of matching the MH370 finding locations and dates. You will see this (as CSIRO, UI, and others have already published) if you properly eliminate both late and early arrivals. In particular, 8S is strongly excluded by the Flaperon discovery at Reunion. So far, the only explanation is that the flaperon was somehow stopped by another intervening island for a year before magically continuing on to reach Reunion on day 508. I don’t think that is sufficiently likely to have occurred to make 8S a plausible MH370 crash site. It’s also a highly unlikely crash location for the other sixteen non-flaperon debris.
@All: I apologize for the disruption of access to the website, but it was due to a Network Solutions issue that affected many sites. This is yet another reason to completely avoid doing business with Network Solutions.
@Victor
From My Modern Safari mobile browser each refresh of your website or click on any link within the website takes approx 11 Seconds to display. Occasionally longer to display making my device inactive and activating the 30 second auto lock I have on. Even when I’m close to my WIFI router. No problems with any other website, or apps. Your actual website resources seem very basic in terms of display and usage. Seems like a Network Solutions traffic flow issue perhaps.
@Joseph Coleman: I am running two other WordPress sites on GoDaddy of similar complexity to the two on Network Solutions. Network Solutions is many times slower. I was told by one technical support person that the slowness is a known issue and engineers are working to fix this. Others deny there is a problem, such as the Network Solutions spokesperson I interacted with recently on Twitter, who said the problem was my fault.
It may be that I have not optimized the site, but I am not doing anything complicated that would account for it being so slow.
If this was a commercial site that I was using to conduct business, I would have left long ago. However, considering the niche nature of this site, and the time and cost to migrate it to another provider, for the time being, I am putting up with the horrible service, and strongly advising others to go elsewhere.
I follow this Blog from time to time and read the articles. Always great work by all participants. I posted to it in July 2018, June 2019, and March 2023 with @TBill, @Victor, @DrB, and @DennisW.
My concern is that a pilot that wants to hide an airplane would NEVER fly a great circle path to fuel exhaustion. This greatly reduces the possible end point from infinite to half infinite.
I would use S-turns or holds along the way to decouple range and endurance.
But with the pilot not knowing about the Inmarsat pings, the maneuvers had to have been a lucky set but it is possible that fuel could have been consumed along a course to a crossing of arc 7 at 25S, for example.
I do not think a pilot could have preplanned a maneuvering route to put the aircraft in the water and plan to meet the exact Inmarsat data, but a pilot could have planned to exhaust fuel at a specific location and maneuver to decouple range and endurance and it just happened to result in the data set.
I do not have any reason to question all of the simulation and analyses that have been done by this group. It just seemed off to me that a pilot would not actively work to hide the aircraft and would just take a great circle route to oblivion.
@Hank,
Nobody can “know” the MH370 pilot’s motivations. However, there is still a slim chance we can know his actions, and this might illuminate his thinking, although faintly.
You are going out on a limb when you say that a pilot would “NEVER” do such and such.
I won’t go nearly that far. I will say that it is highly unlikely the pilot know about the Inmarsat data being archived. Nobody at MAS knew this, and not everybody at Inmarsat knew it. If the pilot knew it, he could have prevented it by not repowering the IFE and thus shutting down the SDU. The fact that he left power on the IFE implies the pilot did not think this allowed aircraft tracking. Then, if the pilot “knows” he can’t be tracked, and especially if even the starting point of the SIO route is highly likely to be unknown to searchers, then his crash location is unknowable, no matter what LNAV mode he used or whether he made additional turns en route. So, he may very well have concluded that his flight could not be tracked, and how he flew into the SIO was immaterial. He was almost right about that. He probably thought about floating debris eventually washing up on distant shores, but that could not be avoided, and that would be unlikely to produce, by itself, a pinpoint crash location.
What we now know is that, if LNAV were flown following a geodesic to a single distant waypoint, that could not have been further south along Arc 7 than circa 35S, and even to get that far requires two fuel-saving measures. One is a reduced speed and therefore a reduced fuel flow from roughly 18:28 to 19:41. The second is reduced fuel flow (but not reduced speed) from 19:41 to MEFE at 00:17 with the air packs off.
We don’t yet know if this happened. The alternative explanation is a synthesized route with a reduced average speed (which removes the necessity of having the air packs off) and multiple speed/bearing changes designed to match the BTOs and BFOs. Numerous routes, which are far from unique, can match both the BTOs/BFOs and MEFE. However, they all end up on Arc 7 well north of 32S where the debris drift probability is generally quite low. Some of the Arc north of 32S is also eliminated by the aerial search non-detections of a floating debris field.
A third possibility is a route similar to UGIB but with very minor bearing or speed changes. I don’t give this possibility much credence, because I would expect a purposeful “evasion maneuver” to involve noticeable turns, and when you include this you must end up north of about 32S, where the debris drift probability is generally low. There is a small chance for a multi-turn route ending near 27S with a relative drift probability of about 1/3 and a relative aerial search probability of about 1/10 of 34S values. So, there is a few percent chance of a crash circa 27S.
You said: “It just seemed off to me that a pilot would not actively work to hide the aircraft and would just take a great circle route to oblivion.”
My guess is the pilot reached his “oblivion” circa 19:41, less than 2 hours after murdering the passengers and other flight crew members. Even if that suicide happened circa 19:41, that would not prove no turns occurred later. They could have been programmed into the FMS beforehand. The inclusion or exclusion of turns in the SIO Route does not depend on having a functioning pilot after 19:41, but a post-MEFE manual glide does. I don’t think this glide occurred because (a) we have no indication of a water ditching based on analyses of recovered debris and (b) the satellite data are fully consistent with an unpiloted crash after MEFE. The final issue is why didn’t the previous searches locate the debris field? The answer to that may be for multiple reasons. As Victor has suggested, the debris on the sea bottom may consist of smaller than normal pieces, and they may lie in difficult terrain that has not yet been thoroughly investigated.
@Victor:
Cloudflare is a free and easy to manage content delivery network. Their free account can cache your website content and deliver it fast globally. All it takes is transferring your DNS to them, which is also free. After a quick setup, your NetSol site will be hidden from hack attacks, and Cloudflare specializes in blocking them through their service. Just take your current DNS settings for mail and such, and copy them over. A bonus is that CF also handles the SSL layer and certificate. Your Netsol site then only needs to deliver http to CF. If you want to get fancy there’s a WordPress plugin to further optimize the CF cache.
@DrB:
To answer your disputes of my meta-analysis, I’ve refined the distance weighting to now be a linear taper of a set window size to zero. No more fog. I’ve done the same for the late and early=reporting penalties. As an extreme test, I’ve shortened the window to just 16 days, which reveals all the arrival waves. It has little effect on the shape of the latitude histogram. It just gets spiky/noisy due to fewer samples. Time plot:
https://drive.google.com/file/d/1DaFM5crmAsEYUV-kzdBX_HHnI_-kUAPq
Histogram:
https://drive.google.com/file/d/1DeDi3D3pm8Ikg-H6CTdZdRVxpEB5IzLp
All the weights are averaged per latitude bin among the discovery sites in the time plot, and between the two datasets.
I don’t agree that the reporting window should exclude early arrivals, but here is a plot that attenuates data arriving a year before it was found by half, and two years is nil. Considering that the majority of the discovery times were in the range of 800-900 days after the 7th Arc, this is a reasonable value:
https://drive.google.com/file/d/1Dgfs7mI712AOBAckrWE-9cVN8yL1XvaO/view
https://drive.google.com/file/d/1Dg9s9Fh0y-p_-5uH6tHIFUWdR_S9pNsm/view?usp=sharing
It’s quite clear here that the main features of my previous histograms are unchanged. The method I’m using is as simple as it gets, and obviously robust to variations in weighting.
I suspect the reason that it doesn’t match yours is that you are multiplying all the fractional site sums together, getting a very noisy result where all discovery sites must agree with significant matches, then smoothing by lumping the far fewer hits in wide 1 degree bins. The other major difference is that I’m using a tapered linear weighting on the time/distance proximities, but you’re using a rectangular window. Boxcar filters usually broaden the result, where any strong hit has full weighting for the duration of the window. Triangular/hanning/gaussian/etc windows give finer results.
I wish you’d stop claiming that my results have a bias or a smear. It implies that I’m not following best scientific methods to avoid a very different sort of bias. As you are now focusing on my spelling errors, I don’t expect you’re going to come around to see a simpler approach as valid, or as a validation test of the meta-analysis.
My point about why a lesser result based on different data needn’t be the highest peak appears to have been lost. Suppose that your candidate site doesn’t sum to the highest probability along the 7th Arc (which is what my plots show). Not the highest, but not nil. If it turns out that your site is correct in the end, it would be clear that it really didn’t matter that there were more optimal drift paths at hypothetical crash sites. The only requirement is that it’s feasible – that a reasonable study doesn’t rule it out entirely. If other studies like mine ruled yours out entirely (which I don’t), then they would have been wrong.
@370Location,
You said: “As an extreme test, I’ve shortened the window to just 16 days, which reveals all the arrival waves. It has little effect on the shape of the latitude histogram. It just gets spiky/noisy due to fewer samples.”
Your conclusion is incorrect for crash latitudes north of 23S (and especially at 8.3S). The likelihood values there were dramatically lowered when you applied a penalty for early arrival, compared to your previous plot, which did not. For example, at 8.3S your previous plot (with no penalty) showed relative likelihoods of 0.4 for both your “low windage” and “high windage” debris categories. Your normalized relative likelihood sum was 0.4 and your likelihood product was 0.16. Your most recent version of the same plots (with narrow time windows) shows zero for the “high windage” debris and about 0.02 for the “low windage” debris, with a sum of 0.02 and a product of zero (why didn’t you show the product plot?). How can you say that there is “little effect” when your proposed crash latitude changes likelihood from 0.4 in the sum to 0.02 (20X lower), and from 0.16 in the product to zero (i.e., from feasible to infeasible)? I see a huge effect north of 23S.
You don’t say how late arrivals are penalized in your final set of likelihood plots, but you allow arrivals which are two years early to contribute, and you penalize by only a factor of two for being one year early. This time window is too wide to provide useful latitude selectivity, especially for the flaperon. The resulting loss of time resolution smears your PDF in latitude space by overweighting crash latitudes which are not well aligned in space and time with the debris reports.
You said: “Boxcar filters usually broaden the result . . .” That is not true in general. It depends on the width of the boxcar. When the boxcar is narrower than the average width of a triangular window, it has higher resolution and it narrows (not broadens) the result.
You said: “My point about why a lesser result based on different data needn’t be the highest peak appears to have been lost.”
I thought we were analyzing the same CSIRO predictions and the same MH370 debris reports, so the “data” are not different. Did you mean different analysis methods? Or a different latitude? I certainly agree that the true crash latitude won’t always appear at the peak of a predicted PDF. If we had a large number of data sets, that would trend to be the case if we did the analysis correctly. Since we have just one crash site and one set of MH370 debris reports, it would be coincidental that the peak of the predicted debris drift probability happened to exactly match the true crash latitude. All latitudes which have significant relative probabilities (which are above the noise floor) are “feasible”.
Your analysis method differs from UI (2023) in three regards. First, you equally weight two debris categories (the flaperon and everything else). In UI all debris had equal weight. Second, you apply smaller penalties for significant mismatches in space and time between the MH370 debris reports and the CSIRO-predicted drift tracks. Doing that lowers the latitude selectivity. Third, you (generally) average the PDFs rather than multiplying them (as one should for conditional probabilities). That multiplication process increases the fractional noise, whereas averaging the PDFs reduces the fractional noise. However, when you average conditional PDFs you no longer have a quantity which is a true probability. As a result, comparisons of average values do not reflect their relative probabilities. You can’t even say that latitude X is more likely than latitude Y, much less by how much.
For example, suppose at latitude X we have two PDFs (for two debris classes) with values of 0.2 and 0.8. Their average is 0.5, but their product is 0.16. Next, at latitude Y we have 0.4 and 0.5 for the two debris classes. Their average is 0.45 and their product is 0.20. Which is more likely? Latitude X or latitude Y? Using the average, latitude X wins out (0.50 versus 0.45). Using the product, latitude Y is more likely (0.20 versus 0.16). So, averaging conditional probabilities does not even guarantee that you can tell which of two latitudes is more likely.
In order to assure that you can determine the relative probabilities, you must use the product when the probabilities are conditional (as these are). You want to know the probability of a particular crash latitude producing drift tracks matching MH370 debris report #1 in space and time AND that latitude also matching report #2 (and so on). That “AND” is what drives the product. We want to know the probability of matching report #2 GIVEN the fact that the same latitude also matches report #1. Thus, the probability of a latitude matching multiple debris reports is P(A & B) = P(A) x P(B). If you don’t use the product, you don’t have a relative probability curve. The inevitable price you pay in knowing the relative probability is higher noise. The product has higher fractional noise than all the components. Trust me, I don’t like that “penalty” of higher noise for doing the computation correctly, but we have to live with it. The benefits are much greater latitude selectivity and the ability to accurately determine relative probabilities of different latitudes.
@Hank
Thank you. Like you I believe we witnessed a deliberate action to the end. I make the “worst case assumption” (to me, realistic assumption) the pilot was savvy and the plan was to hide crash.
A savvy pilot had no idea about the Inmarsat rings, but does know the SATCOM is a possible vulnerability: SATCOM shows he is still flying if a call comes in, and at least he must have wondered if the sat calls gave away GPS info. For that reason, pilot turned off SATCOM (we call that Arc7) and continued flying, probably under the thick clouds with fuel. Prior to that he probably already (before Arc6) had descended to in the range FL150 (based on Arc6 to Arc7 distance and speed). Arc7 we see the BFO dip below to about 5000-ft to become invisible visually. This probably happens in the range 30-32s where these is still fuel. Your guess as good as mine as re: end point, but I am thinking deep, hard to search spot, as far as possible from Arc7.
I suggest the savvy pilot assumed we might figure out the southerly path, so his priorities is to save fuel for hidden flight path after Arc7. Given all that, I do not see a need to fly S-curve or stay in darkness (which is 38 South crowd’s key assumption.
I actually feel the savvy pilot scenario is what the data says and it is not so hard to figure out up to Arc7. After Arc7 is the problem: where and how far from Arc7 is possible in worst case scenario?
@Hank
Thank you. Like you I believe we witnessed a deliberate action to the end. I make the “worst case assumption” (to me, realistic assumption) the pilot was savvy and the plan was to hide crash.
A savvy pilot had no idea about the Inmarsat rings, but does know the SATCOM is a possible vulnerability: SATCOM shows he is still flying if a call comes in, and at least he must have wondered if the sat calls gave away GPS info. For that reason, pilot turned off SATCOM (we call that Arc7) and continued flying, probably under the thick clouds with fuel. Prior to that he probably already (before Arc6) had descended to in the range FL150 (based on Arc6 to Arc7 distance and speed). Arc7 we see the BFO dip below to about 5000-ft to become invisible visually. This probably happens in the range 30-32s where these is still fuel. Your guess as good as mine as re: end point, but I am thinking deep, hard to search spot, as far as possible from Arc7.
I suggest the savvy pilot assumed we might figure out the southerly path, so his priorities is to save fuel for hidden flight path after Arc7. Given all that, I do not see a need to fly S-curve or stay in darkness (which is 38 South crowd’s key assumption.
I actually feel the savvy pilot scenario is what the data says and it is not so hard to figure out up to Arc7. After Arc7 is the problem: where and how far from Arc7 is possible in worst case scenario?
@370Location: I don’t see how Cloudflare would work with pages that are dynamically created and served, which is the essence of WordPress. (I suspect the database calls are also what makes the site so slow.)
@DrB and @TBill
Thank your for your comments. I have always respected the work done by this group and did not intend my comments to be so absolute by using “never” in my language.
Clearly the “straight” “unpiloted” course to fuel exhaustion meets all the Inmarsat data has been clearly shown to be feasible. I have never had an issue with using this assumption because it is clearly viable and needed to be searched.
If the aircraft was programmed to fly directly toward the point where arc 7 crossed at 25S, for example, it would have had excess fuel and all of the other arc crossings would not match the data. So a direct route to that location would not be feasible because it would not fit the Inmarsat data.
But there are many ways to maneuver to burn fuel along the way to put the aircraft into the ocean with empty tanks using s-turns to decouple range and endurance. But out of the infinite routes with s-turns, it had to “just by coincidence” match the recorded Inmarsat data. This could not have been planned – it is just an outcome from a wandering flight out of the infinite maneuvering possibilities. This is an unplannable random route that may have actually been flown.
My question was – “could an active pilot have decided to plant the aircraft with zero fuel at a specific point in the SIO and flown a maneuvering path to get there?” And the actual flight that day would have just resulted in the same Inmarsat data as the straight profile.
Is there only one possible flight path that could have been actively flown that night that could have resulted in the exact Inmarsat data set?
You can’t preplan the flight path to fit the unknowable Inmarsat data, but it might be feasible to pick an endpoint and work backwards to create a feasible solution.
This is not a good outcome if this possible to do with reasonable maneuvering. Better to have a dead pilot.
@VictorI:
Not all of the content is dynamic. Images and js can be cached. If a page hasn’t changed, (sometimes affected by mobile or desktop rendering), the whole thing can be treated as static. There are WP cacheing plugins that help. Here’s more info:
https://www.cloudflare.com/learning/cdn/cdn-for-wordpress
@DrB:
The very narrow time windows were an extreme test to show that there’s no need to make individual plot to discern the modeled arrivals at each destination. It wasn’t meant for histogram analysis, except to show that the shape of the my histogram curve was the same even with narrow time restrictions, but the result is down in the noise.
Here’s a time plot and histogram with time windows closer to yours. The details are in the title:
https://drive.google.com/file/d/1E0ovQqg7_3WBO8rik-RjJWyf5XZBo609/
Histogram:
https://drive.google.com/file/d/1E1secWQB3jKeU2K8N5VZD4OFBaHM3D1e/
We don’t know how accurate the drift modeling is. I set 20 days before the proximity is weighted at zero for modeled debris arriving after it was actually found at a site. We also can’t know the reporting delay added to the drift uncertainty. Debris is still being found. It has 50% weight at 100 days early in this plot, but I think that’s still not long enough to be compatible with the even the flaperon, which was reported there “months earlier”.
I’ve dropped the PDF product from my plots because I think it’s the wrong approach for such uncertain data, resulting in multiplication by zero of useful results. It was only added to make that point.
@DrB: “I thought we were analyzing the same CSIRO predictions and the same MH370 debris reports, so the “data” are not different.”
AFAIK we are using the same two CSIRO datasets. Here’s what got lost: “Again, because my candidate is based on new acoustic evidence, it is not dependent on being the highest peak on an optimized search probability curve from inexact data with narrow assumptions.”
I am not a statistician, but my understanding is that joint probability products are not common except for understanding how distributions relate to each other, and that those datasets should be independent, not conditional. Taking a multivariate joint PDF makes that even more difficult, especially with 19 (or 86K) products. If I understand sk999’s method on the SATCOM analysis, he’s taking the eigenvectors or SVD of the covariance matrices to force each pair to be orthogonally uncorrelated before taking the product. I don’t see any such treatment in your debris drift PDF method.
My simple histogram summing method appears to be quite similar to the method used by David Griffin, who provided the datasets. Here’s an image on his site that’s from his paper “The Search for MH370 and Ocean Surface Drift”, p 18, fig 3.2.1:
https://www.marine.csiro.au/~griffin/MH370/br15_pwent2d/pfromto_1_nonflap.gif
It only covers latitudes 26-42S, but you can see by the same early peaks near 31S and the dropoff near the areas already searched that his probability histograms are being summed.
I do use many more histogram bins along the latitude axis to show the many variations due to the butterfly effect.
@370Location,
1. You said: “Taking a multivariate joint PDF makes that even more difficult, especially with 19 (or 86K) products. If I understand sk999’s method on the SATCOM analysis, he’s taking the eigenvectors or SVD of the covariance matrices to force each pair to be orthogonally uncorrelated before taking the product. I don’t see any such treatment in your debris drift PDF method.”
You are conflating two different problems, which require different statistical treatments. It’s nonsensical to say that the calculation of the joint probability of a given crash latitude matching all the MH370 debris reports should use the same equations/methods as a figure of merit used in fitting one set of SIO route parameters to best match the statistical behavior of the BTO and BFO residuals.
2. You also said: “I think that’s still not long enough to be compatible with the even the flaperon, which was reported there “months earlier”.”
I am not aware of any credible reports of an earlier flaperon arrival. The beach cleaning crew would have removed such a debris when it was first found, so they did not observe it at an earlier date.
3. You also said: “I am not a statistician, but my understanding is that joint probability products are not common except for understanding how distributions relate to each other, and that those datasets should be independent, not conditional.”
If you became familiar with the proper methods for the statistical analysis you are attempting, you would appreciate your error. The MH370 problem is like the coin-flip example. Suppose I ask what the probability is that I will flip heads 10 times in a row. That probability is (1/2)^10 = 0.001, not 10 x ½ = 5 or even AVERAGE (1/2, 1/2, 1/2, 1/2, 1/2, 1/2, 1/2, 1/2, 1/2, 1/2 ) = 1/2. When every trial must have the same outcome (either “heads” in the coin flip example, or a drift trial matching a MH370 debris report), you must use the product of the individual probabilities. Thus, you have to find the Nth product of each individual debris-report(or coin-flip) probability. In the MH370 case, I want to know the probability that MH370 debris landed at ALL the debris reporting sites (just like all the coin flips had to be “heads”). Therefore, I also have to find the product of all the individual probabilities for each debris report.
4. You also said: “My simple histogram summing method appears to be quite similar to the method used by David Griffin, who provided the datasets. Here’s an image on his site that’s from his paper “The Search for MH370 and Ocean Surface Drift”, p 18, fig 3.2.1 . . . It only covers latitudes 26-42S, but you can see by the same early peaks near 31S and the dropoff near the areas already searched that his probability histograms are being summed.”
You are misinterpreting those CSIRO plots. The three panels each show a single PDF (not a sum or a product) which is the probability that debris are predicted to land within certain geographical zones. CSIRO is not combining multiple PDFs in these plots. They simply show the number of trials predicted to be in each geographical zone at the selected time bin. If you wanted to know the probability that non-flaperon debris were predicted to land in Africa AND land in Reunion AND land in Western Australia, you would have to find the product of the three plots.
5. You also said: “I’ve dropped the PDF product from my plots because I think it’s the wrong approach for such uncertain data, resulting in multiplication by zero of useful results. It was only added to make that point.
You seem to be saying two things here: (a) you won’t exclude a crash latitude which has zero probability of matching one or more MH370 debris reports, and (b) that only non-zero probabilities are “useful” (implying that having a zero probability result for a given latitude is not useful and should be ignored).
Those statements defy common sense and illuminate your (admitted) lack of understanding of statistics. Uncertainty (noise) has nothing to do with selecting the proper analysis method. In addition, ALL predicted probabilities are useful, regardless of their numerical value. In the MH370 case, near-zero probabilities for any debris report are quite effective in eliminating portions of Arc 7. That is why CSIRO only plotted latitudes from 26S to 42S, because they had already concluded that the probability of a crash north of 26S had such a low probability that this portion of Arc 7 was inconsequential in making a crash latitude prediction. That is, the area of the Joint PDF curve north of 26S is essentially zero. By NOT excluding zero-probability latitudes, you are applying subjective bias in predicting the joint probability, and, as I have previously pointed out, your result is not an accurate measure of relative probability. It won’t reliably tell you correctly whether Latitude A is more likely to be true than Latitude B.
The video of “teleportation” of #MH370 is now 100% debunked. Jonas De Ro, a digital artist, has released a video showing the match to a cloud photo he took in 2012 over Japan. The raw image shows Mt Fuji, geolocating it far from the Andaman Sea. Japan also matches the EXIF data, which theoretically could have been altered.
@Victor Iannello
Thanks for that link, Victor. Apparently for some of those pushing the faked video, the use of stock visual effects for the “portal flash” wasn’t enough evidence. Hopefully this will put the matter to bed. Brandolini’s Law in action.
With a bit of luck someone will put something together addressing similarly fraudulent claims, such as aeronautical composites having the lightning strike protection layer being separated from the surface material by the honeycomb core, or the number of satellite data units installed on 9M-MRO.
@Mick Gilbert: A symptom of MDS is to double-down on claims that are patently false. As the affliction is rarely cured, limited effort is spent exposing the symptoms.
@Mick,
Please can you confirm that in composite construction, the lightning protection layer is always on the surface layer. Is it possible that it is embedded in certain areas of the aircraft ?
Thanks
@Tim: Did you choose to withdraw your paper claiming that the Tataly-Antsiraka debris was from the inboard flap? The chronology of events surrounding that paper is a bit bizarre. Can you provide us with some color?
@Tim
G’day Tim,
I’ve not been able to find any reference material that shows anything other than the most logical arrangement of the layers; the external surface layer, the lightning strike protection layer, an isolation/insulation layer, the core material layer.
There are a good many illustrations of that layering sequence readily available to anyone who looks (eg the Boeing graphic included in this industry post
https://www.comsol.com/blogs/protecting-aircraft-composites-from-lightning-strike-damage/)
Doubtless, the authors have elected not to include, amongst their twenty various diagrams and photographs, a diagram showing the location of the LSP in aviation composites for a reason, just not a particularly “good” reason with respect to honesty or veracity.
Beyond what research might indicate, just think about the purpose of the LSP in composite materials in aeronautical applications. To quote from a NASA paper on the topic,
“Without proper lightning strike protection, the carbon fiber/epoxy composites can be significantly damaged, particularly at the entry and exit points of the strike. Approaches have been developed to protect the composite structures from lightning direct effects to reduce damage to acceptable levels by using conductive foils or meshes in the outer layer of the composite system.” (Electrical Characterizations of Lightning Strike Protection Techniques for Composite Materials, Szatkowski et al, 2006) – my bolding.
The LSP cannot fulfil its purpose of keeping the current/heat generated by a lightning strike away from the core if it is positioned such that the electrical current has to pass from the surface layer through the core to reach the LSP. It is patently ridiculous to suggest otherwise.
A five minute conversation with anyone involved in composite construction techniques will quickly reveal what the mesh layer on the bag-side (furtherest from the exterior surface layer) of the composite is – it is the resin/media flow mesh.
One can only wonder why the two authors are going to such great lengths to fraudulently misrepresent these pieces as having come from 9M-MRO.
@Victor,
Yes, the Tataly piece will not fit between the fastener lines on the flaps. So I’m still looking to find another possible location.
@Mick,
Thanks for that Boeing report. So for the B787 the LPS is definitely in the outer layers. I’m still considering the possibility that in older designs, and especially near the trailing edges the LPS might be embedded. But, so far have found nothing conclusive.
@Tim
Tim, here’s an industry document from The Gill Corporation specifically for the B777 – https://www.dropbox.com/scl/fi/jm8jlt17ye76qy29lfbjy/TheDoorway-Fall2022-1.pdf?rlkey=91kzjdgse7xgyr17inx5ji2g8&dl=0
Note, the diagram on p.9; LSP (in this case, Strikegrid) immediately under the surface material and outboard of the core material.
And when you say, “nothing conclusive”, you actually mean “nothing at all” don’t you?
You only need a 5 second thought experiment to eliminate the possibility of the LSP having the requisite effect if it were embedded in the resin layer, regardless of where the component was located.
Just from the avoidance of doubt, are you the “aviation professional familiar with the B777” referenced on p.17 of the MH370 Boeing 777-200ER Lightning Strike Protection document?
@Mick,
Thanks for the link for ‘Strikegrid’. They say they introduced this on the 777 from 2004 and onwards. Please help me in my quest to find composite designs in earlier 777s. This will prove one way or the other if Tataly and Broken-O are from a 777.
Perhaps, near the trailing edges, where the current is being taken towards the static discharge wicks it is acceptable to embed the mesh.
And Yes, it is me on p.17 who has been trying to find possible locations for these pieces
@Tim wrote, ‘This will prove one way or the other if Tataly and Broken-O are from a 777.‘
Further bogus and false contrivances are not necessary.
It is already evident from the images posted of ‘Tataly-Antsiraka’ and ‘Broken-O’ that the substrate claimed to be a LPS lies at a depth within the panel where lightning strike damage would be maximised, not minimised (regardless of the manufacturer of the substrate).
@Tim
Needless to say, I have put a reasonable effort into looking for industry references to the use of lightning strike protection layers in composite materials, in particular those used in the B777. Everything I have found tells the same story; the LSP is, of necessity, located just below the surface layer of the composite. It is never embedded in the resin layer below the core.
As I’ve noted previously, a five second thought experiment is all that is needed to eliminate that positioning the LSP such that the core material is positioned between it and the surface layer.
I can say with at least 90 percent confidence that Broken O is not only not from a B777, it is not even an aviation component.
I am at a loss to understand why there is this insistence in trying to demonstrate that these items are something that they manifestly are not? The endeavour sits somewhere between obsessively unhinged and consciously fraudulent.
@sk999
Sorry for the delay in replying; I have been traveling and fully occupied for a while. The file you shared on Nov 28th is really helpful. I guess one of the key questions is “ If the errors are statistically independent amongst all the observables”, in particular with respect to BFO data. I currently assume that the BFO frequency bias may especially change after a cold restart of the OCXO and that drift is limited otherwise. The formal approach through Bayes Theorem is really interesting.
@DrB:
In another attempt to see why my meta-analysis of the CSIRO drift data shows non-nil values near Java and does not show a prominent peak at 34S like yours, I tried utilizing the product of 19 debris find site results for each day and latitude. I got all zeros.
Your Appendix A.5 shows your method of overcoming this is to have a threshold of one (or start counting at one?) in the accumulation bins where the average number of drift hits is around 6. Elsewhere you say that bins with less than two hits are dropped.
I’m using all the data, weighted by distance and time with ranges similar to yours. Weights do get masked to zero outside of the linear triangular windows. It’s not surprising that one of those 19 categories will contain a zero on any given day, and multiplying them all together gives zero.
There is another way to take the product of a large series. They can be summed in log space.
log(x*y*z) = log(x)+log(y)+log(z)
A small epsilon must get added because log(0) = -inf.
To account for the different number of tracks at each latitude bin, my histograms are already using the arithmetic mean rather than just summing. I can now take the geometric mean instead which contains the product of the variables:
geomean = exp(sum(log(dataset+epsilon))/n)-epsilon
(I use the scipy.stats.mstats.gmean function.)
Here is a plot using this method to get the joint probability of the 19 sites, and also the two windage sets:
https://drive.google.com/file/d/1EGU26hvTEdQ_roN9FyRM69SJX-c9MuDX
As you can see, the shape of the resulting curves are again unchanged. There is no major peak at 34S, and values near Java are not nil. If I expand the reporting window to allow much earlier arrivals, the values N of 23S are more intense.
To test if the method is robust, I tried also using the geometric mean when summing tracks by latitude. The weakest latitudes have more zero values with taller spikes, but the shape is again the same.
DrB: “I am not aware of any credible reports of an earlier flaperon arrival. The beach cleaning crew would have removed such a debris when it was first found, so they did not observe it at an earlier date.”
I already addressed this point. It’s a very narrow assumption that the debris being stuck in a region would be washing up again on the same resort beach. Victor and I were tagged Nov 30 on a news report of a prior flaperon sighting at Reunion May 10, 2014:
https://x.com/elizanow1/status/1730213082469908990
You may be disputing the credibility of a local news report, but your narrow assumptions may not apply. It is quite possible that the flaperon was in the area for months or on remote beaches before it hit a popular beach and was recognized as part of MH370.
DrB points 4&5:
You say that the CSIRO plots were not a joint PDF, presumably because only the flaperon had been found at the time. Then you say they ignored half the 7th arc because, “… That is, the area of the Joint PDF curve north of 26S is essentially zero.”
That seems like either a contradiction or more assumptions based on your own results, which I have questions about. The CSIRO was initially tasked with determining whether the active search area was compatible with drift. That was their focus.
I think I’ve shown that my method of weighting the histograms by time and distance is robust to even large changes in the parameters, and it has enough SNR to show the model uncertainty within each latitude division by using finer resolution.
I’ve accommodated all of your constraints but there is still a big mismatch with your result. I could share my python code for others to run the experiment if you still dispute my result. Is your spreadsheet available?