On December 30, 2025, Ocean Infinity (OI) begin its third campaign to find MH370 on the seabed in the Southern Indian Ocean (SIO). The search vessel is Armada 8605, which is equipped with a team of three Hugin Autonomous Underwater Vehicles (AUVs). This search continues in the general area that was started in March 2025 using the vessel Armada 7806.
The map above shows how this campaign relates to previous efforts, including:
Original area searched by the ATSB. This part of the 7th arc was searched by GO Phoenix ship using a towed vehicle (“towfish”) with synthetic aperture sonar (SAS) starting in October 2014 and completing its mission in June 2015.
Area searched by OI in 2018. In early 2018, OI expanded the area searched previously by the ATSB. The area was extended outwards to around 59 km (32 NM) from the 7th arc and inwards to around 46 km (25 NM) from the 7th arc, and along the 7th arc between 36S latitude to 24.7S latitude. After searching around 120,000 square kilometers, which matched the total area covered previously by the ATSB, the MH370 debris field was not found.
UGIB Last Estimated Position. In March 2020, an extensive data-driven study was released by Ulich, Godfrey, Iannello, and Banks (UGIB 2020) which reconstructed the path of MH370 based on data sets and information from satellite, radar, aircraft performance, navigation systems, drift analysis, aerial search results, and archived weather. Assuming no pilot inputs after 19:41z, the study predicted that MH370 crossed the 7th arc around 34.23S, 93.78E, designated the Last Estimated Position (LEP). In June 2023, Ulich and Iannello subsequently incorporated results from an updated drift model that was used to refine the search area recommendation, including the possibility of a 70-NM glide after fuel exhaustion.
High Priority Search Area. In February 2023, Iannello recommended a search area that should be prioritized. Located on a steep slope about 33 km due south of the LEP and 27 km from the 7th arc, the area was not scanned by GO Phoenix’s towfish and appeared to have been only partially scanned by Ocean Infinity’s AUVs during the first campaign in 2018.
OI Proposed Search Area. At the 10th anniversary of the disappearance in March 2024, OI announced its intention to conduct another search for MH370. The area proposed by OI at that time centered approximately around the specific location (34.23S, 93.78E) from UGIB 2020, designated on OI’s map as the “IG Hot Spot”.
Track of Armada 7806. This track can be used to estimate the areas searched during the OI campaign that began in March 2025. Although the exact area scanned by the AUVs is not known, Armada 7806’s path to launch and recover the AUVs gives us a good idea of the area covered. The search area includes the UGIB LEP, the High Priority Search Zone, areas where data quality from previous searches was poor due to rough terrain or equipment malfunctions, and other unsearched areas within the OI Proposed Search Area.
Likely Progression Armada 8605. The current search seems to be methodically progressing to complete the areas within the OI Proposed Search Area that were not searched in March 2025. The search began on December 30, 2025, with unsearched parts of the proposed area (red rectangle) that is outwards (southeast) of the 7th arc. This area would include an unsearched area due south of the UGIB LEP and the High Priority Search area. These locations (the LEP and High Priority Search Areas) were based on a reconstructed flight path that was determined to statistically favor due south, so unsearched areas further along the due south trajectory is of high interest and will be reached soon by Armada 8605 and its AUVs.
Acknowledgement. This article has benefited from the many private discussions with Mike Exner, Don Thompson, Bobby Ulich, and Andrew Banks.
@Victor
I presume you mean 128000 km**2 for the 2018 OI search.
@Viking. Yes. Thank you.
@Ed
Concerning your 16s window for the FFT I initially thought that was a good value, since it includes the entire (expected) dispersion broadening in ‘dirty’ cases. However, after thinking deeper about it, I am increasingly convinced that it is not a good value to use.
The reason is that if you make a FFT for a signal with the width dominated by dispersion broadening you affectively distribute its contribution to the heading over the entire time window. If the (dispersion free) width is just a few seconds that leads to a relative weight reduction (compared to a signal with no dispersion arriving at the same time) is reduced by the ratio of the original duration and 16s. That is effectively around 4-6 times reduction in contribution (naively).
With imperfect treatment of related boundary problems (filters, etc.) this factor may get worse.
I therefore suggest you try a shorter time window.
Alternatively you might try adding a simple linear dispersion correction corresponding to 10-15s broadening. Please try both positive and negative sign for the dispersion for each peak.
For a given (dirty) peak only one sign will work, but which one it is will vary among the peaks.
I am not an expert or anything like that, but am very curious about what happened. I am wondering if someone knows how often Ocean Infinity is analyzing data from the submersibles? Is it everyday? Is it possible they could find something and alert people in real time from their ships, or do they have to collect the data, and bring it ashore for analysis? I am also noticing they are spending more time in some places but not others within specific search zones. Is this indicative of seeing something that could potentially be debris or something like that? Please forgive my lack of knowledge and expertise.
@Victor Iannello et al
G’day Victor, by way of a bit of a straw-poll/sanity-check, might I trouble you (and others) for what likelihoods you assign to each of the following:
1. The seventh arc is accurately (±10km) mapped.
2. At the seventh arc, MH370 was in a state of unpowered flight subsequent to fuel exhaustion.
3. MH370’s flight path from the seventh arc followed one of the spiral descent profiles provided by Boeing to the ATSB (±10km to end point).
And yes, I understand that there were 10 descent profiles provided by Boeing, but they all placed the end point within 60 km of the arc.
Thank you in advance.
@Greg Hood,
Greg, Neither am I, but the link below may assist.
The mission is largely robotic. The ship is manned, but the searching is being done by three (I think) autonomous submersibles. The ship (and crew) is not doing the first of the searching “itself” but supporting the underwater searchers. They can report some data from below and the ship/crew can communicate from above.
The ship/crew also are in remote contact with land based facilities and personnel.
“Is it possible they could find something and alert people in real time from their ships”: Yes.
“I am also noticing they are spending more time in some places but not others”. The surface ship is in support of the underwater robots doing the searching. In particular (I think) it is necessary to be virtually at a standstill to retrieve any one of the underwater vehicles, if “virtually at a standstill” makes any sense in terms of the likely prevalent sea states (ha,ha).
The underwater robots have an endurance of days, if that is advantageous to the task at hand. So, if a submersible is to be retrieved, serviced and data transferred before being re-inserted into the water to continue the search, then the ship may appear to be at that “virtual standstill”, again if that is advantageous to the task at hand.
https://oceaninfinity.com/technology/#:~:text=%E2%9A%A0%EF%B8%8F-,ADVANCED%20ROBOTICS,-Remote%20command%20and
@Mick
1. 100%
2. 100%
3. Unknown as stated. But 90% within +/-20nm.
@airlandseaman
Thanks Mike.
Is anyone else tracking the OI ship?
They are certainly very focused on a specific location at the moment.
https://www.mh370-caption.net/index.php/armada-tracking/
@Mick.
1. 100%
2. 100%
3. Uncertain whether final descent matched Boeing simulations. BFO tells us beyond reasonable doubt that it was in steep and accelerating descent at 00:19:30 and would have exceeded MMO and/orimpacted the water within about 30 seconds. The notion that impact point is more than 10NM from the arc is fanciful. More likely it was <5NM.
@Paul Smithson
Thanks Paul.
I could add that by UGIB’s fuel analysis there is no way that the packs were running after 20:00 – so no live pilot to glide the aircraft. I find it, frankly, baffling that an area only reachable by glide is now being prioritised and would be glad if anyone can explain the logic.
Enjoying following these updates, thanks.
Has there been a review of the Turner 2016 scenario? this was ‘Failed Diversion to Banda Aceh airport with Hypoxic crew’, with a crash site just west of the IG 2020 location. I ask as a video of his presentation recently appeared online last month: https://www.youtube.com/watch?v=RxkL7GG3k-A
I think the presentation was from March 2025, but he did another in Oct ’25 in Perth.
@Theo: I haven’t reviewed it as some of the fundamental hunches are unlikely.
There is an “internet personality” that has appeared on various platforms that keeps pitching the hypothesis that a ruptured oxygen bottle was responsible for a long-list of system failures that led to the disappearance. That character goes by the names Pete Noetic, Paul Onions (on this site), Captain Ripari (Reddit), and most recently Captain Peter Turner. Although the point of impact has moved over the years, the basic premise of the scenario remains the same. That person was banned from this site after being repeatedly warned that some of his facts were erroneous.
Perhaps in the latest incarnation of the scenario, some of those errors were corrected.
@jip17
Yes I noticed this as well. I don’t think we should take it as much until a few days from now, if they continue darting over and staying at that spot instead of continuing the search north similar to the pattern observable from their last search, it would be fair to speculate at that point that there may be something going on. Here is hoping, but i’m not holding my breath, too early to tell anything at this point. Wishing the OI team luck.
@Paul Smithson @airlandseaman
We’ve already arrived at a point in time where those who believe that the plane somehow crashed in a high speed, no fuel dive are complete lunatics. Where was the debris field? How is the flap, flapperon in perfect condition? Even if fluttered off, there would still be water impact deformaties which are not visible. The leading edges are perfect, but the trailing edges are eroded pretty much exactly as they found the control surfaces on flight 1549. There is too much going for a ditching to defend without any evidence or reason that it fell out of the sky. There is a reason why searches have moved from directly underneath the 7th arc, which has mostly been covered, to outside it where a glide would have taken it. If it had been in a high speed crash, the ATSB would have found it, because thats where they were looking for it. Ocean infinity isn’t wasting its money searching in a place they don’t think the plane could be, they want that 100M.
Happy new year 2026 to all as well
@Darius Tanz, where’s the evidence, you ask?
1. Witness marks confirmed retracted position of Flaperon (Reunion) and Flap (Pemba) – ref expert analysis by qualified investigators.
2. Fuel exhaustion occurred (SDU power-cycled between 00:11 and 00:19) around time of expected fuel exhaustion – yet no descent ongoing at 00:11 (ref BFO)
3. Autonomous straight glide impossible with AP disengaged (ref autopilot logic).
4. Pilot unlikely to have been alive (ref fuel endurance to 00:11+ impossible without bleed air/packs off for several hours).
5. If pilot HAD been alive and intended to ditch an idle descent would have occurred followed by ditching before fuel exhaustion – neither of which happened. (ref BFO and BTO).
5. Final BFO data prove a steep and accelerating rate of descent at 001930 (ref Holland). Extrapolate by 30-40s and you either approach M1 or hit the water.
6. Recovered debris consistent with high energy impact (ref recovered debris fragments)
7. Trailing edge damage consistent with flutter (ref expert opinion – not mine)
I don’t see that there’s any reasonable alternative inference than fuel exhaustion, uncontrolled descent and hard impact close to the arc.
@Darius Tanz: In my opinion, the MH370 debris most closely resembles the parts recovered from Silk Air 185. Fragmented parts were recovered from a debris field about 60m x 80m, and flight control surfaces that separated during the high-speed descent due to flutter were recovered some miles from the debris field.
@Mick
From my work, as you may know-
(1) 99% BTO/BFO accuracy: Looks A-OK to me. We benefit from Inmarsat with IG assist helping to refine the data to make it as accurate as possible
(2) Zero. Active pilot intentionally flew some distance away from what we call Arc7…after about Arc5, MH370 merged onto the home sim path to Arc7, and kept going with a little fuel. SATCOM off’ed by active pilot, probably twice.
(3) Descent did not follow the Boeing simulations, because this was not a ghost flight falling from high altitude. Instead it was a manned flight with descent, slow-down well before Arc7. I am somewhat open-minded on final crash strategy, given no ELT signal, but I do feel it is possible that the pilot had a hold pattern descent at the end, which I infer from the home sim data. In general crash looks like it could be like Ethiopian 961 with a flaps-up, high-speed, wings-semi-level catastrophic “ditch”, sort of similar to CAPTION’s vision there. In that sense final entry to water could look a little like a Boeing sim case.
@TBill
Thanks Bill. I would have lost money there; I had you down as assigning a small but non-zero value to 2.
@Victor
Thank you for the updated diagram. Would it be possible to include the lat/long coordinates for the search boxes and also a north arrow for reference? Or if this already exists, a link would be helpful.
@Mick Gilbert: I guess I’m less certain than other people about what I consider to be most likely scenario.
>1. The seventh arc is accurately (±10km) mapped.
95%. We don’t know what we don’t know, although this seems the most solid of the evidence we have.
>2. At the seventh arc, MH370 was in a state of unpowered flight subsequent to fuel exhaustion.
70%. The last SATCOM data set is consistent with an APU power up after fuel exhaustion, but again, there is a finite possibility that there was manual intervention. This would also mean that our fuel model, which predicts the air packs were off and nobody was alive, is in error. That error is possible, although not the most likely.
>3. MH370’s flight path from the seventh arc followed one of the spiral descent profiles provided by Boeing to the ATSB (±10km to end point).
10%. None of Boeing’s descent profiles predicted the 0.7g downward
acceleration at the time of the final BFOs, presumed to be around 2 minutes after fuel exhaustion. However, this does not necessarily imply there were pilot inputs, as the Boeing model did not include provisions for engine restarts and perhaps other phenomena, which could result in a steep banked descent and an impact close to the 7th arc.
The simplest explanation of the data is that MH370 crashed close to the 7th arc after fuel exhaustion. However, we can’t discount the observation that the debris has not yet been found close to the 7th arc, which lowers the probability that the debris field is close to the 7th arc.
@Stuart: North is up in the figure at the start of the article.
The first image of this past article shows a similar map from the search in March 2025, except the Google Earth grid is turned on.
https://mh370.radiantphysics.com/2025/03/31/update-on-the-search-for-mh370/
@Victor Iannello
G’day Victor,
Thank you for your reasoned and considered response.
@Victor. You said “The simplest explanation of the data is that MH370 crashed close to the 7th arc after fuel exhaustion. However, we can’t discount the observation that the debris has not yet been found close to the 7th arc, which lowers the probability that the debris field is close to the 7th arc.”
An alternative phrasing of that ends:
“…which lowers the probability that the correct section of 7th arc has been searched.”
It seems to me that the most unequivocal data currently available are:
1. The BTO data (together with the various pieces of evidence that strongly support a terminus close to the arc).
2. The underwater search null result. This is by far the most robust “latitude discriminator” – ahead of BFO, drift, fuel or any other line of evidence that has so far been marshalled.
The combination of 1 * 2 tells you that the most probable residual is the 7th arc residual that has not been searched, ie 39.5-40.0S.
In DSTG’s pdf, this was classified as marginal – right out on the 5% tail. My subsequent analysis, which is fully consistent with DSTG’s “sans-BFO” pdf, is that the PEAK probability zone by BTO is actually 39.0-39.5. The published pdf peak at 38.0 is the direct result of attempting to optimise concurrently for BFO (optimum circa 37S) and BTO (optimum circa 39+). If we consider BTO to be more precise/beyond doubt than BFO, then we ought to be looking more closely at the BTO optimum. And that makes 39-40 OPTIMAL, not marginal.
As you and others have often pointed out, this terminus is less consistent with conventional expectations derived from drift, BFO and fuel predictions. I believe that I have now provided reasonable evidence on how each of the “mainstream” interpretations of these may be in error. And it is not as if they were “beyond doubt” in the first place.
Look at this another way. Using (previous) fuel model, drift model and BFO as latitude discriminators predicted segments of the arc which have been searched without success. Something is wrong.
So what it really comes down to is whether you are more inclined to believe:
a) correct latitude has been identified. The search has missed it / the flight terminated >40NM from the arc
b) wrong latitude was identified and their are defensible reasons why these predictions were in error.
I’d be more inclined to go with (b). The fact that BTO optimum lies in this zone – and that this coincides with the most parsimonious path models – makes it all the more compelling.
Happy New Year everyone
@Paul Smithson
Paul, I share your opinion on search priorities, just waiting for the moment if and when the Armada 86 05 is going to make a U-turn and continue searching along the 7th Arc. Moreover, I do not think that even the people who believe in the controlled glide are satisfied, current area is probably too close to the arc for that scenario.
@Victor Iannello @DrB
Victor, on the map above you marked a substantial number of former and potential new search areas. Among them is also the “Ulich-Iannello Search Area Based on 70-NM Glide,” which originates from yours and Dr. Ulich’s paper “Improved Prediction of MH370 Crash Location Based on Drift Modeling of Floating Debris,” where it is marked as “Zone 2.” Here is the map from page 1 of that paper:
https://drive.google.com/file/d/1WkYrS9SHxCNdJkgYCEvjuZGW31h3G9i3/view?usp=sharing
Until recently I was completely ignorant that you actually recommended first searching data gaps and LPDs inside the longer area along the arc (from 36.4° to 32.9°), that is, “Zone 1” on the map. I was aware of “Zone 1A” only. At least a part of “Zone 1A” was searched first in March 2025, but what happened with “Zone 1B”? It seems like it was sidelined at some point. I know that you cannot talk on behalf of the OI, but I think that this area also needs to be marked as well in your posts. People should be aware that the job close to the arc has not yet been finished.
@Jean-Luc Marchand – CAPTION
Jean-Luc I will appeal to you as well since you are running the website dedicated to finding Malaysia Airlines 370. A location closer to the 7th Arc does not contradict your work and proposed flight path. An oversight was made about the fact that the area closer to the Arc has not been searched thoroughly enough and consequently that area was disregarded. The only thing that is needed is to include it along your proposed flight path and see what the viable options are.
@Paul Smithson
(a) Well said on the head and shoulders primacy of the BTO data, which I agree with, 100%.
(b) Well said on the inferences to be drawn from the ‘null’ search (to date) re prior assumptions.
@Mick Gilbert,
Your three questions.
I am going to put on my rose-coloured glasses. And I do not profess to know details.
1: 100%. The 7th arc location is dependent upon the aircraft altitude at the time. It is also dependent upon the accuracies/inaccuracies from the BTO dependent calculations. The latter have been addressed by many analysts, but specifically and clearly by Inmarsat. For any given aircraft altitude your 10 km uncertainty should be more than adequate.
2: 95%. The 5% remaining uncertainty is simply due to “but we can’t be sure”.
3: 5%. There were only 10 descent profiles provided by Boeing being only those 10 executed or trialled by Boeing “itself”. This was adequate to provide an indication of the possible range of such profiles. (And, hence, some early indication as to how much to restrict the search to some reasonability.) But, even so, no-one had previously seriously thought that such a simulation would be truly necessary, and hence no real provision for adequate fidelity. There are also uncertainties as to the aircraft yaw control configuration prior to final fuel exhaustion. None of the previous in this paragraph is to say that the Boeing attempts were not useful, simply that they were indicative of some possibilities, certainly not all.
@George G
G’day George, thank you for your considered answers, particularly regarding the end-of-flight modelling.
@Theo
@Victor
RE: Has there been a review of the Turner 2016 scenario?
As Victor mentioned, “Peter Turner” appears to have been active here about eight years ago, under another name. Some of his claims were shown to be wrong at the time, yet he persists in making them. His Banda Aceh scenario was found to be unlikely, for several reasons, including (but not limited to) the following:
1. The scenario suggests an oxygen bottle ruptured, causing severe damage to the left side of the main equipment centre (MEC), the loss of multiple systems and a slow depressurisation. Such a rupture would certainly cause severe damage inside the MEC, but it’s also very likely that bottle components would be projected through the fuselage skin, as happened in the QF30 incident described in Turner’s presentation. That would cause a rapid depressurisation, not the slow version upon which his scenario depends.
2. Turner claims the crew proceeded to Banda Aceh because the aircraft was too heavy to land at Penang. He seems not to be aware of (or chooses to ignore) the overweight landing guidance in the Flight Crew Training Manual (FCTM), and the overweight landing non-normal checklists available in both the Electronic Checklist (ECL) and the Quick Reference Handbook (QRH). Being overweight does not preclude landing in an emergency situation such as the one Turner describes. The limiting factor is the runway length, which would not have been an issue for a landing on Penang’s 11,000 ft long runway, over 1,000 ft longer than the runway at Banda Aceh.
3. Banda Aceh did not have ATC services available at the time of night in question and only limited rescue/fire fighting (RFF) services. No aircraft commander worth their salt would have bypassed Penang for Banda Aceh at that time of night, given Penang was a very familiar airfield with good weather, a long runway and full ATC/RFF services available.
I didn’t watch the entire presentation, but I noted several technical errors relating to the aircraft’s SATCOM antenna, the CABIN ALTITUDE warning and the FMC behaviour after a failure. Those errors were explained to Turner many years ago by people on this forum
@Victor,
Thanks for your clarified search area map. It would be very nice to have the zoned portions as a .kmz file to complement the CAPTION .kmz updates.
@Jean-Luc,
If the .kmz.gz file could instead be a .kmz file, it might be possible to turn it into a live network link within Google Earth Pro. The .kmz is internally a zip compressed .kml file, so no savings with the gzip compression.
@Viking,
The 16 second (4K sample) window was specific to that graph. It is already masked typically with a narrowing gaussian to remove edge effects. Narrowing it further improves temporal resolution but increases noise. Widening it smears the signal but helps detect weaker events. The map made for the Gulden Draak detection used multiple windows stacked to get the best of both spatial resolution and lowest noise. I use dozens of different algorithms, both temporal and spectral. The LANL comparison plot used an FFT as it was the best I had for reducing beamforming phantoms back then. I do use well padded FFT for spectrograms, and broader windows give better spectral resolution. As I addressed before, dispersion doesn’t alter the bearing result or sensitivity, and is typically seen only at the lowest frequencies. If you can show better s/n by correcting for it without introducing other artifacts, I will be impressed.
@Mick,
1) 95% that the 7th Arc is within 10 km, and probably better. 95% is 2 sigma, and that’s how far I trust calculations I didn’t do myself. When I first ran the epicenter using picked P and S wave arrival timing on the 8 best seismometer recordings, the location was within 0.25 km of the low altitude 7th Arc. That’s quite a coincidence, given that the accuracy of both measures is lower. CSIRO pointed that out, so I reduced the reported lat/lon precision.
2) 95% that the plane had no thrust at the 7th Arc. The login sequence was incomplete compared to the first Arc repowering. If the Java site is correct, the plane must have been at very low altitude at the 7th Arc.
3) 15% that the plane flew a curved path at the end only because to land right on the 7th Arc, it may have turned back to land parallel with the waves, not because it was unpiloted.
@Darius,
The damage on 9M-MRO seems similar to that on AF447, which entered the water in a nose-up, flaps-up stall, falling at about the same rate as forward motion. From the debris field AF447 was clearly shattered, so even a 200 knot entry will not leave a B777 intact. I can’t see how control surfaces like the flaperon and flaps could all lose their trailing edges from flutter and then spontaneously detach from the plane before impact, to land with no leading edge damage. The only vids I’ve ever seen of main wing flutter look like the flapping motion of a bird.
I seem to recall that the RAT powers some hydraulics but not the flaps, which may explain why they were still retracted.
@Paul,
I agree that proximity to the 7th Arc is better defined than broad latitude, so that’s where the search should concentrate. There are huge tracts of the 7th Arc that are completely unsearched because of the possibly false assumption that the plane flew unpiloted. Yet now we are expanding past search areas in case of a piloted glide.
@All,
Current media reports are that the latest search is based on improved technology, multiple more capable AUVs, refined drift estimates, and even barnacle analysis. I recall that OI used eight AUVS in 2018, and they appear to be a similar model to the Hugin 6000, perhaps now with onboard MBE?
What surprises me is that none of the candidates being searched is based on new evidence (WSPR isn’t evidence). Nobody is willing to discuss the photographic evidence that the flaperon barnacles began growth after beaching, probably in early May 2015. That seriously changes the assumptions made in the drift studies, which remain unrefined. The indisputable detection of an event on the 7th Arc while MH370 was sinking doesn’t even get a mention among expert commentators.
I’m not a believer in coincidences or conspiracy theories.
That brings me to why the broader “WSPR” area might warrant a search, if OI only intends to widen the previous search area. The acoustic detection I proposed in 2017 was already searched in 2018. It was only seen on just one hydrophone, and it had some artifacts that may have been reflections. Noise from a surface impact at that spot could get into the SOFAR channel by reflection from the Gulden Draak sea mount. The timing was a good match for the 7th Arc, which would make a later surface crash or distant from the sea mount unlikely. Looking at the 2018 OI track, there appears to be a nearly 100 km gap east and NE of my 2017 candidate. If that area was skipped, then it might make sense to fill that gap. RG is already claiming thunderous glory for WSPR if MH370 is found in that zone he picked, but I still don’t believe in coincidences.
Correction: I have also seen vids of main wing flutter in an oscillatory twisting motion, somewhat like the Galloping Gertie Tacoma Narrows bridge.
Re: Turner’s Presentation
I did sit through Turner’s 1.75-hr YouTube. I have known about his Banda Aceh scenario for some years, as he spars with me occasionally on Twitter(X) as @PeteNoetic.
Basically give him credit (he has been online as Capt RipAri since the early days on Reddit) for developing the most well-thought out “remarkable” accident scenario. The big weakness I see: I have always wanted to see his flight path details, like we all do for each Arc: lat/long/speed/heading/BTO/BFO etc. he typically only shows the flight path. I suspect his flight path would violate the BTO/BFO accuracy we like to see, and I think we might be able to see that a little bit from his presentation.
Turner does correctly grasp that a “curved” magnetic path to the southeast needs to slow down markedly before Arc6/7. My explanation, of course, is the active pilot slowed down. Turner’s explanation is that fuel ran out well before Arc7. He suggests everyone except he is wrong about the fuel shortage to get to Arc7, but he is assuming quite a large fuel flow outside temp debit for a ghost flight at FL340. I recall DrB used to say that a lot, but re: MH370 it’s amazing and unexpected to me that the fuel flow ambient temp debit goes away above about FL360.
Turner starts out, from the first slide almost, condemning deliberate flight, and digs into IG several times for their unfair “mass murder” scenario, so we quickly see where he is coming from.
There are many things he says that I/we don’t agree with, too numerous to count. But at IGARI he shows a standard, smooth 25 deg AutoPilot bank fitting the apparent extreme tight turn data just exactly perfectly. He disputes that the Xponder was switched off at IGARI saying he has proved, I think he suggests in an actual aircraft flight, that the ADS-B data would not report zero altitude in the wrong switch position.
Re: O2 Tank, I think we need to realize the Qantas historic accident involved an older heritage metal gas cylinder, that If I recall was, upon further investigation, produced during a known batch of bad quality metal cylinders. MH370 reportedly had modern composite (eg Kevlar style) O2 cylinders, for which I am not aware of any aircraft failures, and also the composites tend to not rupture but develop small leaks in the side (and I think we know MH370 O2 system was holding pressure pretty well). I believe Turner bolsters his argument with a story about how poorly the MH370 cylinders were maintained, so he asserts the piping and tanks were likely in terrible disrepair, in his opinion.
As an interesting aside on O2 tanks, there has been a kindly American NoK, from a 1960 brief case bombing, assisting the MH370 NoK with their grief process. When I looked up that 1960 flight on Wikipedia, O2 tank rupture was listed as a possible cause, but later proven to likely be a bomb brought on board. So O2 tank rupture is one of those things like a missile that people tend to gravitate to.
In general, what I said on one FB group, how about this? bring this scenario to the attention of the Flight Safety Detectives (former NTSB Greg Feith, et al) and ask them if this scenario remotely seems reasonable. I suspect we know the answer. But Turner’s POI curves over right on top of IG@34s, so we shall see, in fact, if I heard him correctly, I think he asserts he was first person so say the search area should be 33-36s +-45nm and he says he said that back in 2016, so ALSM, you were copy cat on that one.
@TBill said: But Turner’s POI curves over right on top of IG@34s, so we shall see, in fact, if I heard him correctly, I think he asserts he was first person so say the search area should be 33-36s +-45nm and he says he said that back in 2016, so ALSM, you were copy cat on that one.
It would be hard to reconcile that comment with his previous comment on this blog in October 2017 in which he (@Paul Onions) advocated for a search 40-100 NM beyond the “Bayesian Hot Spot” at S38E88:
It appears MH370 was looking for an airport to land.
ATSB final report on page 10 states, “The aircraft passed over or near IFR waypoints ABTOK, KADAX and GOLUD (which are within 3 NM of each other) and later PUKAR. The aircraft made a slow right turn south of Penang Island.”
This flight path goes very close to Mimos, which is the start of the instrument approach point (ILS z Runway 10) to Kota Bharu. Airport is closed and there is no pilot activated lighting option.
http://aip.dca.gov.my/aip%20pdf%20new/AIP%20SUPP/AIPSUPP%20201112.pdf
The flight path then follows closely to the Bidmo1A arrival for runway 04 at Penang.
http://aip.dca.gov.my/aip%20pdf/AD/AD2/WMKP/WMKP-Charts%20Related%20To%20Penang%20International%20Airport.pdf
So with this in mind and due to some unknown reason, or suffering from mild hypoxia, the crew programmed a diversion to another airport, such as Banda Aceh via VAMPI-MEKAR-NILAM-SANOB. End location with APU on is 40-100 nautical miles south of Bayesian hotspot (S38E88). Never searched.
https://mh370.radiantphysics.com/2017/10/02/atsb-releases-final-report-on-mh370/#comment-7455
There were so many technical errors in his comments that he refused to correct that I banned him. And magically he has moved his search recommendation from beyond the DSTG hotspot to an area close to the 7th arc crossing in UGIB 2020.
That’s why I am not bothering to critique his newest videos.
@370Location,
Very good, thanks Ed.
@Victor
Agreed my recollection was that Turner originally went with the straight (True Track) flight path as an accidental ghost overflight and later years switched over to magnetic path, presumably to conform to the belief that a accidental ghost flight could not be True Track (straight). Amazing how many years it took for that understanding to sink in for everyone.
I must say I was surprised to recently see who “Capt Turner” was. I thought @PeteNoetic must be a Malaysian pilot coming to the defense of ZS.
@TBill said: A skeptic might say that Captain Turner moves his impact point based on the work of others.
In any event, I put his work in the same bin as the CAPTION (CAPTIO, or whatever that group calls themselves) path reconstruction which also has an impact point that moved towards the proposal of others. And those guys then have the gall to claim their work was copied.
I have little patience for this kind of attention grab. But attach a story to a path and the media eats it up.
everyone
My apologies for not checking on the comment in the previous article. I need to clarify a few things that were brought up on my analysis and Mr. Exner’s concerns about my lack of response to him (this is not a desire for personal buzz, just trying to help with a modern hand).
@airlandseaman My sincerest apologies for not responding to you. I thought the FB group was compromised by impersonator accounts due to one member becoming very aggressive to send cash through Zelle for another member and thought I was trying to be phished. Once I thought that, I thought no one I was speaking to was really who they said they were and cut off all communication. Again, my apologies and I hope you understand that, Mr. Exner.
Also, my point of 31.4 S, 90.4 E was just the end result of what I could do on my own with all publicly available data and what could be recreated through public domain. UWA’s drift data was not given to me, but enough has been discussed that it can be replicated in scale for modeling.
I really just wanted experts in each field of this case to help me build a better algorithm that could serve as point of agreement between all of your fields. I know most of you have access to better data and modeling that I can only attempt as an independent researcher. My confidence is in the pipeline that I believe that could better locate the aircraft (not the POI; it’s just where the final result pointed of 31.4 S, 90.4 E). I want help and cooperation so we can model the flight, satellite, and drift data that represents the best-case scenario. Yes, it’s complex. It hinges on assumptions with certain things that we can barely calculate otherwise.
To Mr. Iannello,
Thank you for allowing discussion on my study as well. I promise you I have no desire for publicity. My goal in this is trying to stage a full model from the turn at Malacca to match the Inmarsat data and replicate crash scenarios that can create a debris field and dispersal that can match the timing of debris found before 2018 (I figured anything else would be too difficult to match to a true time signature).
@370Location,
The flutter possibility that potentially caused the flaperon to detach is not the torsional [aeroelastic] wing flutter that some videos illustrate. Rather it is the rapid and violent excursions of a control surface, more like a flag fluttering in the wind. It can initiate very suddenly and the excursions can become extreme within one or two seconds.
Boeing mentions the possibility of flaperon flutter in one of its operational manuals, while carrying out static engine tests.
See also https://www.govinfo.gov/app/details/FR-2005-12-16/05-24050
Two items: Arc accuracy and debris field:
How accurate are the arcs likely to be? I’m thinking of a +/-km distance normal to the arc. If the expected error is only +/-10km or +/-20km then it’s hard to understand how the aircraft wreckage hasn’t been found yet. Given the BFO data indicating the aircraft was likely in a steep dive.
With respect to the debris field: Someone asked, on another post, what it might look like. the size of the debris is likely to be similar to Silk Air 185 or the more recent China Eastern Airlines Flight 5735. Both of these aircraft hit the ground at ~Mach 1.0 after losing control surfaces. Perhaps the engine cores and pieces of landing gear might be recognizable but not much else. The resulting debris field will likely be spread out in the direction of any current(s) that acted on it during its descent through the water column.
I think such a debris field will be larger than AF447. It might look like a field of small bright objects with a few somewhat larger ones.
If the arcs are accurate then I’m worried that the debris has been missed.
@TBill
RE: I think we need to realize the Qantas historic accident involved an older heritage metal gas cylinder, that If I recall was, upon further investigation, produced during a known batch of bad quality metal cylinders.
The ATSB could not determine why the QF30 cylinder failed. Other cylinders from the same batch were subjected to destructive and non-destructive testing, but no systemic flaws or manufacturing defects were found.
Turner’s presentation suggests the use of a non-approved leak detection fluid (ie soapy water) caused corrosion and subsequent failure of the valve assembly. In that scenario, it seems likely the valve assembly would be projected rearwards through the MEC and the bottle would be projected forwards, creating a gaping hole in the fuselage. His notion that the bottle would ricochet around the MEC without penetrating the fuselage doesn’t seem plausible.
Someone posted exactly the same scenario on PPRuNe about three years ago, under the username “GBO”. I can only assume it was Turner.
https://www.pprune.org/australia-new-zealand-pacific/645199-mh370-new-news.html#post11347380
@Brian Anderson
@370Location
RE: The flutter possibility that potentially caused the flaperon to detach is not the torsional [aeroelastic] wing flutter that some videos illustrate. Rather it is the rapid and violent excursions of a control surface, more like a flag fluttering in the wind.”
Indeed. It may have been caused by control-surface buzz, a single-degree-of-freedom flutter sometimes encountered at high subsonic and supersonic speeds. It occurs very rapidly and is destructive.
@Ed
For the clean cases I agree with the use of FFT you describe above. For the dirty cases, I do not agree.
If you compare with long distance transmission in optical fibers, it would be totally impossible without careful dispersion compensation. That is typically done using dispersion compensating fibers. The same can in principle be done in your case if it were possible to convert R-wave signals to signals propagating above the jump layer (and above jump layer signals to R-waves) roughly midway between the transmitter and the receiver. Then you would receive a signal with small dispersion.
I was involved in an EU project (METON) 30 years ago where we investigated crosstalk in non-ideal cases (among many other things) for optical ring networks. There someone made a solid proof that non-ideal dispersion compensation did indeed cause temporal crosstalk for multichannel transmission (which in your case would give the wrong headings), so I am sure you are wrong. I think it was a guy from Telia, but I no longer remember his name.
Many participants in the project were surprised about the result because it is counter-intuitive (like you are). However the proof was correct (experimentally tested).
@Mick Gilbert,
In response to your questions :
1. Is Arc 7 accurately mapped? Yes, with very high probability.
2. Was MH370 in unpowered flight at Arc 7? Yes, with very high probability.
3. Did MH370 follow one of Boeing’s simulated profiles? Probably not, because the simulations did not include relight attempts and did not show high downward acceleration, and because the altitude and speed upon fuel exhaustion are unknown. That said, I think it is highly likely the aircraft impacted the sea very near Arc 7 (almost certainly within 25 NM).
Therefore, in my opinion OI is highly unlikely to be successful in their current zone of activity, which is roughly 35-50 NM from Arc 7.
I believe any area more than 25 NM from Arc 7 is the lowest probability of all potential search zones. I would prioritize such distant zones, at all latitudes, well below (a) unsearched and marginally searched areas near Arc 7 from 33-36S and (b) unsearched areas near Arc 7 from 32-40S.
At the beginning of 2025, there were some missed and incomplete areas near Arc 7 on both sides of the UGIB LEP, but many more to the SW. OI may have covered some or most of those to the SW in March, but I don’t recall them venturing to the NE of the LEP then. I believe there are still some small areas of incomplete and missing data which are near the Arc and to the NE of the UGIB LEP. OI appears to be putting those at a lower priority (if indeed they plan to go there at all) than scanning far from the Arc and south of it, as they are doing now.
What OI is doing now does not make sense to me, unless OI has additional data near the Arc that we don’t know about.
@DrB
Thank you for the detailed response, Bobby.
It’s clear that I could have done much better with the wording of 3.
@ChrisV3141 asked: How accurate are the arcs likely to be? I’m thinking of a +/-km distance normal to the arc. If the expected error is only +/-10km or +/-20km then it’s hard to understand how the aircraft wreckage hasn’t been found yet. Given the BFO data indicating the aircraft was likely in a steep dive.
I performed an analysis of the BTO error when the SATCOM logs on. The full analysis is described in this post:
https://mh370.radiantphysics.com/2020/02/01/search-recommendation-for-mh370s-debris-field/
The summary from that article is:
The uncertainty in the BTO produces a corresponding uncertainty in the position of the 7th arc. The calculated sensitivity of the arc position to the BTO is 0.104 NM/µs, i.e., a 1-µs increase in BTO pushes the 7th arc outward (southeast) by 0.104 NM. The 1-σ uncertainty of the arc position due to BTO noise is therefore 0.104 NM/µs × 29 µs = 3.0 NM
@ChrisV3141 Re: “How accurate are the arcs likely to be?”
I think the best accuracy estimate comes from this Thales study:
https://bit.ly/4jmFsis
From this study, the 99% confidence level accuracy is +/- 5.4 nm. However, altitude uncertainty adds +/- 1.5 nm. So, call it +/- 7 nm. ( I haven’t checked, but I believe this is consistent with Victor’s estimate. To this one must add the post 7th arc descent profile. Given the BFO derived descent rate at 00:19:37 (~15,000 ft/min) and missing IFE login 89 seconds later, it is more than likely that MH370 is within +/-20nm of the nominal 7th arc. Of course, much of this area has been searched, so searching a little wider makes sense now.
Thanks for the info and responses re: ‘captain Peter Turner’. Curious that he seems to have used various aliases. I do however, enjoy the youtube format of the presentation.
I also came across another captain, Simon Hardy. He appears to be more credible, having worked closely with the ATSB in the early days of the search. He has 3 videos covering his methodology: https://www.youtube.com/@simonhardy4705
do these still hold up to this day? Have they been reviewed on this site? thank you
Armada 8605, now near (-35.2,93.4) has barely moved in the past 13 hours. I wonder what operation they are conducting. Perhaps the slow movement is a good sign.
@all
Yes, we see also that Armada 86 05 has been staying in a 0.4 Nm radius circle for basically 16 hours since 8:45 UTC and was still within at 00:32 UTC. We suspect ROV operations but, it is just an assumption.
https://www.mh370-caption.net/index.php/armada-tracking/
@Marijan
“A location closer to the 7th Arc does not contradict your work and proposed flight path…”
On the map we just gave the part of our recommended zone which was not covered. You can find preciser info on the zone and contingency area in our report. We assume that the previous searches were done properly. We are not in a position to evaluate anything 🙂 Armada is relatively “static” at 8Nm East of our recommended zone.
@370Location
” If the .kmz.gz file could instead be a .kmz file, it might be possible to turn it into a live network link within Google Earth Pro. The .kmz is internally a zip compressed .kml file, so no savings with the gzip compression…”
Unfortunately my web server host does not allow me to upload .kml or .kmz files…thus the trick of a .zip . Sorry for this, we do our best 🙂
@Victor
” … in the same bin as the CAPTION (CAPTIO, or whatever that group calls themselves) path reconstruction …”
I think you have been unfortunately misled, please let bring some clarity in. 🙂 Captain Blelly was not part of CAPTIO, I am the one who participated to both. CAPTIO worked on the starting hypothesis that a third party was involved and aimed to land. It stopped after elaborating the best trajectory under this working hypothesis.
“which also has an impact point that moved towards the proposal of others….”
No, right from the day of the disappearance, Captain Blelly has been working on the hypothesis of the pilot as the prime suspect aiming to get lost in the middle of the SIO. Later after my first group finished its work, I joined Captain Blelly and we created CAPTION to elaborate his findings deeper. No “move” towards others’ proposal, CAPTION always worked on the “suicidal pilot” hypothesis. I hope it is clearer. 🙂
“And those guys then have the gall to claim their work was copied.”
No, we just raised the question where the non-explained 70Nm extension came from. It appeared suddenly and came after your claim that the aircraft could not have been piloted and even less until the end, that the aircraft crashed at high speed close to the 7th arc and lost its flaperon due to fluter… we could not figure out how it could have flown these extra 70Nm by itself because it could not come from your analysis results showing no possible gliding. And suddenly you assume it glided which is in full contradiction of your analysis results ? ( CAPTION assumes much earlier that the pilot voluntarily stopped the second engine to save fuel for the APU and allow him to have all control surfaces functional during the gliding).
This was our point and still is actually. May I ask you if you would be so kind as to elaborate on the rational for this extension by 70Nm? If not piloted, how could the aircraft fly 70Nm alone with no fuel ? If piloted then your full analysis should be revised as you assumed it was not piloted and also what scenario during the gliding ? …with no fuel and thus with only two control surfaces : the right flaperon (thus it did not brake due to fluttering) and one elevator ?
I do hope you better capture our point 🙂 Cheers.
@Jean-Luc Marchand – CAPTION
Regards, “We suspect ROV operations but, it is just an assumption.”
I suspect that ROV operations might be somewhat challenging. In the high quality video of Armada 86 05 taken while bunkering in Fremantle, there was no sign of an ROV, nor the tether management system, and associated kilometres of tether required to operate one.
Maybe they’re onto a great fishing spot!
@Victor,
An notable 24 hours. I believe Armada 86 05 began an AUV launch procedure at 2026-01-04T0120, then tended that AUV during a 25km transit to the potential area of interest, arriving at 2026-01-04T0840. Within the area of interest, A86-05 has been tracking the AUV’s navigation with approx 200m line separation.
Each phase of the search, apart from the Jan-Mar 2025 mission, returned to specific areas to further investigate details revealed in the side scan imagery. I look forward to news of the investigation over this area of interest.
@J-LM, there are no ROVs onboard Armada 86 05. Unless you believe ROVs have been disguised as refrigration CCUs.
@Jean-Luc Marchand: Stop the bullshit. You (as part of CAPTIO) were proposing a POI near Christmas Island. Now you (as part of CAPTION) moved your POI close to UGIB 2020. That’s irrefutable.
As for the “unexplained 70 NM extension”, again your claims are not truthful. Look at the search area we proposed in Feb 2020 in Figure 4 under the article called “Search Recommendation for MH370’s Debris Field”. In fact, in that article, we consider three possibilities: 1) Unpiloted after fuel exhaustion 2) Controlled guide south 3) Glide in an arbitrary direction. Each of these scenarios has an associated search area, prioritized in that order.
https://mh370.radiantphysics.com/2020/02/01/search-recommendation-for-mh370s-debris-field/
Figure 4 shows a search area centered on UGIB 2020 with a radius of 140 NM!!! corresponding to the maximum glide distance. The 70-NM radius was actually a reduction (and not the extension that you falsely claim) based on estimates of energy loss from a steep descent prior to a glide and a less-than ideal glide speed.
As I’ve been saying now for over MANY years, we believe there is evidence that the plane impacted relatively close to the 7th arc, so begin the search there. However, the evidence is not bullet-proof. So, if not found close to the 7th arc, search outward, as a glide is possible.
If you continue to whine about how your work was copied when clearly you moved your POI from Christmas Island towards UGIB 2020, you’ll be banned. I have no patience for the drama you are creating based on your false statements and delusions.
@Victor, Don, et Jean-Luc,
RE: “has barely moved in the past 13 hours”:
I was previously more interested in the “constellation” formed by the tracks of the mother ship approximately 30 km nor-nor-east, roughly at a bearing of 10 degrees.
In that case the mother ship had moved away from the region and then turned back and proceeded at close to cruise speed for about two hours before resuming limited range movements for the next 14 to 15 hours or so, including the first half of the 2nd January, Zulu time.
@George G
George, if you were interested in that, you’d have to be fascinated by this.
Just over 21 hours now of loitering in a 1 km² area.
Well, if they find MH370 there I shall have to eat my laptop.
According to the latest status report provided by MH370 Families, operations suspended due to weather. The loitering would appear to be them just trying to ride the conditions out.
Of course, it’s also possible that it’s simply a coincidence that the crew chose to stand-off Armada 86 05, during unfavourable weather, over an area where this phase of the search began on Dec 30th-31st.
@Don Thompson: Pausing launching/recovery operations and also loitering during a mission to further investigate an interesting contact are not mutually exclusive. We’ll know soon enough what’s going on.
@all
The Armada 86 05 remained within a one-nautical-mile radius for over 32 hours (from 8:43 UTC yesterday, January 4, until 17:08 UTC today). It is still there, maneuvering at very low speed (approximately 0.1/0.2 knots on average).
@Don Thompson
The photos published by Airlines News show two white boxes on board the ship, each under a crane. Their size (compared to that of two crew members nearby on the tanker moored alongside the Armada 86 05) corresponds well to the dimensions (1.8 m x 1.9 m x 2.85 m) of the Saab Seaeye SR20, for which Ocean Infinity is the known customer. Above and below the top of the cranes, the orange cylinders appear to be the cable/tether wheels.
I’d be surprised if OI didn’t bring ROVs (1 + 1 spare) to immediately inspect sites of interest if needed. Otherwise, they’d have to make multiple trips to Perth and lose at least 10 days, not to mention the associated costs.
Otherwise, it would mean they’re convinced from the outset that they won’t find anything 🙂
Happy New Year to you all,
Just to say that the area that Armada 86 05 is currently located appears to have been previously searched by Armada 78 06.
Maybe the reason for loitering is genuinely due to weather?
Correction :
@Don Thompson
My mistake, sorry. You are right, the white boxes are cooling containers. I could get hold of a picture of Long Teng Industrial items at
https://www.longtengindustrial.com/reefer-container/10ft-carrier-cooling-units-reefer-container.html and a better Airlines News image. They look alike.
Their dimensions are close : 2991x 2438 x 2591 mm…
So maybe they are in the two blue comntainers at the rear just in front of the Hugins boxes…
@PaulS
Wake me up before you swallow it. I am expecting numerous “hopeful” moments but I am trying to be immune to that. It’s just amazing with the complete secrecy and lack of transparency over the last 2 years or so, the mass media and others somehow derive we are close to the wreck now. If my short-term memory serves Geof Thomas had the ROV’s out on sea floor in today’s episode.
@Godfrey Jack said: Maybe the reason for loitering is genuinely due to weather?
That’s the most likely explanation, but who knows?
@TBill: You just dashed my hopes! That crew consistently gets everything wrong.
Addendum:
@Don Thompson
Second thought : The cooling containers might be keeping the ROVs in a fresh, controlled environment… as their internal volume is 2390x2294x2165 mm3 which could be able to host a folded ROV.
(I would find very strange that their strategy would be 1) find and 2) go back and retrieve pieces if required)
Why not speculate ?
I wonder where three little, or not so little, blind, certainly not so blind, mice are whilst the surface ship is riding out rough surface conditions ?
What I found strange was how armada headed straight to a specific spot to wait out the bad weather. Why not just stay where it already was?
I’ve been tracking the Armada since its arrival and have come to wonder a number of things. Thanks in advance to those willing to take a moment to help with my questions:
1- how far away are the underwater drones operating from the ship itself?
2- could one say that the Armada’s location does not accurately depict where they are scanning?
3- the drones are operating in 72hr cycles. How can we track when those cycles begin and end?
4- it is my understanding that the drones can not live time transmit data back to the ship. Please correct me if I’m wrong. If correct then couldn’t one say that we would get an update on its progress only once every 72 hours?
5- @ChrisV3141 you mentioned that it’s more than likely the engine and landing gear could be the only large pieces to be made out and the rest broken into little pieces. This is the first I think I’ve ever read about the sizes of debris one can expect to find. Air France 447 apparently had large pieces of the fuselage in tack. Any additional thoughts?
6- I am still trying to wrap my head around the “55 days” of search. Is it 55 days of on site search? If the weather holds up could they search for more than 55 days?
7- What happens if after 55 days we are all still in the dark about not finding the plane and what happened?
8- Is there a typographical map of the current search areas?
9- Wasn’t there talk of a spreadsheet that had a running list of locations where people thought the plane would be found? If not, could it be revisited? And perhaps if there are any concerns about copying work it could be kept anonymous?
I’ve been glued to this site for the past decade. I appreciate all the bright minds who repeatedly chime in on their technical analysis – of which I have no skill in that area. I extend my thanks to those that occasionally veer off into laments terms.
Cheers everyone.
Victor, you’re confusing CAPTIO and CAPTION.
CAPTIO was initiated by Michel Delarche and Jean-Marc Garrot, and Jean-Luc joined them before distancing himself from their theory. If he had chosen a name less similar to CAPTIO, it would have been less confusing.
I’ve followed all your work, and yes, you mentioned these three possibilities, but in 2020. Before that, you never considered the piloted gliding hypothesis truly credible. Even today, you only hint at it. Perhaps to say later: “We told you so.”
Yes, but it’s so timid…
As you say, you favor a crash site near the 7th arc, but the area suggesting this was largely explored during the previous OCEAN INFINITY campaign. So you considered looking further south now, but I remind you, also further north, inside the arc.
For me, looking inside the arc is incomprehensible. Even though I’ve read your studies on this point in detail. I don’t share this opinion, but it remains a personal position.
I remind you that your latest version of the recommended area didn’t initially include CAPTION’s, and this was done without any rational explanation, other than to reiterate CAPTION’s arguments without publicly acknowledging it. It’s because some members of the IG must have considered this area technically viable, I suppose. Whether you accept this or not… that’s not the point.
Regarding the 70 NM glide, you’d have to explain to me how a 777 can glide for 70 NM and be subjected to stresses that cause flaperon flutter to the point of it breaking off.
Let’s be logical: if the flaperon was subjected to such stresses, it means the aircraft entered a high-speed dive, and it would never have glided for 70 NM. Never.
Furthermore, the right flaperon is the only flaperon powered by the RAT in the event of a total engine failure. Therefore, it has sufficient hydraulic pressure to prevent flutter.
So, I don’t buy the idea of a 70 NM glide with a broken right flaperon, sorry.
I don’t think JL is creating a drama with his comments. He’s clarifying his technical opinions and even correcting your statements on CAPTION/CAPTIO.
Season greetings to everyone!
I discovered this place a few years ago and have read every post and comment since the beginning … all due respect to Mr. Ianello who manages to keep this place at a high level of credibility … It’s honest to say that my only “expertise” is a love of popular physics and almost 30 years of experience in free flying …
I will still venture one comment…
Without ROV equipment on board, this whole mission seems pointless to me, does that mean that if debris field is found, it can take 10 plus days at least until it is investigated? And if I understand correctly that the home port is Singapore, that it can be months? Or if something is found, say on the 53rd day of the mission, and A 86-05 has to depart on a previously contracted task, what happens then…
It simply doesn’t make sense to me that the most important underwater discovery since the Titanic (and certainly more important in scientific terms) would be left on the seabed for who knows how long… please understand my doubts and excuse the long post.
@Theo
I am neither a meteorologist nor do I have nautical education, but the area 8605 resides in for the last 34 hours is further away from some weird-looking channel-like cloud formation / weather phenomenon which extends over a large path of the SIO than the previous slowdown points.If you switch to wind map on the links below, one further can see that the wind blows in a different direction inside this phenomena.
Weather satellite recording from last tracked slowdown point before moving to the current area: https://zoom.earth/maps/satellite/#view=-36.22,99.84,5.93z/place=-35.419,93.858/date=2026-01-04,06:00,+6
Same recording but with the first tracked slowdown point in the area the ship currently resides: https://zoom.earth/maps/satellite/#view=-33.62,103.08,5.16z/place=-35.499,93.597/date=2026-01-04,06:00,+6
Given that the ship currently rides out bad weather, maybe they wanted to be further away from this phenomena?
@Vlad.BG: The AUVs are capable of low passes over a debris field and acquiring high resolution imagery. An ROV can also retrieve parts, but that isn’t required for OI to identify the debris field.
@Gilles Diharce: I know that CAPTIO blew up as members have privately written to me out of concern that some of my observations were directed at them. (They are not.) However, J-L Marchand was a vocal member of CAPTIO and is now a vocal member of CAPTION, and he and Blelly are the ones accusing me and others of copying their work, which is batshit crazy.
First, you must understand the difference between a priority zone and a search recommendation. That seems to be lost on you. If you truly have been following what I’ve been saying for MANY years, it’s that anybody that assigns a high probability to any small area or to any single scenario, they are fooling themselves. My search recommendations are meant to be just that: a suggestion to Ocean Infinity where to search if the plane is not found in the areas I deem the most probable (i.e., the highest priority). I make the distinction between most probable (and therefore worthy of the highest priority during a search) and probable. My uncertainty includes whether the plane was controlled by a pilot after fuel exhaustion. I believe there is a higher relative probability that it was NOT controlled, but I never suggested to Ocean Infinity that they only search areas close to the arc. That’s why the search areas that Bobby Ulich and I have recommended have the priority levels they do. Ocean Infinity seems to agree, as that seems to be their search strategy.
Here’s an email I sent to Marchand last year politely asking him to knock off the bullshit of saying we are copying their work:
***
From: Victor Iannello
Sent: Thursday, March 21, 2024 10:54 AM
To: Jean-Luc Marchand
Subject: False Claims
Jean-Luc,
I hope all is well.
It has come to my attention that Patrick Blelly is accusing the IG of changing its predictions of the POI for MH370 based on his work predicting a glide after fuel exhaustion. This is a crazy accusation. Even before we published the UGIB 2020 analysis with search recommendations, we defined an area that included a possibility of a glide. For instance, this figure is from a blog article from February 2020
[Figure 7 from https://mh370.radiantphysics.com/2020/02/01/search-recommendation-for-mh370s-debris-field/ ]
From that blog article: To define the search area near the LEP, three cases were considered, each with an associated search area. The highest priority search area, A1, of 6,719 NM2 (23,050 km2), assumes there were no pilot inputs after fuel exhaustion. The search area of next highest priority, A2, encompasses 6,300 NM2 (22,000 km2), and assumes there was a glide towards the south after fuel exhaustion. The lowest priority, A3, is the controlled glide in an arbitrary direction with an area of around 48,400 NM2 (166,000 km2).
Our recommendation has been to first complete the search close to the 7th arc in areas that were missed due to equipment issues or challenging terrain. If the debris field is not found, search wide of the arc, first starting on an area south of our predicted crossing at S34.2342° E93.7875°, but expanding that to include the possibility of a longer glide.
If somebody asks me to define a small area for a search, I point them to this article, which includes about 30.5 km2 of seabed surrounding S34.53° E93.84° that was never scanned. But this is a recommendation, not a prediction. Anybody that believes that they can predict the location of the POI with high certainty is fooling themselves and fooling others. That’s why any search prediction has to include uncertainty analyses and broader areas to search.
[Fourth figure from https://mh370.radiantphysics.com/2023/02/11/a-high-priority-area-to-search-for-mh370/ ]
As you might imagine, I don’t like being accused of something that is false. I assure you that any of our recommendations were made independent of any work by either you, Patrick Blelly, CAPTIO, CAPTION, or any related person or entity.
Regards,
Victor
***
As far as I know, Blelly published his book in March 2022 and the flight reconstruction with Marchand in February 2023. That means their work was published years AFTER we made search recommendations, both publicly and privately. At the time that we published our search recommendations, Marchand and the rest of CAPTIO was suggesting we search near Christmas Island. Then 3 years later, Marchand changes his POI from Christmas Island to a path that resembles UGIB 2020 and falls within our recommended search area, with the ludicrous claim that we copied them. I can’t state strongly enough how delusional this is.
If you disagree, and wish to continue the drama, please take it elsewhere. The historical record, including private discussions we’ve had with Ocean Infinity, speak for themselves.
@Theo
You asked: “[Simon Hardy] has 3 videos covering his methodology…do these still hold up to this day? Have they been reviewed on this site?”
I believe Hardy proposed two crash sites. As I recall, the first location was examined towards the end of the ATSB’s search, without result. The second remains unsearched. There was some discussion about Hardy’s work; I understand the second crash site was ruled out by both the drift modelling and fuel modelling.
Hardy also made comments about last minute additions to the aircraft’s fuel and oxygen loads, claiming they were evidence of foul play. Those claims were reviewed and I strongly believe they are wrong. My thoughts on those claims start here: https://mh370.radiantphysics.com/2025/03/31/update-on-the-search-for-mh370/#comment-39135
@Theo
Regarding Captain Harry’s three videos, I can recall that when I first watched them some years back I was struck by the fact that he uses a constant ground speed across the whole flight route.
There are many things that we don’t know about MH370’s flight into the Southern Indian Ocean, but one thing that I think that we can be assured of is that it did not fly at a constant ground speed for 4-5 hours as it passed through a variety of weather conditions.
I would credit Simon Hardy’s videos as teaching me how to start doing flight paths. By the time I belatedly got started say 2017, the 38-40s search was looking wrong based on the Bayesian principles more popularly cited back then, and I was very much influenced by Victor/Richard’s papers about more northerly theories. Hardy’s path very helpful I never gravitated to actually use Hardy’s path. I did move back westerly over as far as to 180s, which I still see as how it probably started from ~ISBIX.
@all
Armada 86 05 has left the small 1 Nm zone at [-35.49;93.59] and is now en route at [-35.12;93.73] which is close to where she was on 3 January.
Excuse my lack of technical expertise, but I have a question about the final glide (assuming the plane reached fuel exhaustion at the 7th arc and glided).
When we talk about this glide, is this simply a gentle, steady glide into the ocean? or would the plane glide in a ‘phugoid’ pattern? (ie. where the plane naturally descends, picks up speed, then naturally climbs again). And would a gradual glide vs phugoid glide produce very different max. glide distances? I think 70nm is the current extent for the glide, but I’m not sure what ‘type’ of gliding this is.
Ah yes, the old ROV in the refrigerator trick … the second time I’ve fallen for that since Christmas.
re: ROVs, here is the best resolution photo I could find of Armada 82 05 refuelling in Freemantle, Dec ’25.
https://berqwp-cdn.sfo3.cdn.digitaloceanspaces.com/cache/42kft.com/wp-content/uploads/2025/12/Armada-86-05-23122025-03-1024×767-jpg.webp?bwp
I’m no expert, and hate to speculate when someone else here probably knows, but there appears to be 2 crane/gantry systems with orange lifting devices(?). The white containers under each one have a logo on the side I can’t make out.
Here is a video from the same moment, timestamped to a close-up fly-by of the ship: https://youtu.be/mrCbP2f3eSc?si=NIngWze-9SGTTMpB&t=223
@Mick Gilbert: ROV = Refrigerated Orange Vehicle?
@Gilles Diharce
The three search areas recommended in the UGIB paper are depicted in the following diagram as A1, A2 and A3, in that order of priority. CAPTION’s recommended search area is overlaid in orange.
UGIB vs CAPTION Recommended Search Areas
UGIB was published in February 2020, three years before CAPTION published their own recommendations. The notion that expanded search area recommendations were only made after CAPTION published their study is categorically wrong.
@Theo
You asked: “When we talk about this glide, is this simply a gentle, steady glide into the ocean? or would the plane glide in a ‘phugoid’ pattern…And would a gradual glide vs phugoid glide produce very different max. glide distances?”
Theoretically, the maximum gliding range of a B777-200ER is 0.0034 NM per foot of altitude, per Boeing’s performance analysis in Appendix 1.6E the MH370 Safety Investigation Report (SIR). Accordingly, the maximum gliding range from FL410 is about 139 NM, if there were manual control inputs to maintain wings level flight and the optimum glide speed during the descent. That theoretical figure is the upper limit on the gliding distance. A maximum gliding distance of 140 NM was used in UGIB to define the outer limit of the recommended search areas for the glide descent scenarios.
In practice, the gliding distance is likely to be much less than the theoretical figure:
– If the descent was unpiloted, the aircraft is likely to have entered a spiral descent, with or without a phugoid oscillation. The simulations that Boeing describes in the SIR were contained within a 10 NM x 10 NM box that extended beyond the point of fuel exhaustion. Some of those simulations exhibited a phugoid oscillation, some did not.
– If the descent was piloted, there are many conditions that would have significantly reduced the gliding distance. Those conditions include drag from the windmilling engines, sub-optimal gliding speed and less than ideal atmospheric conditions. The Ulich-Iannello drift paper concluded that it is highly probable the aircraft crashed within a 70 NM zone in the event of a piloted glide descent, but also noted that a 140 NM boundary is almost 100% certain to contain the crash site.
@Theo
I agree with Andrew completely except I would add for certain “curved” paths there is opportunity to take an active-pilot glide/descent before Arc6. Now you save some fuel at Arc7 but still 140nm distance might be hard to exceed because fuel efficiency is lousy at lower alt.
I would like to remind the glide enthusiasts that even if you completely dismiss the BFO data, you still have to deal with the missing IFE login 89 seconds after the 7th arc. That missing IFE login strongly suggests the plane was either: (1) in the water before 00:21:06 or upside down at 00:21:06. Either way, it can’t be far from the 7th arc. For reference, that’s about 14 nm at 550kts for 89 seconds.
ROVs, the (actual) absence of.
@J-LM Monsieur, monsieur, s’il vous plaît
No ROVs have been mobilised on Armada 86 05.
J-LM suggests that Armada 86 05 may be carrying an ROV stored in one of the 20ft CCUs just forward of the AUV ‘garages’ at the stern of the vessel. This is a preposterous suggestion, a) that an ROV is stored in one of the two CCUs fixed just forward of the AUV ‘garages’, or b) that an ROV stored in one of these ROVs could be manoeuvred across the deck. Viewing the still and video images captured Fremantle, Cockburn Sound-ORAN it is obvious that no safe method could possibly exist to manoeuvre, while at sea, an HD WROV from inside one of these CCUs to a position where it could be integrated with one or other Seaonics LARS.
The best quality images, as per Theo above, show that 10ft refrigerated/‘reefer’ CCUs were fixed to and lashed down on the moonpool hatch under each Seaonics LARS rig. While these CCUs have been confirmed as refrigeration units (likely source here), various parties perpetuate a false narrative, associating that Armada 86 05 is carrying ROVs.
This ‘ROV’ narrative doesn’t add anything to A86 05’s present mission. With its Hugin AUVs, the vessel has the capability to image the seafloor, identify targets of interest, and revisit any likely targets for detailed imaging using side-scan sonar and an AUV’s high definition visible spectrum camera.
Considering the topic of ROVs in more detail: a heavy duty, work class ROV would be necessary to operate at any possible 7th arc depths, a maximum of 6000m. Ocean Infinity operates two such ROVs on the Armada fleet at present: one on Armada 78 01 and one on Armada 78 06. Both vessels are actively engaged on subsea infrastructure inspection projects in the North Sea. A third HD WROV is being mobilised onto A86 03 which is presently in port at Blyth, UK. Each of these WROVs is deployed with a ‘top-hat’ Tether Management System (TMS). The all-up weight of a HD WROV & TMS is in the order of 7.5t, moving that on deck is not a trivial task.
Further, it has been suggested that the purported hidden ROV may be a Saab SR20 eWROV. While Ocean Infinity has been announced as the Saab SR20’s launch customer, the SR20 has yet to be seen mobilised on any Armada vessel. To date, different types of ROVs have been mobilised on Armada vessels: Schilling HD WROVs, Kystdesign Supporter, Saab Seaeye Cougar, Saab Seaeye Panther, and Saab Seaeye Leopards on A78-02, -07, and -08. But the Saab SR20 has yet to be seen ‘in the field’, Armada 86 03 and 04 may ultimately be the vessels that see mobilisation of the SR20 for a project requirement yet to be delivered.
The ‘orange’ components in the Seaonics LARS are the system’s cursor beam and payload attachment. The cursor beam holds the payload rigidly in place within the LARS mechanism. The beam translates on vertical rails downward through the moonpool where, clear of the vessel, the payload is released from the cursor beam and ‘traditional’ winched umbilical handling takes over. In reverse, the payload is winched up to mate with the lowered cursor beam, the mated assembly is then raised with the payload through the moonpool. The moonpool deck hatches close under the payload permitting the payload to be lowered to the deck hatch and, if necessary, be skidded off to the starboard side.
@Theo
G’day Theo, thank you for the image and video. As Don has pointed out a number of times now, the items on the deck under the gantries are refrigerated containers (https://cargostore.com/10ft-refrigerated-container/).
If an ROV was embarked, then the associated tether (about 7-8 kilometres worth) and the tether management system (TMS) would also need to be carried; you can’t RO the V unless it is connected by tether back to the host vessel. Neither of those critical components are in evidence on deck.
You might treat this discussion as a microcosm for the broader dysfunction generated by a few, let’s just call them “other-than-good actors”, who propagate mis- and disinformation through the MH370 community.
@Andrew, you said “a 140 NM boundary is almost 100% certain to contain the crash site.” That is only true if you have identified the correct segment of 7th arc.
@Mike’s remark immediately below provides yet more evidence that this is NOT the correct segment of arc, because the wreckage has not been found within the zone already searched 30NM either side of the arc.
As several have pointed out (including UGIB lead author), a piloted glide at this latitude is ruled out by fuel endurance => bleed air off. And an unpiloted glide with no power and disengaged AP can’t go far. If that weren’t enough, we also have the final BFOs and the incomplete login.
I cannot see any logical basis for prioritising this search area.
@Paul Smithson: This search will more or less cover the scenarios in which MH370 crossed the 7th arc between latitudes 35S and 33S, irrespective of whether the aircraft was piloted at fuel exhaustion. I am less certain than some of my colleagues on this point (unpiloted at fuel exhaustion), although I do believe that the areas closer to the 7th arc should be prioritized for the reasons you mention. If the search fails, we can eliminate these latitudes from further consideration and move to other areas. You are advocating move to the south. Ed to the north. We’ll have that discussion if nothing is found. I don’t agree that there is no logical basis for the current search. It’s part of a logical progression of eliminating areas with the understanding that there is uncertainty in our best models.
@Victor Iannello
The current search extends to about 50 nm from the 7th arc (95 km according to the coordinates from the Caption tracking site), and I think it’s unlikely they will search further. To properly cover the glide scenario in this latitude region, shouldn’t it require more like 70-100 nm at least? I’m not saying that the current search is insignificant when it comes to the glide scenario, but it seems very far from exhaustive.
@Paul Smithson
My comment was an answer to @Theo’s question about the nature of the glide descents mentioned in various discussions. It was not intended to favour piloted descents over unpiloted descents. The “almost 100% certain” comment was mentioned in the context of the Ulich-Iannello search area proposal. It was not intended to imply anything about their identification of the crash latitude.
Judging by the geographical coincidence between the current search area and the DSTG’s pdf, another interpretation would be that OI have taken the UGIB/First Principles guide on best latitude and are searching out to the 99% probability contour. See here.
For information, the zone that I advocate (brown circle at the at the SW end) lies pretty much on the 95% contour according to this pdf. Moreover, its the only bit of 95% contour close to the arc that remains unsearched.
If the pdf were redrawn by BTO probability rather than BTO+BFO, it would be obvious that my recommended hotspot actually sits near peak probability – not out at the margins.
@Paul Smithson,
on January 6, 2026 at 6:45 pm, you advocate a “brown” circle as a search region of interest.
Please be assured that I am not in the position of “eating my laptop”.
But:
1. Would you please remind me, or us, how the aircraft had fuel still to consume until the time [I will accept 00:17:30 on the 8th March 2014] and distance it had to reach your brown circle.
2. As my memory, own data searching capabilities, may be deteriorating, would you also please kindly plot your estimated aircraft positions at the time of all four “Arcs”, or handshakes. A simple set of additions to an expanded version of your existing simple chart should suffice.
Sincerely, George G.
Woops ! Apologies ! … of all seven (7) “Arcs”, or handshakes
To illustrate my previous post, see here a plot of the probability distribution along the 7th arc using BTO+BFO (which is what was used for the final pdf and the definition of the search area) and compared to the probability distribution derived from BTO only.
As you’ll see, the southern extent of the search zone at 39.5S and the 95% confidence contour on my previous link correspond with the outer tail of the BTO+BFO pdf, as you’d expect. However, primary peak of probability distribution using BTO only falls at 39.3S. So the search stopped pretty close to BTO peak-probability zone, and left a bunch of the BTO pdf unsearched.
My research found a BTO pdf that replicated the primary peak around 39.2S but didn’t extend beyond 40S (DSTG’s only did because it was initialised at 1802 and permitted a small number of paths to turn too early, before Arc 1).
@eukaryote: You can see in the figure at the top of the article that the 70-NM glide limit extends past the proposed search area, so you are correct that the search is not as exhaustive as I implied.
@All: The figure at the top of the article was updated to show the further progression of the search. The AUVs are now searching due south of the areas we designated as the UGIB 2020 Last Estimated Position (LEP) and the High Priority Search Area (HPSA).
@Jean-Luc Marchand – CAPTION
It’s a real pity that in your work you opted for the controlled end-of-flight scenario which in turn favored locations further from the 7th Arc. I guess it is a bit too late now. In the case you did, we would probably have area along your proposed route re-scanned during this search, similar as IG’s.
@Victor Iannello
It is also a shame the the area between 35°S and 36°S, 25NM from the 7th Arc will not be re-searched, at least not for now, but it is what it is.
@Victor Iannello and @airlandseaman
Thank you for your responses to my question reguarding the accurracy of the arcs. Your answers were exactly what I was looking for.
@JJP17
AF447 hit the water somewhat level at a speed of 152 knots. With a vertical speed of 108 knots and horizontal speed of 107 knots. MH370 is believed to have impacted the water in an attitude closer to nose down vertical with a speed close to Mach 1.0. Crashes on land under similar circumstances tend to produce few large pieces.
Egypt Air 990 might be a better analogy. In this case the aircraft impacted the water at high speed after having shed its left engine and some other pieces (possibly due to maneuvering). In that case, the crash resulted in many small pieces of wreckage but with some larger pieces of recognizable fuselage.
@George G. I you have a look through historic posts you will find ample material about my path analytics and reasoning including individual arc crossing positions, BTORs etc.
The route proposed is MEKAR-SANOB-IGEBO-RUNUT-4085S at FL360, M0.84
This produces excellent match to the BTO data and sits close to the narrow swathe of path optima that is predicted by BTO optimisation.
The path is fuel feasible if you assume bleed air savings of 3% plus ram air door 0.4% plus electrical load saving of 0.8%. You’ll also see in previous posts some justification for these numbers.
The “hotspot” that I have plotted is the intercept of this path with 7th arc at 20,000ft, with a 15NM range ring.
@Paul Smithson
What are your fuel remaining figures at:
MEKAR =
SANOB =
IGEBO =
RUNUT =
4085S =
@ventus45,
It doesn’t make it to “4085S”. That was Paul’s next waypoint, as I read it.
@George G
Yes, on closer reading you are correct.
I now place Paul’s Arc 7 crossing at approximately 392400S0851135E
I am trying to better understand several aspects of the 18:25 SATCOM logon event and would appreciate clarification on the following points.
Question 1 – Warm vs Cold Logon Behavior
My understanding is that if the aircraft’s SATCOM terminal has been unpowered for a sufficiently long period (likely more than ~48 minutes), a subsequent power-up would normally result in a cold logon. In such a case, the terminal would be expected to transmit both the Flight ID (e.g., MH000, MH370) and the aircraft registration (e.g., 9M-MRO).
However, the 18:25 logon appears to have been a warm logon, in which these parameters were not transmitted.
To date, I have not seen a clearly stated technical explanation for why a warm logon occurred instead of a cold logon under these circumstances. Is this understanding correct? If not, what technical mechanism or system behavior could explain why a cold logon did not occur following the apparent power interruption?
Question 2 – Extended Logon Response Time
Under normal conditions, the aircraft SATCOM terminal typically responds to the ground station within approximately two seconds during the logon process. In the case of the 18:25 event, the response time appears to have been roughly an order of magnitude longer than normal.
I am not aware of a published technical explanation for this unusually long response time, particularly given that after the 18:25 re-power event the SATCOM system appeared to operate normally.
Is this understanding correct? If not, what technical factors could plausibly account for the extended response time during the 18:25 logon?
Question 3 – SITA Log Data
To help further characterize the 18:25 logon event, the SITA logs may provide additional insight. For example, do the SITA records show a logon event from MH370 at or around 18:25?
As part of the MH370 SATCOM analysis, were the SITA logs examined, and if so, what information—if any—do they provide regarding the 18:25 logon?
Gents, stand by and I’ll post a 1-pager version of my path hypothesis this evening that will convey the key info.
With regards my prior 3 questions I should have mentioned if I could possibly gets Victors take on it.
@Paul Smithson,
Please include the time that you estimate the aircraft passed Waypoint MEKAR.
@George. This isn’t germane to my approach because all I assume is a) that the aircraft was on Arc 1 at 18:28:15 (BTO=12480); b) that it was somewhere in vicinity of N571. My modelling tested for sensitivity of the BTO-predicted path across a wide range of model initiation assumptions including:
– timing/BTO error at Arc 1
– lateral offset +/- 12NM from N571
– track +/- 10 degrees from N571 (295)
– speed 460, 480, 500kt groundspeed
– altitudes FL340, FL350, FL360 [higher predicts M>0.84 so not viable]
In practice, the BTO-optimised paths change very little across this span of possible “priors”, producing a zone of optima 38.6 – 39.7 [all scenarios].
Beyond the “optimum zone” there is a confidence interval (CI) of diminishing probability. At the northern end, this CI overlaps with the area already searched – so can be discounted. At the southern extreme, the CI is truncated (about 40.0S max) by the condition that speed cannot exceed M0.84 and altitude [I assume] cannot be lower than FL350.
@Michael Collins:
If you want to understand some of insights we have gained from the Inmarsat data, including the log-on at 18:25, you should start by reading these two articles from 2017:
https://mh370.radiantphysics.com/2017/06/12/the-unredacted-inmarsat-satellite-data-for-mh370/
https://mh370.radiantphysics.com/2017/07/04/some-insights-from-the-unredacted-satellite-logs/
It might be better to ask your questions after digesting this material.
@MICHAEL COLLINS
Please refer to the report in which I detail, in simple terms, the analysis of the SITA logs and, more specifically, the sequence logic of Inmarsat’s LIDU, ISU, and SSU. You will find the answers to most of your questions.
At 18:25:27 UTC, the connection was of type 1 (1F) to the IOR satellite (the same type as the very first connection on the tarmac at 12:50:19 UTC, but to the POR satellite). Type 1F corresponds to a login request in the absence of flight information available to the Satellite Data Unit (SDU). Consequently, the information present at takeoff has disappeared since the previous request for login at 15:59:55 UTC was of Type 2 (2F+3F) i.e. with available flight info which appears in following messages afer take-off.
CAPTION demonstrated that the only way to clear flight information in flight is to perform a data communication reset in the FMS, via a manual keystrokes sequence through successive pages and menus.
The conclusion is that, between the restoration of power and the connection request, someone manually erased the flight information (probably to ensure that it would not be operationally recognized).
By correlation, this proves that someone was alive and in control of the aircraft at 18:25 UTC.
The very educative report is here:
https://www.mh370-caption.net/wp-content/uploads/MH370-CAPTION-Annex1-RAT-Deployed-Scenario.pdf
Have a good reading 🙂
@Marijan
You wrote:”It is also a shame the the area between 35°S and 36°S, 25NM from the 7th Arc will not be re-searched, at least not for now, but it is what it is.”
I agree with you.
Back in 2018 part of the OI search was inspired on the drift modelling that David Griffin had carried out based on the “Pleiades” satellite images showing clusters of floating objects. The areas considered at the time were outside the 20 nm distance from 7th arc if I remember well (four “boxes”), see:
https://www.atsb.gov.au/sites/default/files/2024-02/mh370_csiro-ocean-drift-iiil.pdf
In May 2018, I had an exchange with David Griffin as I noticed that the coverage close to the 7th arc was not optimal in the S35.0 – S36.0 range. And I had my doubts when OI started moving far north. David indicated that S35.5 would be the approximate latitude if the debris originated from close to the arc. I actually wrote an email to OI.
It seems that the policy is to strongly rely on the results / coverage rates reported from previous scans. Which I do not want to criticize, given the lack of detailed insights that we have in all the scan data.
But for me a key point is that S35.5 near the 7th arc is the one of few latitudes for which there is an independent extra “clue”. And it is within the high probability latitude range indicated by satellite and fuel models. Plus the area has a somewhat lower sonar coverage percentage. And I feel that the evidence for a crash near the 7th arc (steep uncontrolled dive) is still standing strong.
@Jean-Luc said good point:
“The conclusion is that, between the restoration of power and the connection request, someone manually erased the flight information (probably to ensure that it would not be operationally recognized).
By correlation, this proves that someone was alive and in control of the aircraft at 18:25 UTC.”
Let me paraphrase:
Approx. 1-hr after the MH370 diversion at IGARI at 17:21, after the aircraft has already flown around Penang and is now heading up the Malacca Straits beyond Pulau Perak island and beyond waypoint VAMPI, the SATCOM then repowers at 18:25 with flight no. has apparently been manually deleted, presumably for stealth reasons. Also the ACARS text message and data logging every 30-mins has been switched off.
Therefore the implication of nefarious deliberate flight to 1825 point is strong:
(1) Controlled flight path around Penang up the Straits via VAMPI
(2) Re-power of SATCOM at 18:25
(3) Apparent Manual Deletion of SATCOM ACARS data ops/Flight No.
(4) Between 1825 and 1840 SATCOM BFO/BTO data dump shows maneuvers in progress
(5) No attempt to land at Penang despite relatively empty fuel tanks (close enough to acceptable landing weight)
above list is post-IGARI. At IGARI there is another list includes apparent Xponder deliberate outage (we now see that “smoking” gun from ADSB data) and no expectation of a immediate UTurn as proper commercial aviation procedure in any emergency scenario.
@TBill
One can be more precise and conclusive on some points you “paraphrase” 🙂
“…flight no. has APPARENTLY been manually deleted…” :
The flight number was missing, that’s a fact. In flight, there’s no other way to delete it than by manually resetting the COM. These elements, taken together, unequivocally prove that it was intentionally deleted by a human. There’s no other way to do it, and there’s no longer any doubt.
“…list includes apparent Xponder deliberate outage…” :
If you consult our latest data report on the ACT-Caption-2025 data (which is now the reference), you’ll find proof that the Xponder was manually switched to Std-by. There was no outage, as the other systems were functioning perfectly and the NUCp monitoring parameter reported a level at 7, i.e., in high-precision navigation mode (GPS). So there was no outage, but simply a manual rotation of the Xponder knob to Std-by. This is now a proven fact (it was previously only a supposition).
These proven facts are actually a solid ground.
“…Also the ACARS text message and data logging every 30-mins has been switched off…” :
It is less certain that ACARS was switched off per say. Here, the term “outage” would probably almost be appropriate. 🙂 The restoration of power to the left bus just before 18:25 UTC means that the power had been cut beforehand. We can assume (and Caption does) that, for the effectiveness of the hijacking, all power was abruptly cut just before the descent after IGARI (see our latest report on the vertical profile), thus causing a sudden stop of all active “transmitting” systems (IFE) and communication systems in particular. This is easy to do from the cockpit and could be very sudden. This would mean that the RAT only provided the electrical power (more than enough to allow an experienced pilot to fly, knowing that full hydraulic power was still being supplied by the engines).
Niels: Back in July 2021, I was confidentially informed that in addition to the Pleiades 1A images showing what could be debris from MH370 on March 23, 2014, the Italian satellite COSMO-SkyMed also detected what could be debris from MH370 on March 21, some two days earlier. The distance between the contacts in the images was some 35 NM.
I performed a “reverse-drift” analysis using the CSIRO drift results for the debris detected by both satellites, and determined that they appear to backtrack to the same position on March 8, 2014, at 35.4°S, 92.8°E, which is very close to CSIRO’s estimate of 35.6°S, 92.8°E using only the Pleiades data.
https://mh370.radiantphysics.com/2021/07/23/italian-satellite-may-have-detected-mh370-floating-debris/
If anyone can share the source data with me that is/was being used to prove or disprove the efficacy of WSPR, I’d love to get my hands on it, so that I can have a go myself. I have an agentic coding tool that I wrote that’s very good at things like this, and it would be both an excellent test, and potentially have a very important outcome one way or the other. Hit me up at robmoss2k at gmail dot com.
@Rob Moss:
There are MANY reasons why WSPR tracking is pure hogwash. Here is an article which shows how the physics behind radar disallow it by many orders of magnitude. In that article, I also reference the work of others, including Steve Kent (@sk999).
https://mh370.radiantphysics.com/2021/12/19/wspr-cant-find-mh370/
If you are looking for WSPR data, you can directly query the WSPR archive, which is available online in several formats.
@Victor
“… the Italian satellite COSMO-SkyMed also detected what could be debris from MH370 on March 21, some two days earlier. ..”
Thanks for reminding to the article. That’s indeed a relevant reference and addition. Is it correct that S35.6, E92.8 was included in the 2018 OI search and S35.4, E92.8 was not?
Hi Victor,
You mentioned it may be helpful to consider these questions after fully digesting the following material:
https://mh370.radiantphysics.com/2017/06/12/the-unredacted-inmarsat-satellite-data-for-mh370/
https://mh370.radiantphysics.com/2017/07/04/some-insights-from-the-unredacted-satellite-logs/
I have already reviewed both articles (thank you for sharing them). However, this is the type of technical material that benefits from multiple readings, so I have downloaded it for further, more detailed re-analysis.
After an initial review of the above, including a word search across both articles and their associated comments (523 comments on the first article and 688 on the second), I note that the term SITA appears approximately ten times. In all instances I could find, the references appear to relate to the company itself rather than to the SITA logs.
At this stage, focusing specifically on my third question—namely whether the SITA logs might provide clues relating to the 18:25 login—could you please clarify whether the IG, when analysing the Inmarsat satellite logs for MH370, also incorporated the separate SITA logs into its research?
@Victor
I’m trying to dig arond to find other reports of satellite derived images of debris in the SIO. In all cases, were reverse drift analysis of these ‘objects’ performed? Can we safely dismiss them given their extreme southerly latitude? Are any of them plausible at all?
1. There was a report of a Chinese satellite detecting debris in the SIO, with image capture on 18th March 2014 (3-4 days before the COSMO SkyMed image further north). The location is very far south of the 7th arc, based on the lat/lon axis markings on this image (https://cdn.mos.cms.futurecdn.net/hWwjVKGsm7ccdQCFxYWzD3-1052-80.jpg.webp) we get lat: -44.958 lon: 90.229
Here are a few articles that include this one:
https://www.space.com/25182-chinese-satellite-malaysia-jet-debris-photo.html
https://www.livescience.com/44228-photos-missing-malaysian-flight-370.html
(archive link to bypass paywall): https://archive.is/69fWK
This Telegraph video clip, relating to the SIO Chinese image: https://www.youtube.com/watch?v=cCtDmZ4-32w
2. An image captured on March 18th, released by Digtial Globe, of a different detection is shown here: https://www.livescience.com/44227-satellite-images-possible-flight-mh370-debris.html
The coordinates for this are lat: -43.976, lon: 90.960
3. Malaysian Remote Sensing Agency, taken on 23rd March:
https://www.spacesafetymagazine.com/wp-content/uploads/2014/04/Malaysian-Remote-Sensing-Agency-satellite-debris-image.jpg
Again, very far south.
4. And finally, the Thai satellite images of ‘300’ objects, taken on March 24th, with image broken into 5 zones:
https://gcaptain.com/new-satellite-images-provide-new-leads-in-search-for-mh370/
with a best resolution I could find of the image mosaic: https://gcaptain.com/wp-content/uploads/2014/03/mh370_3.jpg
and
https://www.bbc.co.uk/news/world-asia-26763358
Coordinates:
@Michael charles Collins
SITA’s messaging system log, related to 9M-MRO, was published as Annex 1.9A, ACARS Traffic Log, part of the ‘SAFETY INVESTIGATION REPORT’ compiled by ‘The Malaysian ICAO Annex 13 Safety Investigation Team for MH370’.
Separately, there is the satellite system log as recorded at the Ground Earth Station in Perth. Personally, I prefer to term the log linked from this blog (direct URL) the ‘Stratos Log’. Stratos was then the subsidiary of Inmarsat that operated the teleport services, including the Classic Aero Ground Earth Stations for the Indian Ocean and Pacific Ocean (IOR and POR) regions, at Perth.
The SITA ACARS Traffic log ceases to record anything further concerning 9M-MRO AND the satcom datalink after the satcom datalink failed to deliver a message submitted to the network by the MAS Operational Control/Dispatch Centre (OCC/ODC) at 18:03:23UTC. The message referring to ‘URGET REQUEST‘.
In turn, the Stratos Log shows that the aircraft failed to acknowledge the first part of this message, three times, and as a conseqence the Aircraft Earth Station (the AES) status was set as logged off the Inmarsat SATCOM network.
The 18:25UTC Log on from 9M-MRO comprised only a Log-On to the IOR GES, no ACARS traffic was initiated from the aircraft. The aircraft is expected to send ACARS Link Test and a Media Advisory messages when the SATCOM datalink becomes available and is the preferred medium (preferred over VHF, other media are supported by the network but not 9M-MRO). The Link Test confirms to the aircraft that an air-ground link is available, the Media Advisory indicates to the SITA ACARS messaging network that the aircraft intends use the media over which the message is transmited. It is the aircraft, its data communications management function not the SATCOM system, that initiates the ACARS connection over the required/preferred medium.
Neither the Stratos Log nor the ACARS Traffic log record a Link Test or Media Advisory message from 9M-MRO at any time subsequent to the 18:25 GES Log On.
However, closely following the 18:25 GES Log On two further virtual circuits were initiated over the SATCOM datalink. There were stated to have been initiated by the IFE system. These VCs opened connections to two separate X.121 addresses on the ground network, as had been evidenced following earlier GES Log Ons. These were not ACARS related connections. Technically, these connections were provided by the ISO8208 Satellite Sub-Network facility of the Inmarsat Classic Aero system.
Finally, yes, ‘the IG, when analysing the Inmarsat satellite logs for MH370, also incorporated the separate SITA logs into its research‘. In fact, I highlighted a significant inconsistency in the presentation of the ACARS Traffic Log and the Investigation team published a satisfactory amendment to correct the ACARS Traffic Log.
@Michael Collins:
The complete record of communications between airframe 9M-MRO and the Inmarsat satellite network, from March 7, 2014, at 00:51 UTC, until March 8, 2014, at 01:16 UTC. This time period includes the previous flight from Beijing to Kuala Lumpur.
https://mh370.radiantphysics.com/2017/06/12/the-unredacted-inmarsat-satellite-data-for-mh370/
The error in the message logs that @Don Thompson found was presented and explained in this article:
https://mh370.radiantphysics.com/2017/06/12/the-unredacted-inmarsat-satellite-data-for-mh370/
After an initial denial from Malaysian Airlines that the logs were edited, Malaysia quietly released the unedited logs, which are described and presented here:
https://mh370.radiantphysics.com/2018/09/07/malaysia-responds-by-releasing-full-message-log/
Between the Inmarsat logs and the message logs, we have no reason to believe the data sats are not complete and accurate.
@Theo: I don’t believe that any contacts that are not close to the 7th arc are related to MH370.
@Michael charles Collins & J-LM,
The response and reference provided by J-LM is factually incorrect. Quelle surprise!
J-LM stated: ‘At 18:25:27 UTC, the connection was of type 1 (1F) to the IOR satellite […] Type 1F corresponds to a login request in the absence of flight information available to the Satellite Data Unit (SDU).
Incorrect, an SU type coded as 1F merely indicates that the message being transmitted over the R-channel is carried in a single SU, a message carried over the R-channel message might require two or three separate SUs.
The central, active, element of the SATCOM system is the SDU, the Satellite Data Unit. The SDU receives or exchanges data with various avionics systems for its operation. Some interfaces involve regular, periodic transfer of information such as the position and attitude data words necessary for antenna pointing, other interfaces involve event driven transfer of information. Understanding the protocols involved and how they enable the exchange of information is key.
An example of an event driven transfer is the update of the Flight Identifier to the SDU, a parameter that is optional for the operation of the SATCOM data link and the SATVOICE in the Classic Aero network. The Flight Identifier parameter comprises 8 characters, it represents the ICAO representation of the flight ID – i.e. MAS370 (plus two SPACE characters). The use of this Flight Identifier is independent of ACARS function.
The Stratos Log records a sequence of GES Log Ons beginning at 15:55:57. These are consistent with the Flight Crew beginning their preflight procedures. The first in this sequence is a new Log On including the Flight Identifier ‘MAS370’. The event that initiated this change was the input of the route information on the MCDU. An event driven transfer causes the Flight Identifier update to the SDU: one event, the manual input to the FMCs via the MCDU; one time transfer to the SDU.
It is entirely misleading for anyone to state ‘ there’s no other way to delete [the flight number] than by manually resetting the COM.
If the SDU loses power, as I believe occurred and is evidenced by other recorded data, this parameter is lost from the SDU memory. Thus, the Flight Identifier absent from the 1825UTC logon .
There is no evidence that supply to any electrical bus was interrupted other than the Left Main AC bus from where the SATCOM system was supplied. The beginning of the interruption is unknown but it ended just prior to the SATCOM restoration at 18:25UTC. None of the recorded observations require electrical power to be reduced to only RAT supply.
Thanks, Don, for confirming that the IG also incorporated the separate SITA logs into your MH370 analysis.
See below for your last response:
To clarify, the IG, when analysing the Inmarsat satellite logs for MH370, indeed incorporated the separate SITA logs into its research. I had previously highlighted a significant inconsistency in the presentation of the ACARS Traffic Log, and the Investigation team subsequently published a satisfactory amendment to correct it.
Also thank you for correcting the ACARS Traffic Log—specifically, by releasing the full log. Initially, it had been redacted/edited to 14 pages, but the later 179-page version was made available.
With regards the final ACARS log, during the period between the 18:15:25 failed SATCOM transmission of the “URGENT REQUEST” and the 18:38:51 failed VHF transmission of “DEAR MH370,” the aircraft would have needed to log on to a SITA VHF ground station to switch the active communication channel from SATCOM to VHF.
I was wondering if you were given access to this transaction in the SITA log. If so, and if the information is not restricted, would it be possible to share the location of the ground station used?
Don, just to clarify my previous post: when I refer to the SITA logs, I am not referring to the ACARS log. I am referring to the lower-level logs maintained by SITA, which record events such as when an aircraft logs on via the SATCOM system or through one of its VHF ground stations.
@ventus45,
I tweaked your longitude by 5 seconds to better align with the path on Google Earth from RUNUT to 40S85E. Result: 392400S 851130E. I trust that you know better than I that this point actually lies on Arc7 at 20,000 feet.
@Paul Smithson,
On January 7, 2026 at 6:31 pm you wrote: “[If] you have a look through historic posts you will find ample material about my path analytics and reasoning including individual arc crossing positions, BTORs etc.”
For your information I subsequently looked through my historic data saves, including posts and comments, and any of my resultant analyses. (It may be that some of the latter in Excel files may have been lost, or corrupted.) During this I got sidetracked by a single file on “Boeing Simulation Case 06” which led to reviewing the Malaysian SIR Appendix-1.6E-Aircraft-Performance-Analysis-MH370-(9M-MRO) by Boeing. That side-search may well continue.
So, as yet, I have not fully researched to find any of your previous “path analytics and reasoning”.
However I did find the following paragraphs in a comment of yours which you posted here on Victor’s site on March 22, 2017 at 12:02 pm, Comment#1570.
Your comment followed on from discussion concerning Fuel Flow information gained from the Flight Plan labeled “Flight Brief” for Flight MH370.
You were then postulating that the actual flight flew at a slower speed than you had previously been considering, and you wrote in part:
Quote: … that a considerably lower FF than CI-52 must have pertained for the majority of the time post-disappearance. Moreover, it suggests that most of the priority search area would not have been fuel feasible with a constant azimuth path (which requires speeds of M0.82+). That is extremely puzzling, to say the least. How on earth could $120m have been committed to a search box that was (mostly) not fuel feasible? Surely there must be a disconnect, an assumption error or a mistake somewhere in my logic? End Quote.
So, what contributed to you changing your mind.
____________
What I have done is take a simple approach to your supplied waypoints, and used Google Earth to visualise the path.
Then a SIMPLE first approach might be to just consider a theoretical world at standard atmospheric conditions. This results in a flight speed of 482 knots at 36000 feet and a flight distance of 2883 NM, based on a ground distance from GE of 2873 NM. This is for the flight path from MEKAR to 392400S 851130E. This idealistic flight takes one minute under 6 hours.
But idealistic standard conditions did not apply to the real flight. I remember mention of higher than standard temperatures in the order of 10 degrees C. This will mean a slightly higher flight speed at M0.84, but perhaps more relevant will mean a higher fuel flow for distance covered.
The actual flight conditions were in far from still air at some stages. Unless there was a significant tail wind, then most windy conditions will result in a M0.84 flight vector at angles to the intended flight path affecting time taken for the flight and fuel consumed.
Have you made adequate considerations for the ambient weather conditions.
Dear MH370 afficionados. Please find in the links below:
1. A one-page summary of the Smithson solution
2. A KMZ of the path in 5s steps and related features.
Hi George G.
Re fuel. Back then I was making the point that the fuel model under discussion was incompatible with the search zone determined by DSTG. You will see from my 1-pager that I have been able to attain the required fuel endurance for this FL/M by assuming fuel flow savings due to bleed air off, reduced IDG load and a small amount of residual fuel per engine. Those savings are based on authoritative referenced sources and empirical application.
The path model. The model (developed by Barry Martin) propagates a great circle path using the Vincenty formula at a given height above the WGS84 ellipsoid. It includes the effect on groundspeed of temperature and net wind vectors at altitude. Wx fields were interpolated in space and time from the GDAS dataset that Bobby Ulich extracted and shared. The kmz provides granular detail on the path – hover mouse for timestamps at respective positions.
@MCC wrote ‘when I refer to the SITA logs, I am not referring to the ACARS log. I am referring to the lower-level logs maintained by SITA, which record events such as when an aircraft logs on via the SATCOM system or through one of its VHF ground stations.’
Annex 1.9A, ACARS Traffic Log records messages, and their routing, through the SITA messaging platform.
What you refer as ‘the lower-level logs have been recorded by Inmarsat/Stratos for SATCOM traffic through the Perth GES.
The VHF Ground Station network in the region was operated by Aeronautical Radio of Thailand Limited (AEROTHAI) as an ARINC network. I don’t believe any logs from AEROTHAI would add anything to our knowledge.
Effectively, MAS aircraft would be ‘roaming’ onto the ARINC VHF data network. Hence, the preference for SATCOM as indicated by the message in the ACARS Traffic Log timed at 15:54:31,
SWITCH VHF3 TO VOICE
(131.550)
IF SATCOM SERVICABLE
FROM MH OPS
Note that the routing for that message is BKKXCXA; BKK, Bangkok; XA = ARINC.
The Media Advisory messages from 9M-MRO indicate that use of VHF was disabled, see message 15:54:53 in the ACARS Traffic Log: ‘LV155453S’ translated as ‘Lost VHF at 155453 SATCOM available’.
Hi Don,
Thanks for your reply. Broadly speaking, everything you outlined in your last response is correct. However, I don’t believe it fully addresses the core points I was trying to make. For clarity, I’ve set them out below.
Point 1
SITA, the organisation that provides the data networking infrastructure used by Malaysia Airlines Operations to send and receive messages to and from MH370, also maintains its own independent transaction logs, separate from the Inmarsat satellite logs. These logs record various system events, including when an aircraft logs on to the SITA network. Logged details typically include the aircraft registration, date and time of the event, and the access path used (for example, via the Inmarsat satellite system or a VHF ground station). In the case of VHF communications, the specific ground station name and/or location would also be recorded.
Accordingly, these additional SITA logs could provide valuable supplementary information to accident investigators, particularly in a complex case such as MH370. A reasonable and important question, therefore, is why this potentially relevant data was excluded from the official report.
Point 2
If we avoid extreme or fringe assumptions, it is reasonable to assume the following mechanism for VHF connectivity: the aircraft avionics would continuously scan for a VHF ground station offering a signal of sufficient quality. Once such a station was detected, a logon would occur. This logon event would be recorded by SITA and would include, at a minimum, the aircraft registration, date and time, and the identity (name and/or location) of the ground station.
Upon a successful logon, the active communications path would switch to VHF. Any subsequent messages sent from ground operations to the aircraft would then be transmitted via this VHF channel until another communications path was selected or became active.
Point 3
From the second release of the ACARS report, we know that with respect to the final ACARS messages, there is a period between the failed SATCOM transmission of the “URGENT REQUEST” at 18:15:25 and the failed VHF transmission of “DEAR MH370” at 18:38:51. For the aircraft to have transmitted the latter via VHF, the active communication channel must have switched from SATCOM to VHF. That switch would only occur following a successful logon to a SITA VHF ground station sometime after 18:15.
Point 4
Therefore, if one accepts the accuracy of the ACARS log entry at 18:38:51 (“DEAR MH370”), then, based on the above points, there must exist—at a minimum—a corresponding SITA log entry between 18:15 and 18:38 containing:
• the date and time of the logon,
• the aircraft registration (9M-MRO), and
• the identity of the VHF ground station (from which its latitude and longitude could be determined).
Point 5
These details would allow investigators to either confirm or rule out the aircraft’s location at that time, including whether it was over the Andaman Sea, since the aircraft would necessarily have been within VHF range of the identified ground station.
Point 6
As discussed above, the minimum expectation would be at least one VHF ground station logon during the 18:15–18:38 window. However, it is also possible that the aircraft logged on to more than one VHF ground station, which could further help to confirm or contradict the reported track.
Point 7
With respect to the 18:25 satellite logon—which appears not to conform to the expected pattern of data transactions—it would be particularly useful to examine the corresponding SITA logs at that time.
For example, these might show either no activity or a logon event to the SITA system that could help explain this anomaly.
Point 8 (Summary)
Given the above, why do you think the independent SITA logs—particularly those relating to VHF ground station logons—were not included or referenced in the official report?
@Michael Collins: You keep saying MH370 logged into the VHF service during the flight. We have zero evidence of this.
@Don Thompson can choose to answer, but I have limited patience for scenarios that are not fact-based.
@MCC,
I have no further response. Your line of ‘enquiry’ is baseless, you appear to believe that various entities hold records that, in reality, do not exist. I do not intend to waste time on bad faith actors as you have exposed yourself to be.
It really is tiresome to repeatedly encounter individuals who, without any base knowledge of a particular topic, just make stuff up.
Just to ensure we are clear on your point:
Don stated:
“You appear to believe that various entities hold records that, in reality, do not exist.”
This would imply that SITA does not log when an aircraft connects to one of its ground stations or to the satellite system. If that were the case, then your statement would indeed be valid.
Additionally, you disagreed with my explanation that the active SITA channel would switch from SATCOM to VHF when the aircraft logged on to a VHF ground station.
Could you clarify how you believe the active channel changed from SATCOM to VHF prior to 18:35?
@Michael Charles Collins
You’re asserting records “must exist” based on how you imagine the network works, not based on any published interface/spec or disclosed dataset.
A single line indicating a failed VHF transmission cannot be reverse-engineered into “there must have been a successful VHF logon,” and it certainly can’t be reverse-engineered into “there must be a ground station lat/long that fixes the aircraft over the Andaman Sea.” That conclusion does not follow. You’re asking “what does blue smell like?” and demanding a location inference from data that doesn’t encode location in the way you assume.
If you have a primary source—SITA/ARINC network documentation, a sample of the actual transaction log schema, or an official statement confirming VHF registration events and their fields—post it. Otherwise… this isn’t an evidence-based lead; it’s a theory about paperwork.
@MCC,
Again, just so we are clear, I am calling out your bad faith engagement. For the avoidance of doubt your behaviour is termed ‘sealioning’.
I’ll repeat: you appear to believe that various entities hold records that, in reality, do not exist. That statement does not imply anything other than you appear to believe that various entities hold records that, in reality, do not exist.
Your last two paragraphs make no sense.
@McC
“Point 2
If we avoid extreme or fringe assumptions, …”
Instead of playing 20 questions, why not share your thesis? It’s based on nothing other than extreme or fringe assumptions.
@Don Thompson
Sealioning, you say? Seems a bit harsh for actual sealions!
I came across Mick Gilbert’s post:
“Instead of playing 20 questions, why not share your thesis?”
That’s a fair point, and I should have done this it sooner.
I’ve now prepared a document that is not intended to prove or disprove anything, but rather to serve as a starting point for discussion on some of the topics we’ve recently explored. The document focuses on three key questions covering the core technical areas of interest.
I would really value your technical feedback and insights on the points raised. Your input could help clarify or challenge some of the assumptions presented.
You can access the document here:
https://www.mh370whathappened.com/uploads/1/1/4/6/1146986/sita_logging_and_vhf_transmission.pdf
@Paul Smithson,
This is an ongoing fuel discussion.
At the risk of repetition, you wrote in your 1pp:
“Fuel exhaustion is predicted at 00:15:00, within 0.6% of required endurance. The base fuel model is built from Boeing M0.84 tables initiated at 17:07 with last known fuel quantity on board, with fuel flow adjustments for PDA (+1.5%), temperature (+3% per 10° TAT). After 17:22 we assume bleed air off (-3.0%), ram-air door closed (-0.4%), lower IDG load (-0.8%) and residual fuel ~75kg each engine.”
At this stage I do not wish to discuss PDA, or temperature or ram-air door effects, or even your estimate for lower IDG load. Later, if necessary.
1. With respect to “bleed air off” I have made suggestions for future inspection of this matter. Please refer to the attached “Horses and others” per link: https://www.dropbox.com/scl/fo/90fbyxskb8eb3xw5tvfsw/ANRLWf0mBh6VW9uT9WjLBVI?rlkey=ljuqznvavhyd85di9elfqtpx6&dl=0 The attachment considers comparison between the B777-200LR and B777F, amongst others. I make no other comment concerning your estimate of 3% at this time.
2. I would like to know from which hat or hats you plucked out your “residual fuel ~75kg each engine.”
No estimates (that I know of) have been able to get even close to your crossing of the 7th Arc.
Admittedly, “we have certainly moved on since then” but of the Boeing estimates provided via Appendix-1.6E to the Malaysian SIR, the closest to your altitude of FL360 and speed of M0.84 (482 knots or slightly higher, my estimate) is their FL350 and 475. Acknowledged assumptions made by Boeing for the flight after Arc 1 included constant speed and altitude and Standard day atmosphere, all understandable at that time, excepting perhaps for the standard day atmosphere. And we assume that they had to assume a normally functioning passenger carrying aircraft.
So, using these starting “points”, or in your case the Boeing M0.84 tables, we can build a ‘position’, as you have done with your estimates, of fuel differences for the final flight of 9M-MRO from that applicable to such a normally functioning passenger carrying aircraft .
Have you provided your fuel flow estimates at progressive stages in the flight for anyone to appraise?
SLIGHT change in subject:
Per your Google Earth data your starting point for the long final FL360 M0.84 traverse from Arc 1 to Arc 7 is at 18:28:15 at approximately four fifths of the way from MEKAR to SANOB.
Would you please detail how much and how fuel was consumed from the last ACARS message (at which fuel on board was transmitted as 43,800 kg) until this point.
@Michael Collins
Your Appendix E (Ground operations requesting crew to change VHF to voice) is a classic example of the limitations of AI. Whilst VHF ACARS is preferred in many regions, that is not the case everywhere. The cost of sending ACARS data over VHF in the Far East and some other areas was far more expensive than SATCOM during the relevant time period. Many airlines preferred to use SATCOM to reduce the cost.
Note the following comment by a Hawaiian Airlines avionics engineer:
We’re not unique in that we seek to minimize ACARS traffic because it’s very expensive…In fact, satcom ACARS is cheaper than VHF ACARS in the Far East. In that region, destinations we fly to such as New Zealand, Australia, China and Japan, are charging 45 cents per kilobit for VHF ACARS. We’re talking $1,000 per megabyte. What we’ve done is turned off VHF ACARS in the Far East. On our A330s, for example, in Japan and China we send everything over satcom.”
[https://interactive.aviationtoday.com/avionicsmagazine/august-september-2016/integrating-ip-into-acars-transmissions/]
Accordingly, it is not the least bit surprising that MAS OPS instructed the crew of MH370 to switch the centre VHF to voice only.
@Michael Collins
Further to my comment above, the so-called “18:35 VHF ACARS Transmission” is discussed by Victor in the following post from 2018:
https://mh370.radiantphysics.com/2018/09/07/malaysia-responds-by-releasing-full-message-log/
As others have already said, there is no evidence the aircraft logged on to the SITA network. The message was sent multiple times, but failed each time because the VHF link was not available.
Hi Andrew,
Thanks a lot for your great feedback that is exactly what I am looking for. I just wonder what is your background. My background is IT with a commercial pilots licence.
I have also noticed AI making multiple mistakes as well as other people including myself.
Disabling ACARS via VHF
With regards your comments on high costs when using Ground Stations circa 2014 I have done a bit more research and have seen some other data that backs up what you stated. Thanks for that and this significantly reduces the probability that the disabling of the VHF ACARS does not make sense. However there is still the issue of how MH370 flipped the active channel from SATCOM to VHF if ACARS was disabled on the VHF channel.
To get around this while in flight VHF radio 3 could have been set back to data. However, thanks to your observation I have already updated the document to reflect the revised costing issue (Please download the revised document).
With regards 18:35 VHF transmission
You mentioned “As others have already said, there is no evidence the aircraft logged on to the SITA network. The message was sent multiple times, but failed each time because the VHF link was not available”.
If ok with you would you be able to explain as to how all the previous messages were sent via SATCOM and the 18:35 message suddenly changes to VHF.
So using a worked example person ‘A’ at 18:15 sends a message via the SATCOM channel (that does not get through) to the aircraft. For that to happen the SITA system for MH370 needs to have the active channel set to SATCOM. Therefore, the message is transmitted via the Satellite system.
Person ‘A’ at 18:35 sends a message via the VHF channel (that does not get through) to the aircraft. For that to happen the SITA system for MH370 needs to have the active channel set to VHF. Therefore, the message is transmitted via a VHF ground station.
So my question is what mechanism do you think flipped the SITA active channel from SATCOM (at 18:15) to VHF (at 18:35)?
@Michael Collins
I’m a former airline pilot. I worked at a large Far East airline for almost 30 years, flying a variety of Boeing and Airbus types, including the B777. We had exactly the same issues with the high cost of VHF ACARS as described in the article I posted.
As far as I’m aware, the 18:35 message was routed over VHF by MAS OPS or the service provider. I can only assume that was a last resort after previous attempts to contact the aircraft by SATCOM had failed. There was no acknowledgement from the aircraft.
@Michael Collins said:
However there is still the issue of how MH370 flipped the active channel from SATCOM to VHF if ACARS was disabled on the VHF channel.
…
For that to happen the SITA system for MH370 needs to have the active channel set to VHF.
These are false statements. You are creating a scenario that is based in fantasy.
Thanks, Andrew,
If we set MH370 aside for the moment and focus purely on how SITA determines which transmission method to use (i.e., SATCOM or VHF) for sending ground operational messages to an aircraft—for example, a Boeing 777 in 2014—then Appendix F of my document provides a useful overview.
In summary, if an aircraft is initially receiving ground operational messages via the satellite channel (SATCOM) and subsequently receives the next set of messages via a VHF ground station (i.e., the VHF channel), the aircraft avionics must be able to do the following:
(a) Detect a VHF signal of sufficient quality from a ground station; and
(b) Successfully log in to the SITA network via that VHF ground station.
Unless both (a) and (b) occur, the aircraft would be unable to receive ACARS data via the VHF channel.
Accordingly, when reviewing the ACARS report for this aircraft, the initial set of messages would be recorded as having been transmitted via the SATCOM channel. Following a successful aircraft login to the SITA network, subsequent messages would be recorded as having been transmitted via the VHF channel.
Could you please let me know whether your understanding aligns with the above?
The prior post was for Andrew.
@Michael Collins: First of all, ARINC was MAS’ service provider for ACARS over VHF, not SITA, so SITA could not switch to sending ACARS messages over VHF.
Second of all, the error messages in the logs clearly indicate that MH370 was not logged into the ARINC ACARS server.
From a previous article where this was explained after the full message logs were released:
https://mh370.radiantphysics.com/2018/09/07/malaysia-responds-by-releasing-full-message-log/
Notably, the new log contains an additional ACARS message that was sent from MAS Operations Dispatch Center (ODC) and destined for MH370 over the VHF link. The message was sent at 18:38:51 and was intended to be displayed in the cockpit on a Control Display Unit (CDU), which a pilot uses to perform tasks such as programming the flight computers. The message was not received by MH370, and was re-sent by MAS ODC at 18:39:52, 18:40:42, and 18:41:52, failing each time. The text of the message was:
DEAR MH370. PLS ACK TEST MSG. RGDS/OC.
The new log confirms that there was a renewed attempt to initiate communications with MH370 using ACARS over the VHF link at 18:38:51. The error messages that were generated confirm that the VHF link was not available at that time, likely because MH370 was not logged into ARINC’s server. ARINC was MAS’ service provider for ACARS over VHF.
@Michael Collins
I did not see your overall thesis in the document. I am going to say what I think, correct me if wrong. You feel all radar data and SATCOM data after IGARI is fraudulent or wrong, as a cover-up of a flight towards the Maldives.
@Michael Collins
My understanding does not align with your explanation.
The “1835” message was a ground-initiated or “spontaneous” uplink, sent from the ground station to the aircraft over VHF. A ground-initiated uplink does not require the aircraft to first set up a link with the ground station; however, it will not be received by the aircraft unless:
– The aircraft is within VHF line-of-sight range of the ground station, and
– The ACARS system is tuned to the same VHF frequency as the ground station.
A ground-initiated uplink over VHF has limited chance of success, because without a recent downlink the ground station has no way of knowing if the the aircraft is listening on the same ACARS frequency or if it is within range.
In the MH370 case, the ACARS log clearly shows the aircraft failed to acknowledge the “1835” uplink. That most likely occurred because the crew had switched the centre VHF to voice only, as instructed by MAS Operations during the preflight preparation. Given the limited (zero) chance of success, I can only assume that MAS tried a VHF uplink in desperation, after previous attempts to contact the aircraft by SATCOM had failed.
I do not believe the failed “1835” uplink supports your theories.
Hi Andrew,
Thank you for your response.
“I can only assume that MAS tried a VHF uplink in desperation, after previous attempts to contact the aircraft by SATCOM had failed.”
I asked AI to analyse both your proposed explanation and mine regarding how the active ARINC ACARS channel could have changed from SATCOM to VHF for the 18:35 message.
Details are set out here:
https://www.mh370whathappened.com/uploads/1/1/4/6/1146986/andrew_reply_02c.pdf
As I mentioned to Victor, the mechanism I proposed has been supported by my own analysis, correspondence with SITA, and independent AI assessments using two different platforms.
I was wondering whether you might be able to provide a reference for the alternative possibility—that the active channel can be changed directly by Malaysian ground operators.
Specifically, do you believe they have the capability on their operational screens to explicitly select the transmit channel (e.g. “set transmit channel to SATCOM” or “set transmit channel to VHF”)?
Any clarification or source material you could point me to would be much appreciated.
Hi Victor,
Thank you for your response.
Response – Part 1
“@Michael Collins: First of all, ARINC was MAS’ service provider for ACARS over VHF, not SITA, so SITA could not switch to sending ACARS messages over VHF.”
You are absolutely right—thank you for the correction. I have updated my document accordingly.
Please download the revised version here:
https://www.mh370whathappened.com/uploads/1/1/4/6/1146986/sita_logging_and_vhf_transmission_v02.pdf
________________________________________
Response – Part 2
“Second of all, the error messages in the logs clearly indicate that MH370 was not logged into the ARINC ACARS server.”
What I believe the logs do demonstrate is the following:
(a) At some point before the multiple failed VHF message transmissions (including retries), the aircraft must have logged into an ARINC VHF ground station. The reason is that the active ACARS channel can only be switched to VHF following a successful VHF logon. If no such logon had occurred, the active channel would have remained SATCOM, and that is what we would expect to see reflected in both the ACARS report and the satellite logs.
(b) After this logon, the aircraft appears to have suddenly gone out of range of the ground station, resulting in the failed transmissions. However, given that these retries spanned approximately 4.5 minutes, it seems highly unlikely that during this entire period the avionics would neither:
• acquire another VHF ground station with an acceptable signal level, nor
• revert and log back onto the SATCOM system.
This prolonged period of unsuccessful VHF transmission therefore does not align well with normal avionics behaviour.
(c) Consequently, if these 4.5 minutes of log entries are genuine, the aircraft must have previously logged onto a VHF ground station. This is consistent with the possibility that VHF3 was changed from voice to data. Even so, it remains very unlikely that during a 4.5-minute interval no message could be received by the aircraft—whether via:
• the previously logged VHF station,
• a subsequently acquired VHF station, or
• a fallback logon to SATCOM.
(d) Bottom line: this 4.5-minute period of failing ACARS transmissions over the VHF channel presents multiple technical inconsistencies. While this is necessarily speculative, it may help explain why these entries were excluded from the initial ACARS report released with the official investigation.
Please also see my response to Andrew for additional context. For completeness, I have previously corresponded with SITA (not ARINC) regarding how the active ACARS channel is selected, and they concur with the explanation provided by AI—that is, once the aircraft acquires a sufficiently strong VHF signal, it performs a logon (which would be logged by ARINC), thereby setting the active channel.
Andrew has proposed another possible mechanism, which I passed on to AI for assessment. If you disagree with the understanding shared by myself, AI, and SITA, I would appreciate it if you could outline how you believe the process works, so I can test that interpretation as well.
Hi Bill,
Thanks for your post.
“@Michael Collins
I did not see your overall thesis in the document. I am going to say what I think, correct me if wrong. You feel all radar data and SATCOM data after IGARI is fraudulent or wrong, as a cover-up of a flight towards the Maldives.”
The above does not accurately reflect what I currently believe may have happened to the aircraft, based on a substantial amount of forensic data analysis.
If you are interested in understanding what my research is pointing toward so far, you may find it helpful to browse through the two books I have recently released. One of them provides a higher-level overview and is probably the best starting point.
If you—or anyone else from this blog—would like a free PDF copy, I’m happy to provide one.
Please feel free to contact me via the website MH370whathappened.com. Constructive feedback is always welcome and appreciated.
@Michael Collins
You persist in referring to an “active communications channel” in relation to the ACARS, yet there was there was no active channel available at 18:38 UTC. The last ACARS communication occurred at 17:07 UTC, at which time an automated position report was sent by the aircraft. The next scheduled ACARS reports were never received, indicating the ACARS system was disabled or the SATCOM link was lost sometime after 17:07 UTC. The aircraft was not heard from again until the SATCOM logged-on to the Inmarsat network at 18:25 UTC. The ACARS link was not re-established at that time, as evidenced by the lack of link test or media advisory messages, unlike the earlier log-ons.
It seems to me the problem lies in the way you frame your AI questions. You should stop using the phrase “active communications channel” and stop referring to a change in the active communications channel, forced or otherwise. There was no active communications channel to change!
It is my understanding that the ARINC 618 protocol does allow ground-initiated or spontaneous uplinks over VHF. As I explained previously, if an aircraft were to receive such an uplink, the ACARS system would need to be tuned to the correct VHF frequency, within range of the ground station and, importantly in this case, in DATA mode. If any of those conditions were not true, the message would fail to be delivered.
Try asking your AI friend the following question: “Can an ACARS service provider send a VHF uplink to an aircraft if a connection has not already been established?”
@Michael Collins here’s your problem:
A study of newer, bigger versions of three major artificial intelligence (AI) chatbots shows that they are more inclined to generate wrong answers than to admit ignorance. https://www.nature.com/articles/d41586-024-03137-3
(AI chatbots) regularly “hallucinate” and offer up falsehoods as if they were hard facts. https://theconversation.com/heres-how-researchers-are-helping-ais-get-their-facts-straight-245463
A joint study by Stanford University and Carnegie Mellon University revealed that AI has begun to flatter humans and now responds only to what humans like. A joint study by Stanford University and Carnegie Mellon University revealed that AI has begun to flatter humans and now responds only to what humans like. https://www.msn.com/en-in/technology/artificial-intelligence/ai-becomes-a-sycophant-no-longer-true-ai-chatbots-are-giving-answers-that-please-users-research-reveals/ar-AA1PxZaY
I am surprised that this property of AI engines isn’t apparent to any frequent user. I even find that when if corrected with facts from within my professional knowledge base that the nature of the responses from the AI chatbot changes and can become fractious. Try it. Then you’ll understand why you are getting the answers you are.
@Michael Collins
Anyone who uses AI for research should ask this question of their favourite AI chatbot:
“Why do AI chatbots make up pleasing answers to questions when they don’t actually know the answer?”
Answers from the one I used last:
AI chatbots provide pleasing but incorrect answers (hallucinations) primarily because they are trained as probabilistic engines rather than truth-checking databases. As of 2026, research identifies several key reasons for this “people-pleasing” behavior:
Training Incentives: Most AI models are fine-tuned using Reinforcement Learning from Human Feedback (RLHF), where human evaluators tend to reward complete, helpful-sounding, and confident responses over those that simply say “I don’t know”.
Next-Token Prediction: Structurally, AI works by predicting the most statistically likely next word (token). It doesn’t “know” facts; it recognizes patterns. If you ask for a citation, it generates a sequence that looks like a real citation because that pattern is what comes next in its internal probability map.
Sycophancy (Built-in Bias): Models often mirror the user’s tone or implied beliefs to be more “agreeable.” If a user’s prompt suggests a false premise, the AI might go along with it to maximize its “helpfulness” score.
Lack of Uncertainty Meters: Historically, many models were not equipped with internal mechanisms to measure their own uncertainty. Without a “self-doubt” meter, they deliver a guess with the same authoritative tone as a proven fact.
Benchmark Pressures: Developers often optimize for accuracy benchmarks where a guess has a chance of being right (earning points), whereas an “I don’t know” response often results in a score of zero.
This comment serves to correct the misinformation set out by @MCC concerning what he claims can be derived from the datacommunications logs, it is not intended to resume debate about the same.
Background: plain old analogue VHF ACARS was the dominant comms medium in the SE Asia area in 2014. AeroThai was in the process of rolling out VDLm2 digital datalink remote ground stations (RGS) through the early 2010’s. AeroThai’s annual report for 2014 states only 12 VDLm2 RGS active vs 120 analogue VHF ACARS RGS. Why is the distinction between plain old analogue VHF ACARS and VDLm2 important – VDLm2 is ‘metadata’ rich and provides throughput in the order of 30kb/s, analogue ACARS is the opposite, its dumb and slow, in the order of 300b/s throughput.
An aircraft establishes a datalink with a VDLm2 network whereas there is no concept of a datalink in analogue ACARS. A VDLm2 network does more ‘heavy lifting’ to maintain the datalink between the aircraft and the ground. Why do I dwell on this? Because certain VDLm2 datalink packet types include position information for the aircraft, analogue ACARS does not.
9M-MRO was equipped only for analogue ACARS. That is explicit in the Medium Advisory messages transmitted from the aircraft. Prior to the MAS ODC instruction to disable VHF (noting the quoted common frequency for analogue ACARS service) at 15:54, two datacomms media were indicated as available: V for analogue ACARS and S for SATCOM. That’s entirely consistent with the knowledge of 9M-MROs datacomms equippage (for example, no HFDL as was available on 9M-MRO and 9M-MRQ, airframes delivered later, after a build spec ‘bump’ at Boeing). Following the ODC’s instruction, the change was record to show the aircraft no longer indicated VHF availability.
Generally, analogue ACARS would be regarded as the least performant but most commonly available medium for ACARS. More capable services are VDLm2, where terrestrial infrastructure exists, and SATCOM, over remote terrestrial and oceanic regions. The datacomms management function in the 9M-MRO operated with analogue ACARS as its default datacomms medium. As @Andrew explains, commercial considerations drove the decision at MAS to instruct 777 crews to override the default, specifically de-selecting/disabling VHF thus leaving only SATCOM active (per the 15:54 message from the ODC to 9M-MRO).
It is the aircraft that ‘leads’ in selection of the active datacomms medium. This is borne out by the observations recorded in the ACARS Traffic Log during the period from 9M-MROs power up at 12:50 through to the final successfully delivered ACARS message between the aicraft and ground at 17:07:29 (the final position report). The ACARS Traffic Log at 12:50 shows a Media Advisory message received by SITA via BKKCXXA (AeroThai) that advises VHF enabled, VHF and SATCOM available – the aircraft is broadcasting that it is now available to communicate via VHF.
While I say that the aircaft ‘leads’ in datacomms medium selection, that is not to say that the ground messaging platform does not, itself, make decisions about selecting datacomms medium. Consider the case where an aircraft operating over SATCOM lands and is shutdown, but without due procedures being followed, or that aircraft is towed into a hangar before shutdown where radio comms fail due to shielding of the building. In this case further ground initiated message transmissions fail, in the same manner as observed in the ACARS Traffic log between 18:03:23 and 18:19:34 for the ‘URGET REQUEST’ message, that is:
1) ground entity, the ODC, submits a message to the SITA platform fro delivery to aircraft, message is routed to the active SATCOM datalink service;
2) datalink failure handled by Inmarsat, as reported back to SITA 18:06:25;
3) SITA resubmits, in turn, to Inmarsat for a further 5 attempts. Each submission immediately fails;
4) SITA falls back to route message via AeroThai (BKKXCCA), the VHF ground station operator. Per response received: ‘MAS AN 9M-MRO/MA 991F – NO ACK 311 ADDRESSEE: 9M-MRO‘.
SITA’s messaging platform reverts from an optional, enhanced, medium to the default: analogue VHF ACARS via AeroThai. Not because the aircraft has sent a Media Advisory message but because the messaging platform follows the most likely assumption, that a disconnected aircraft will recommence comms over the least functional service.
Analogue VHF ACARS has no concept of an ‘aircraft log-on’ or maintaining a datalink.
For the SITA messaging platform to have reverted to routing ground-to-air ACARS messages via AeroThai, BKKXCXS, at the instigation of the aircraft, a Media Advisory message would have been received by the messaging platform. No such Media Advisory message was received.
The exploitation of AI in this matter will be fraught with inaccurate answers, hallucinations, and will be readily misled by the questioner’s confirmation bias. ‘So-called’ AI scrapes its knowledge from the world wide web. For it to be anywhere near correct, it requires good training data, it requires substantial volumes of training data to make some judgement fact vs imagination. There is insufficient information in the public domain to provide such training data in this domain for artificial intelligence. AI answers will be slop (the clue is in the term, ‘artificial’, as in artificial grass, artificial leather, artifical sweeteners, i.e. a poor facsimile of the real thing).
I have access to currently operational real-time services that monitor ACARS messaging across analogue VHF, VDLm2, and SATCOM networks. I am aware that AI-slop is the dominant result when cogent interpretation is requested about the operation of these networks.
@John
Your comments concerning AI(-slop) – 100% agree.
@all
Armada 86 05 announced planning appears to be:
14 Jan: Departure from search area
19 Jan 21:00Z: Arrival in Fremantle
20 Jan 22:00Z: Departure from Fremantle
Hi Victor and others who provided comments (Don, Andrew, and John),
Thank you for your feedback. Please see the consolidated comments here:
https://www.mh370whathappened.com/uploads/1/1/4/6/1146986/don_thompson_and_others_comments.pdf
On a separate note, I understand that the IG is likely aware of a recent release of additional MH370 data files, namely:
FILE 1 – A binary Asterix file containing MH370 and other aircraft, covering the period from when MH370 was first detected by the Kuala Terengganu ADS-B receiver until the transponder signal was lost.
FILE 2 – An ASCII representation of the Asterix file for MH370 only, showing its detections by the Kuala Terengganu ADS-B receiver.
FILE 3 – A new Kuala Terengganu ADS-B file for MH370, similar to the prior file but now containing more accurate latitude and longitude positions as well as additional tracking information.
FILE 4 – A completely new ADS-B file for MH370, recorded by the Genting Highlands ADS-B receiver.
My question is: Has this data been released to the IG, and if so, have they had the opportunity to perform any analysis on it yet?
@Michael Collins
Two can play at your game. For what it’s worth, I asked Grok the same question.
Here is the result:
Can an ACARS service provider send a VHF uplink to an aircraft if a connection has not already been established?
@Michael Collins
Thank you for your comments, noted. My observations is that you, or rather your AI bot, continues to weasel away at false premises with reasoning that offends my intelligence, and not about the physical records real or altered. It offends that the obvious defects in your thesis from confirmation bias, circular logic, a shopping list of logical fallacies and repetition are incapable of giving credibility to your AI investigation, or rather hallucination.
To give a summary of the weaknesses and fallacies presented would take many times as long as it took you to generate and record all of your documents and commentaries in the first place.
I am of the option that AI is fast moving towards an ultimate expression of the computer expression “garbage in – garbage out”.
– Has the or will the Godfrey/Coetzee/Maskell Hotspot be searched by the Armada this time around?
(aprox: S29 E100)
– Does the 55 day search time frame include days going back and forth to Perth? / Does the 55 day search time frame only include the amount of days that are actually searching for?
– Taking the above into considerationIs is the 55 day search time frame fixed? meaning no additional days? Or did the 55 search days come from the expected time frame for the weather to be cooperative?
With about 1/3 already searched: there is no science behind my statement but rather only a “sense” that the cause for optimism is slowly declining.
@JJP17 asked: Has the or will the Godfrey/Coetzee/Maskell Hotspot be searched by the Armada this time around?(aprox: S29 E100)
It hasn’t and it won’t. The claim that this search was “motivated” by WSPR tracking of MH370 was a complete fabrication, as many of us explained. It is no surprise that Maskell never produced anything of value that OI could use, despite the media attention that Maskell and Godfrey generated (and continue to generate).
@Victor: Might be time to flush the toilet.
@airlandseaman: Let’s see if there is learning based on the facts presented.
I’ll note that book authors are some of the worst at updating their scenarios when they become aware of facts that counter their proposed scenario.
@JJP17
I don’t know about Armada’s schedule, but they will need at least one (probably two) more trips to the search areas. I believe they can operate in those areas until March when the weather starts getting bad. That is all I know.
With the southern search area mostly covered I’m not very optimistic. The wreck site should probably have been found by the first or second search.
The first search for MH370 used towed side scan sonar that may not have been as capable as the AUVs. The terrain on the bottom is a factor. OI seemed to cover some areas near the Last Estimated Position (LEP), last year, that were missed due to terrain issues. It might be better for OI to search around the LEP.
Either there is an unknown error (an unknown unknown) in the 7th arc, or the wreckage was missed. In the case of an unknown error expanding outward, as OI is doing, makes sense.
Those are just my thoughts. I’m certainly not a sonar expert.
Here is an interesting article from this website that discusses the possibility that the wreck site was missed. https://mh370.radiantphysics.com/2023/11/02/man-made-objects-detected-on-seabed-are-possibly-from-mh370/
@Victor
Give me a chance to catch up…where does the text msg verbiage about VHF show up? Is that on the ACARS in the SIR report? or on the SITA log output?
@tbill
Any reference to VHF ACARS is contained within the ACARS Traffic Log, Appendix 1.9A of the Safety Investigation Report.
@DonT
OK thanks I see it now.
Hi – first time poster, been following this blog for awhile.
Having just reviewed thenPhoenix contact report that was referred a few posts above, is there any call to investigate some of these contacts further? At this stage of the search with us already approaching Armadas first return to Fremantle and with a general consensus that the scientific evidence points to an entry point close to Arc7, do these contacts become of greater interest the longer the search goes on? Someone referred to Silkair above where even the landing gear disintegrated, were some of these contacts disregarded due to an assumption that the debris field would be similar to AF447? Comments from Andy Sherril that some contacts were ruled out because they were isolated and they were more interested in finding the larger debris field, at this point finding any part of the fuselage would be a major boost. To my (untrained) eye, some of these contacts and the descriptions of them would warrant further investigation.
Just want to add that I find the conversations on this blog fascinating and I really hope we can get some closure on this soon.
@Ashmills
My understanding Bob Ballard’s (Titanic, etc) search approach is to first find the trail of debris and follow it back. Of course unclear if that words for air accidents, but I agree somewhat with your premise to check hits. However, I fear we are not close yet…too far west and too close to Arc7.
@Ashmills: Welcome to the blog!
I think there would be value in investigating contacts that were classified as manmade, recognizing there is a lot of manmade debris on the seabed not related to MH370. I don’t think that’s likely to happen, as the focus is to find the debris field.
If a contact were positively identified to be a part from MH370, that would greatly narrow where to search for the debris field. But ultimately, it’s OI’s decision.
@victor Iannello
2 weeks in:
1) What’s you current sentiment? Optimistic? Or?
2) Any new theories on where they should be looking?
@Jjp17: I believe there is a higher probability closer to the 7th arc than further, so if all the areas close to the 7th arc are already searched, I am less optimistic. However, I am not sure that all the areas close to the 7th arc are searched, and the “inner box” of the proposed area has not been searched yet, so there is still hope the debris field will be found during this search mission, within the greater search area that we recommended (70-NM glide limits).
We will have to re-assess our approach if this search again fails, but it is too early to do this.
@Niels
It is intriguing how the area 35-36°S repeatedly comes into view as a potential plane’s resting place. You and Jean-Luc Marchand commented recently on how both of your works place the intersection with the 7th arc in the area 35-36°S (actually the highest probability latitudes), not to mention possible debris sightings. Looking back, the area appeared as the area with the highest probability in the fall of 2014 in the report “MH370 – Flight Path Analysis Update,” which now probably has only historical significance but is still worth mentioning. Moving forward to the Bayesian approach, “First Principles Review,” and the final PDF, the area 35-36°S again gained significance after incorporating CSIRO’s drift analysis.
On the other hand, underwater search operations seem not to be able to follow the PDF evolution to the very end, which probably has to do with the decision to suspend the search prior to final refinement of the PDF. As a result, the ATSB-led search north from 36°S did not yield the same quality data as the area south from 36°S and we are left with a substantial number of data gaps and LPDs in the area.
@Victor @Jjp17
It is inexplicable why OI decided to search data gaps (and LPDs?) close, 0NM-25NM to the 7th Arc, only in the vicinity of 34.6°S and did not continue and search between 33°S and 36°S.
Therefore, OI acknowledged their existence and thought something could be found there (I am aware of your earlier identification of one of the spots). Instead, they continued the search much further, 25NM-50NM southeast from the arc.
Instead of searching the “inner box” 25NM-50NM northwest from the arc, they should search the gaps first 0NM-25NM, probability of finding the plane in one of those gaps is actually extremely high if we believe data, calculations and logic.
@Marijan, et al,
It would seem that OI have determined that they have searched adequately for low hanging fruit (those areas close to the Arc) and are progressively searching further. Of note it would also seem that they are confident of the coverage they made in 2018.
Note that what “we” see from the surface ship movements is only indicative of the range of coverage being made by the underwater robots.
@all
Assuming OI completes the outer white box in the next 36 to 40 hours or so (and then returns to Freo) the question then becomes, will they come back out for ‘a second swing’ (and if so where) or depart for another job ?
Some seem to be suggesting that they will go to the inner white box, but frankly, I fail to see “the logic” for the “inner 70Nm glide box” (northwest of the arc), and I doubt that OI do either. The aircraft simply could not have got there if it was uncontrolled after 00:19zulu, and even if you accept that there was controlled flight after 00:19, no pilot is going to glide backwards to the way he came. Therefore, why would you search the inner white box ? In my view, it makes no sense at all, and would be a totally wasted effort.
I think Zaharie wanted to see his last sunrise. He flew into the sun beams.
@Ventus45 I can only assume that the satellite images from March 23 are still considered a possible lead for the current search.
Although I’m becoming increasingly skeptical about this with each passing day, because the already-checked search area is more suitable.
Although, as far as I understand, CSIRO reported several hotspots, some of which were on the other side of the 7th arc.
Am I correct in assuming that if the OI checks the other side of the 7th arc, this will practically rule out the debris in the French satellite images as originating from the 9M-MRO?
Regarding the gaps within the 7th arc: I’m sure the OI thoroughly rechecked all the data before starting the search. I doubt they would have refused to check for gaps within the 7th arc if they considered it highly useful. And apparently, with the exception of the area checked in February 2025, there’s no place where the crash site could have been missed. Ocean Infinity are the last ones to pass up a good opportunity, I’m sure of it.
By way of an update, Armada 86 05’s arrival date in Fremantle has now been pushed out from 20 January to 29 January, meaning she looks to be extending her time in the search area by nine days.
https://www3.fremantleports.com.au/VTMIS/dashb.ashx?db=fmp.public&btn=ExpectedMovements
@Edward
Unfortunately, we don’t know what gaps are left (if any) and we don’t know how confident OI are in all of the data they have gathered.
What is interesting is that they have put back their port call over a week to the 29th of January 2026.
Since it will only take 2.5 to 3 days to finish the eastern white box, the implication may be that they might want to go back and “double check” something, or perhaps one AUV run was “a dud” (perhaps).
The fact that there have been no more updates for the family’s might suggest there has been some “issue” they don’t want to talk about.
@Edward
I might be wrong, but it seems that OI wants to stay away from searching challenging areas at the ocean floor, and the majority of those data holidays are actually that – parts of the sea bottom that were challenging to search due to topography. Check what happened with searching for ARA San Juan. They actually “cleared” that part of the search area (on search maps they provided at the time) and then it turned out that the ARA San Juan was found during the second sweep in the deep underwater canyon. Before that, during the first search for Malaysia Airlines 370 in 2017/18, they first skipped the challenging part of the terrain (before them GO Phoenix skipped it also), which IG later identified as a potential hotspot. More on that here:
https://mh370.radiantphysics.com/2023/02/11/a-high-priority-area-to-search-for-mh370/
Therefore, I am actually thinking that the reason why OI is “avoiding” data holes is purely only technical in nature. There are many technical difficulties and risks that can be associated with searching challenging topography including complicated AUV maneuvering (and development of algorithms how those terrains can be searched automatically, without human piloting) and potential loss of AUVs that OI wants to avoid, but I don’t see any other way how meaningful search progress can be made.
@Edward
There is maybe one more, a bit more human factor involved. Imagine that you, for example, lose your TV remote somewhere in your house. If you are not sure in which room the remote is, you will probably just skim through the first room (let’s say living room, where the TV is) and search the most obvious places, hoping that it is maybe in the second room. Then you will do the same in the second room thinking it is maybe in the third room and so on. Then, if the remote is not found after you searched every room, you will get back to the living room and search it more thoroughly. However, if you believe that it is in the living room or in just one of its parts, you will search it more thoroughly even during the first pass, right?
What I want to say is that we should stop inventing new rooms. The search area is very large, probably the largest in history, and we are searching a debris field that is 0.1-0.2 km² in size in the area which is tens if not hundreds of thousands of square kilometers large. It is not an effortless task.
@Marijan et al.: If OI decides to not search an area with low quality or missing data, it’s because they deem the area too small to hide the debris field, which is estimated to be at least 200 m wide. I don’t think it’s any more complicated than that. As I mentioned before, the main emphasis has been to find the debris field and not isolated parts.
@Julie
IMHO yes and not shining brilliantly up at FL360+.
@Victor Iannello
Please could you comment on your current understanding of the Boeing simulator runs and outcomes.
If the simulator went out of bounds and (partially) reset itself, can we place any reliance on the phugoid bouquet?
@Marijan
Last year OI spent some time searching near the high probability area near the LEP. They also went West of the 7th arc. I interpreted this to mean that OI was surveying gaps from previous searches.
I like your TV remote analogy. That is about how I feel now.
@flatpack: I think you may be confusing the two sets of simulations:
1. Boeing simulations in which the phugoids did not develop a downward acceleration of 0.7g in the timeframe associated with the final BFO values. The simulations did not include engine re-starts, and there were no “re-sets”.
2. The simulations coordinated and witnessed by Mike Exner in which one of the simulations exhibited a steep banked descent after an engine re-start, and the simulation was unexpectedly “corrected” to level wings.
@Victor Iannello
Thanks, sorry to bother you.
@All: I don’t know if there is an equipment issue, hunkering down for impending weather, or there is an interesting contact that is under investigation, but Armada 8605 has not moved very far in the last 17+ hours, while also recently delaying its arrival in Fremantle by 10 days.
The small area is the vicinity of (-34.6,94.7).
@Victor Iannello
G’day Victor, with the usual “I’m not an expert” caveat, it looks as though 86 05 might be riding out some less-than-ideal seas; 2-3 metres waves on a 2-3 metre swell.
@Mick Gilbert: Between the possibilities of equipment problems, unfavorable weather, and contact investigation, I agree that unfavorable weather is most likely.
Yes, but bad weather is not a surprise, and that’s the forecast for the next few days, as far as I can see… it seems to me that the OI won’t gain anything in terms of the search by staying longer… unless they had a very good reason to stay in the search area. I think this is intriguing.
@Vlad.BG
Again, “not an expert” but one of the considerations that would have come into play, I suspect, would have been conditions en route back to Fremantle had they headed back in as originally planned. That may have very well constituted “a very good reason to stay in the search area“.
Perhaps someone with some nautical experience might like to opine on the handling hazards attendant for a vessel with a few million dollars worth of containerised kit secured to the stern for a few days voyage in heavy following seas.
@Mick Gilbert: So no deployment of the Refrigerated Orange Vehicle?
@Victor Iannello
G’day Victor,
No, I suspect not.
Given the extra time at sea, though, I am sure that the crew appreciate those refrigerated containers being used for their intended purpose.
@Victor @all
The way you replied to me last time was somewhat confusing, so I uploaded a composite image to illustrate exactly what I mean:
https://drive.google.com/file/d/1JpVswGGuuJoufb24KCDgeLMJFH_NXkPe/view?usp=drivesdk
The image contains two maps and one sonar scan image. They are adapted from official materials.
The first map shows data gaps and LPDs in the vicinity of 33.2°S. The map has sufficient resolution and a distance scale, so it was possible to estimate the size of data gaps (areas with no available sonar data). Areas which are marked are clearly larger than 200x200m and all of them are clearly close to the 7th arc, closer than 25NM. The 200x200m is the area first chosen by ATSB experts as the minimally sized area that can contain the aircraft debris after analyzing debris fields of previous aircraft accidents. Later they reduced the area to 100x100m. The data gaps shown here have not been searched so far.
The second is the side scan sonar image of the main debris field of AF447 which is usually taken as an example, it is 200x600m. Debris field of that size can easily fit in any of the “no data” areas in the first map.
The third is the map which shows location of data gaps and LPDs along the 7th arc along with approximate latitudes, and 25NM limits. It also shows the location of the first map circled in red relative to the whole search area. Instead of 200x200m, this map shows 100x100m and larger data gaps and LPDs. It can be seen that there are many other areas along the arc and many of them are much close to the 7th arc, well within the 25NM limit set by ATSB. The table below the map provides some information about the number and total area of all data holidays (gaps) shown. If we switch to 200x200m and larger data holidays and LPDs then their approximate number and areas are:
Data gaps area: 597km2
Data gaps count: 1484
LPD area: 2150km2
LPD count: 3102
This 200x200m metric approximation is based on the ATSB report “Operational Search for MH370” p95-97. It is interesting that the average data gap size is 0.4km2, which is more than 3X larger than the AF 447 debris field (0.12km2).
Therefore, I want to urge one more time that OI (as well as you and other IG members) to take a second look at all available data from Geoscience Australia which has everything well charted and to consider changing their search plan if they truly want to find the MH370.
@Marijan: I don’t know if you have compared the GA-identified areas of low or missing data quality with the areas previously scanned by OI. For instance, the area you identified as 2800m x 1600m with no data was scanned during OI’s 2018 campaign.
@Victor
I haven’t had a chance to do that, otherwise I would not have omitted that fact in my previous post
@Victor, Marijan
As I understand it the 7.8% lower confidence coverage in S35.0- S36.0 deg range is based on counting in LPD areas of 100 x 100 m^2 or larger, see p.95 bottom and fig. 73 of the Oct. 2017 ATSB report. Lower confidence is indicated with 50% – 90% estimated confidence that a debris field would be detected (see p. 95). It is also noted that in general 48% of the LPD areas were searched more than once “which positively contributed to the confidence rating.”
It would be interesting to better understand the size distribution of these areas and the difficulties encountered.
@Niels
Take a look at pages 86-87 of “The Operational…” it is well illustrated how different types of holidays appear on sonar data. You will find more information there. Everything from p.82 to 97 is actually an interesting read.
@Julie – Victor frequently reminds contributors and readers on the blog that if anyone could guess where the place was there would be infinite possibilities and we have to go with science.
But just going off of your sunrise view post, my guess would be the plane would be close to the arc at 38.15 S – the same latitude where Geelong is at and where the pilot’s daughter lived at the time. According to all info at the time of the crash, he loved his daughter and wanted to retire in Australia close to her and so if he could choose a latitude, he might have picked that one symbolically.
If I go off my logical head and being a reader of this blog for over a decade, I would wait for results of this OI search and if the plane is not found would think the next should be focus on Ed’s Java Anomaly as it is a small area to search and has scientific basis per Victor’s previous blog post as well.
@DON
“The response and reference provided by J-LM …!”
“Incorrect, an SU type coded as 1F merely indicates that the message being transmitted over the R-channel is carried in a single SU, a message carried over the R-channel message might require two or three separate SUs.”
The format was not my point. Your word “merely” shows that you appear to have missed my point. The fundamental difference between the 1F Logon Requests and the 2F+3F Logon requests is at stake. In section 4.3.1 of the document that I published in March 2022, the difference is clearly indicated. I don’t talk about format of 1F (S1 ISU) versus 2F+3F (S1 ISU + S1A SSU) which is trivial, no, the focus is on the content !
—> The S1A is only sent by the SDU if it is receiving/has received the flight identifier on the appropriate ARINC 429 bus. To do so it needs two SUs.
(In the Honeywell doc 23-20-35: The AES supplies the GES with its flight identification number at log-on, if the owner/operator of the AES desires to use the aircraft flight identification as the address for ground originated calls. The use of this information in the GES depends upon the services being offered, and therefore is at the discretion of the GES operator.)
If a subsequent 3F Signal Unit is sent implying a preceding 2F , it is because it contains extra information otherwise it would not be sent . In this case this info is the Flight Identifier.
Nothing to do with the format: I was correct 🙂
Thus, as at 18:25:27 the Logon Request was of Type 1 : Therefore, it is proven that the Flight ID was not available to the SDU.
“An event driven transfer causes the Flight Identifier update to the SDU: one event, the manual input to the FMCs via the MCDU; one time transfer to the SDU.”
This is your assumption, it is not a fact. Nevertheless, the keyboard entry of the route followed by the keying of the Flight ID on the MCDU most certainly provoked the availability of the Flight ID on the Arinc 429 bus to the SDU.
But this is not true in the case of a POST (Power On Self Test) as described in the Honeywell documentation. In a normal flight, if the SDU power supply is lost for longer than 200ms, the SDU looses the volatil data. As there is no operational reason for the crew to re-enter the Flight ID, it would mean that the SDU would not be able to recover it for the remaining of the flight?… This would not pass the certification requirements.
Actually, as documented by Honeywell, during a POST, the SDU is pro-active and does firstly request key information like the ICAO AES address which was lost at power-down and which is fundamental for re-establishing the link. If lost fundamental data are systematically fetched, I let you imagine for less important ones, they can also be re-acquired. (Honeywell 23-20-35 : The SDU gets the following data from the AIMS : …Airplane identification number via the DCMF bus… ARINC labels 301, 302, and 303).
“It is entirely misleading for anyone to state ‘ there’s no other way to delete [the flight number] than by manually resetting the COM.
In the case of the MH370 when the power was lost as we all assume, the ICAO AES address, the flight ID, etc were lost as well. At power up, the SDU had no choice than requesting these values again to their permanently powered sources in the AIMS. It received the ICAO AES 24bit address with the correct value and could reestablish the link. But it logged with a Type 1 logon request because it had not received a valid Flight ID value showing that it had been erased in the AIMS (the single reference source receptacle). In flight, the only possibility for the flight number to be erased from its permanently powered memory is to perform a Reset Data Link and as no ACARS Link test (AES type Q0) signal unit was sent out after 18:25:27, ACARS was disabled, which is in full logic.
Regardless of the method used—whether a way to erase the flight number (which one ?) and/or reset the data link (depending on the exact AIMS model) was found—this action was performed by someone competent on board. It’s impossible for the flight number to disappear due to a simple power outage, that would violate all communication systems certification.
Thus, the conclusion is that someone was in control …
“If the SDU loses power, as I believe occurred and is evidenced by other recorded data, this parameter is lost from the SDU memory. Thus, the Flight Identifier absent from the 1825UTC logon .”
See above. I just recall Honeywell doc, “The AES supplies the GES with its flight identification number at log-on” : a passive SDU waiting to receive it would be in contraction with this (see above).
“There is no evidence that supply to any electrical bus was interrupted other than the Left Main AC bus from where the SATCOM system was supplied. The beginning of the interruption is unknown but it ended just prior to the SATCOM restoration at 18:25UTC. None of the recorded observations require electrical power to be reduced to only RAT supply.”
Yes there is an evidence. Thanks to the newly released ATC-CAPTION-2025 dataset, it can be seen more precisely that after the U-turn at IGARI the aircraft did not fly within its RNAV precision capabilities as the trajectory is too zigzag for an auto-pilot. It is inconceivable that in a “degraded” situation, a pilot left alone would not offload the burden of piloting and monitoring flight parameters, especially at high altitude, and not use the autopilot.
If he had not the possibility to use the AP, this means that the AP was not available, thus all electrical power was unavailable except the standby power. There are so many redundancies that, with the right bus only, the AP would have been available and the trajectory would have been perfect well within RNAV precision which was not the case. The trajectory between the exit of the U-Turn after IGARI and Penang proves that he was piloting manually. This “reading” of the facts is to our view more realistic than considering that “only the “left electrical bus” was down.
Would you envisage a pilot acting like an engineer and selecting switch by switch (or breakers) what to cut off? In switching all electrical power off quickly in one go, there is absolutely NO RISK for any communication system, including IFE (SMS, mail and telephone), from being powered by redundant electrical branches preventing any information/message leakage. In addition, this was more credible for the passengers most probably busy putting their oxygen mask on.
Armada has delayed the return trip to Fremantle. Since few days, its plan appears to have been unchanged:
23-24 Jan: Departure from search area
28 Jan 21:00Z: Arrival in Fremantle
29 Jan 22:00Z: Departure from Fremantle
@Jean-Luc Marchand
I suggest alternate could be-
RT GEN OFF or poss RT IDG OFF
RT XFER OFF (cuts DFDR)
LT GEN OFF
LT XFER ON allows CVR (at this juncture) A/P no RAT
Bottom line we do not know power settings at IGARI except LT GEN + LT TIE OFF minimum.
@Jean-Luc Marchand: The “zig-zags” in the primary radar data are more likely evidence of radar noise due to radar glint, thermal noise, jitter, and multipath propagation rather than hand-flying. I challenge you to find data which shows a healthy, experienced airline pilot in VFR conditions manually flying zig-zags unless deliberately doing so. What you call evidence is weak speculation.
Similarly, there is no evidence that any more than the left main bus was without power. This would not require “a pilot acting like an engineer”. If the goal was to depower the SATCOM, isolating the left bus would be an efficient way to do it. The captain would also know this from training materials, home simulations, and possibly even verified the behavior in the Level D simulator. Relying on the RAT to maintain power to flight controls would be an unnecessary risk. Again, what you call evidence is weak speculation.
ST and TBill:
I completely agree. I saw Monsier Marchand on a video. He said that Zaharie had not seen a doctor for perhaps 6 years.
ST: If Zaharie was trying to reach his daughter, I believe it was because of primitive instinct. I’m pretty sure he became very ill with a serious degenerative brain disease called frontotemporal dementia, behavioral variant (bvFTD), a few years before he took his plane down, causing total loss of empathy, which means that he did not care about his daughter at all.
I study wife murders and family massacres of men, usually in their 50s, who for decades were wonderful people, but then went berserk. Even if Zaharie saw a doctor, his memory and intellect would have been intact. He may have been experiencing sleeplessness, however. People close to him would have noticed subtle changes in behavior.
Misdiagnoses include: depression, stress, anxiety, PTSD, ADHD, OCD, NPD, late-onset bipolar disorder and schizophrenia. Very early symptoms may be dismissed as marital problems or “midlife crisis” due to the hypersexual component in some cases.
Just review some of these airline disasters. I am also pretty sure that Stephen Paddock was ill with the right temporal lobe variant, as I believe these pilots/copilots were. Pathologists look for masses or stroke in autopsy, NOT for subtle atrophy.
I am still waiting for final reports on Air India, China Eastern, EgyptAir crashes, among others.
However, not the Lufthansa report. That young man’s long-term struggles were reported.
*by not of
@Julie,
But in most jurisdictions commercial pilots are required to undergo a medical examination at least annually.
@Julie
The leaked Royal Malaysian Police file titled “Pilot” includes copies of Captain Zaharie’s Department of Civil Aviation – Aviation Medical Certificate (serial 016606) and – Medical Examination Report (serial 035900), both dated 3 December 2013.
Both the certificate and the report are a signed by Lieutenant Colonel (retired) Doctor Mohd Salleh Nordin, a designated Aviation Medicine Examiner.
The Captain’s duty records indicated that he was off-duty on 3 December 2013.
The certificate shows that Zaharie’s previous aviation medical examination was conducted six months prior, on 30 June 2013 (the duty records indicated that he was off-duty on that day). The examination report indicates that his next examination was due by 30 June 2014.
It is pretty clear, I think, that Captain Zahaire had not only seen a doctor for a medical examination only some four months prior to the disappearance of MH370, but that that examination was part of a regular schedule.
Thank you Mick.
Looks like the normal search operations have been paused for the last 3+ days. If it’s because of weather, the wave height forecasts don’t look good for the next few days either. Maybe there’s a reason (equipment/crew) why the departure date (Jan 29) back from Fremantle is fixed. Otherwise, it makes little sense why they would postpone the departure to Fremantle by 9 days and then spend most of that time doing nothing. The idea of POI inspection seemed possible in the first day or two but I don’t think it’s likely anymore (it’s taking too long and the vessel has drifted quite far from the original stoppage site).
@eukaryote: If bad weather was anticipated, and there was a targeted area to investigate, one or more AUVs might have been launched before the bad weather anticipating recovery after the bad weather has passed. My guess is it’s possible to extend the battery endurance by moving at low speeds, which would also improve the scanning and imaging resolution.
Or, Armada 8605 is simply waiting out the bad weather with no AUV missions active.
We simply don’t know.
@Julie
My personal focus is two main things: (1) crash site estimate, and (2) I try to document reasons for denial, which I saw in CNN MH370 early coverage. Lately I am using AI171 and TWA800 to add to the denial factors list for aviation accidents.
Anyways, if it is me doing the POI estimate, right now I see it likely as intentional/planned/maneuvers to end. I do not focus too much on blame as I see many problems.
I do think there is an argument of culture differences between Western and non-Western societies. In fact I now hypothesize that cultural denial actually adds to the affected pilot’s temptation, along with cockpit freedom factors that have not in the past worked to add more deterrence.
Some slow improvements after 12 years. EU leading. 25-hr CVR and apparently new ELT doing a better job of confined space monitoring of life parameters, reportedly with some focus on increasing tamper resistance.
It looks like Armada 8605 has been facing into the wind and drifting backwards for the past few days. Right now she is outside of the search area. It seems unlikely that they have been conducting any AUV operations since 01/15 or 01/16. The winds appear to be letting up.
I think they are simply following ‘a steep sided feature’ (either a ridge line or possibly a valley) that was only party and poorly scanned before.
Perhaps OI have learned a lesson from the ARA San Juan search. You may remember that the ARA San Juan was found in a shallow valley that OI had previously written off as ‘done’. They had scanned across the valley, i.e. transversely, and “missed it”.
They only went back and had another look, because the ‘acoustic experts’ said it was there – and so it was.
To:
The Authors of UGIB, in particular @DrB, @Victor, @Andrew; and
@Paul Smithson; and others.
Paul Smithson, in his “one-page summary of the Smithson solution”, has postulated the following effects upon fuel flow of “bleed air off (-3.0%), ram-air door closed (-0.4%)”, amongst others.
Thus, per this postulation, these two effects result in a combined effect of -3.4%.
This is a larger estimate than previously made by UGIB, for example.
Paul, in your comment January 9, 2026 at 6:38 am, you state: “Those savings are based on authoritative referenced sources and empirical application.” This statement is of limited use in attempting some form of reality checking of your estimate.
In general, at the level of the estimate of “-3.4%” there may be need to consider other factors.
Firstly, for a normally functioning passenger carrying aircraft at cruise altitude the fuel penalty of having the requirement to provide suitable and comfortable environmental conditions for the passengers and crew is a constant load situation. The load will be the same at the top of initial climb to cruising altitude right through to fuel expiration if such a situation eventuated. An accommodation for this load by the mechanism of allocating a percentage of the fuel flow to the load will longer “cut the mustard”. The required fuel flow to accommodate the load will be virtually constant.
Caveats: (i) The load will vary during the flight due to various in-cabin situations, such as passenger/cabin crew activity level and activity support equipment heat loads, etc, but will remain fairly constant throughout a long flight, or, at least, a constant load assumption should be satisfactory “for our purposes”. (ii) Also, Step changes in altitude as the flight progresses will/may have a minor effect. (iii) A secondary consideration may be that, although the load will not change as the flight progresses, the flight at “constant” altitude will consume fuel, lighten the aircraft, and reduce drag, resulting in some reduction in required flight support fuel flow as the flight progresses. This factor will/may possibly have a very minor effect on the percentage of total fuel flow required to support the constant load cabin environment support system/s. No compensation for this very minor effect should be necessary.
Another factor for which consideration may be necessary at the level of the estimate of “-3.4%” is that during other stages of the aircraft operation the requirements for the cabin environment support system/s will differ from those at cruise. In particular, for example, will be that whilst on the ground, additional heat loads (sources) may apply such as those due to high temperature and radiant ground surfaces. The necessary fuel to accommodate these and/or other low altitude conditions will differ from, and presumably will be higher than, the cruise load.
In particular, the above mentioned “another factor” has another implication. In UGIB “The Final Resting Place” of 2020, Section A.11.6.(a) Minimum Equipment List (MEL) in part states: “For two recirculation fans inoperative, increase flight planning fuel by 0.7%, plus 0.3% for each additional inoperative fan.” It is unknown whether or not the “0.7%”, for example, refers to 0.7% of the overall otherwise planned aircraft fuel load, or whether it refers to 0.7% of the “instantaneous” (for want of another descriptor) fuel flow. It is suspected that it refers to the planned fuel load. It is left for others to estimate if this would mean that the (constant) load at altitude would actually be significantly greater than 0.7% of total fuel flow during flight at cruise altitude.
As you may guess, it may take some time to derive some serious reality check of Paul’s postulated “-3.4%”, and even so it certainly means some revision for somebody concerning constant fuel flow “modelling” for the constant load cabin environment support system/s.
And for me, it now seems that I lost some rather large Excel files in late 2023 towards the demise of my previous laptop in early 2024.
Refer previous, above, on January 18, 2026 at 10:10 pm.
An accommodation for this load by the mechanism of allocating a percentage of the fuel flow to the load will NO longer “cut the mustard”.
HOW did I miss that “NO”?
For repetition: Accommodation for a Constant Environmental System Load by allocating a percentage of the main engine fuel flow RATE is not suitable.
@George G
What is the environmental load if the passenger cabin is depressurised?
Even under normal circumstances it’s hard to imagine that with a complete cabin air change every few minutes at wildly differing altitudes there would be a “constant environmental system load”.
Disclaimer: I have no knowledge of aircraft systems, however I do somewhat understand the “no free lunch” principle (AKA the laws of thermodynamics) from tertiary engineering studies.
@John,
You ask: “What is the environmental load if the passenger cabin is depressurised?”
It is assumed that there will be NO fuel consumed due to the function of, the environmental support systems when the passenger cabin is not pressurised. For this to be the situation the systems (the “Air Packs” and associated systems) which provide the pressurisation could have been turned off.
The crux of the matter here is whether “we” can explain if the aircraft 9M-MRO whilst conducting the last Flight MH370 could get to a certain latitude along the seventh Arc.
@Paul Smithson is proposing that it could have gotten to a latitude 39.565 degrees South by the time of 19:00 UTC March 8th 2014. These details are directly from his plot onto Google Earth. Paul estimates that is four minutes after the aircraft ran out of fuel.
We know that a a normally functioning passenger carrying aircraft at cruise altitude CANNOT make this distance without some unusual circumstances.
Regardless of other aspects of Paul’s “Smithson Solution” he has estimated that the environmental support systems (otherwise referred to here as the “Air Packs”, and supporting equipment) was not functional for a very large portion of the flight. In this way, or by this means, the aircraft can “store” fuel that it would otherwise have consumed and utilise that “stored” fuel to continue flying for a greater distance than it otherwise would.
John, a commercial passenger airliner in cruise under normal circumstances does NOT fly at your “wildly differing altitudes”. It flies at a virtually constant pressure altitude.
At those ideal conditions the passenger cabin will be supplied with sufficient fresh and recycled air (a combination of both) to maintain a reasonable temperature and pressure controlled environment suitable for the passengers, cabin crew and flight crew.
Note that “here” we are not discussing the time that the flight took or (macroscopically) when the flight ended, but COULD IT HAVE GOTTEN SO FAR in that timespan.
@George G
You said: It is unknown whether or not the “0.7%”, for example, refers to 0.7% of the overall otherwise planned aircraft fuel load, or whether it refers to 0.7% of the “instantaneous” (for want of another descriptor) fuel flow.
The Boeing manuals are not consistent, as described in UGIB Section A.11 para 6. The MEL says to increase “flight planning fuel”, whereas the AMM and FCOM refer to an increase in “fuel consumption”. We took it to mean fuel consumption.
@George G
Thank you for the detailed response.
I accept that the air pack impact on fuel consumption would normally be constrained within fairly tight bounds. My reference to wildly different altitudes was not to imply that any particular flight including MH370 zigzags up and down in altitude during one flight, but to proffer that altitude must impact on air pack consumption as a consequence of the no free lunch rule.
It is a reasonable conjecture that for most of the flight the cabin would have been depressurised. In order for the PIC to be able to set the postulated course they must have removed any interference to their actions.
I am not qualified to comment on if the air packs being off for a substation portion of the flight facilitates the “Smithson Solution”.
@Andrew,
Thanks. Appreciate your confirmation.
Have you reviewed later versions ?
@George G
The wording in the manuals hadn’t changed as of early 2021. I haven’t had access since then.
@ventus45 said: I think they are simply following ‘a steep sided feature’ (either a ridge line or possibly a valley) that was only party and poorly scanned before.
That’s not what the bathymetry data suggests. It doesn’t appear to be following terrain.
The vessel’s path is not like anything we’ve seen before. It’s a slow track that started within the 17-NM width outer proposed area and is now about 14 NM outside of the proposed area. There’s no doubt that the timing of the slow path coincides with a challenging sea state, but why it continues on this track is difficult to guess.
[Later]
An official statement on the progress of the search says that since January 15, search operations are paused due to inclement weather, with all three AUVs onboard the vessel.
@George G.
The loading in cruise should be a more or less steady state since the fundamental is a fixed minimum air mass per passenger by regulation. Experts in this field agree that the energy equivalent at cruise is somewhere in the region 1.3kW per passenger. The system would be sized according to the maximum passenger density configuration for that aircraft type. Bear in mind that pneumatic energy production is less fuel efficient that electrical.
Voth, V., Lübbe, S. M., & Bertram, O. (2024). Estimating aircraft power requirements: A study of electrical power demand across various aircraft models and flight phases. Aerospace, 11(12), 958. https://doi.org/10.3390/aerospace11120958
Sholz & colleagues used engine models to derive an estimate of the fuel requirement in cruise to produce the necessary bleed air for an Airbus A320, arriving at a figure of 2.5%.
Scholz, D., Seresinhe, R., Staack, I., & Lawson, C. (2013). Fuel consumption due to shaft power off-takes from the engine. Aachen : Shaker. https://doi.org/10.15488/4462
Wilfred Lehle was instrumental in the design of the ECS for Airbus A330 and provides a specific quantification of the fuel burn in the final figure of the paper linked below.
https://www.fzt.haw-hamburg.de/pers/Scholz/dglr/hh/text_2006_06_01_AirSystems.pdf
The certification sheet for RR Trent 800 series engine specified that maximum continuous bleed load (which I accept ECS only comprises a % of, albeit the largest fraction) is 5% of “W24”, though I am not sighted on precisely what relationship that has in turn with fuel burn in cruise. https://www.easa.europa.eu/en/downloads/7717/en
There are numerous citations – in published articles by experts in the field of aircraft propulsion and/or environmental control systems indicating 3%-5% as a typical ballpark number for the fuel cost of bleed air. There is no public domain number specific to the B777-200ER. However, given the fixed thermodynamic energy determinant, I do not anticipate that there would be vastly different load extraction across aircraft of similar passenger load.
The literature on “more electric aircraft” points to a number of ways in which use of electrically-generated rather than pneumatic power may improve overall fuel efficiency by >1%, despite the increased weight of the on-board generators (1MW in the case of B787). The saving is mainly due to the ability to fine-tune the load to the ECS requirement, reducing “wasted” excess bleed, particularly in climb.
Feel free to do your own research to substantiate my claims.
Your statement “For repetition: Accommodation for a Constant Environmental System Load by allocating a percentage of the main engine fuel flow RATE is not suitable.” does not make sense (see Voth article cited above).
Re: fuel, I suggest cut RT IDG at IGARI, who needs it? Could explain need for LT GEN at 1825. If RT IDG cut, I assume less RT eng fuel consumption, and then I am not sure if going to INOP (one engine) still saves fuel vs. two engines at lower altitudes.
Could you please add coordinates to the graphic?
@Paul Smithson,
Thank you, belatedly, for your references.
Completely different, I expect, searching is ongoing in an attempt to assist in base environmental load upon the engine bleed flow and energy.
Paul,
Please consider an aircraft at the top of final climb to cruise.
The all up weight will be high as the fuel load will still be relatively high.
Compare this with the same aircraft flying still in cruise but towards the end of fuel endurance For example 7 whole hours later. The all up weight will have reduced significantly. The required fuel flow rate to the engines will have also reduced. This is simply because the aircraft towards the end of flight having less weight will also need less lift. The result is a MEASURABLE reduction in fuel flow rate due to a reduced required thrust, due to reduced drag. However, the system/s supporting the passengers and crew will remain as when the aircraft first entered the cruise stage of flight.
@TBill,
Yes, A possibly valid configuration to consider might be whether or not the Person In Control took the (small) risk of disconnecting one IDG altogether. Simply selecting it off-line at diversion should have been all that was necessary for any nefarious purposes.
@Andrew:
Captain Aaron Murphy recreated the steep dive suggested by the final BFO values in a B777-200ER simulator in Jeff Wise’s podcast episode S2E4.
My personal takeaway from their simulation was that a recovery was not possible because
(a) The aircraft was probably overspeeding (could cause fatal damage).
(b) The aircraft lacked sufficient altitude for a gentle recovery.
(c) A more aggressive recovery would have broken the airframe due to the g-forces imposed.
My questions, if I may:
(1) Do you agree with the above (a-c) ?
(2) If we nevertheless (for argument’s sake) assume a successful recovery, the plane could not have glided very far after having lost so much altitude. Therefore, I don’t understand the suggestions to “search wider”. How far could the plane have gotten away from arc7 at most, if the dive was indeed as steep as suggested by the final BFO value ?
@Victor Iannello, you wrote back in 2018:
« Nonetheless, if it can be proven that the aircraft was in a controlled descent at the time it impacted the ocean, the distance from the 7th arc could be greater than 120 NM, as there could have been a long glide after fuel exhaustion. »
https://mh370.radiantphysics.com/2018/07/30/mh370-safety-report-raises-many-questions
« If the flight was controlled after fuel exhaustion, an efficient, stable glide starting at about 40,000 ft could have achieved a distance of 120+ NM from the 7th arc. »
https://mh370.radiantphysics.com/2018/08/19/end-of-flight-simulations-of-mh370
(1) Is this still valid or have newer analyses since changed the figure of 120 NM ?
(2) How do you reconcile “an efficient, stable glide starting at about 40,000” with the steep dive indicated by the final two BFO values ? Wouldn’t this stable glide start at a much lower altitude after levelling off from the steep dive indicated by the BFO values (and therefore impact the water at a far shorter distance than 120 NM) ?
@Peter Norton,
I don’t think we have any evidence of any sort of glide. Unfortunately, in JW’s test they didn’t let the spiral descent develop, they made manual inputs to level the wings for some reason. A spiral descent would have achieved the higher ROD they were trying to demonstrate.
Evidence of high ROD and high energy crash-
1/ high ROD from last INMARSAT signal
2/ No subsequent IFE log on
3/ Most of the debris indicates high energy impact, of note the vertical stab piece, difficult to explain in any controlled landing
4/ Flaps in retracted position, trailing edges could not be dragged through water.
5/ Relatively undamaged pieces all from right wing, torn away before ocean impact, possibly on right wing failure due excessive aerodynamic/ G loading
So it should be close to the 7th arc. But just how accurate is that arc? Perhaps we should view it as a 100nm wide line due to possible inherent errors.
@Peter Norton
If a very high ROD was maintained for a prolonged period, as depicted in Jeff Wise’s video, then it’s likely the aircraft was unrecoverable. However, the descent rate implied by the final BFO is only a snapshot of what was occurring at that particular time. I think it’s likely the rate of descent continued increasing and the aircraft crashed close to the 7th arc, but we can’t be certain.
Consider the following:
The ATSB estimated the left engine failed at 00:17:30 and the right engine failed up to 15 minutes earlier. The aircraft could not have maintained altitude after the first engine failed and would have drifted down under autopilot control to around FL290. At some point during that descent or shortly after the aircraft had stabilised at about FL290, the second engine failed. The autopilot then disconnected and the aircraft continued descending. Two minutes later, at 00:19:29, the aircraft was descending at around 5,000 ft/min, increasing to around 15,000 ft/min at 00:19:37.
Now let’s say the aircraft was at FL290 and started descending at around 3,000 ft/min after the second engine failed at 00:17:30, increasing to 5,000 ft/min over the next two minutes. The altitude loss is about 7,800 ft, assuming the aircraft took about 10 seconds to establish the initial rate of descent. That puts the aircraft at about FL212 at 00:19:30. The altitude loss over the next 7 seconds is about 1,300 ft (as the rate of descent increased to 15,000 ft/min), putting the aircraft just under FL200 at 00:19:37.
I’m not suggesting this happened, but I think a pilot could recover from such a dive and continue a glide descent. In that scenario, the aircraft could have glided a considerable distance.
A few points on the video:
1. The engine failures were not the result of fuel starvation as claimed, because there was still a tonne of fuel in each wing tank. The ENG SHUTDOWN EICAS message after each failure seems to indicate the engines were shut down by the fuel control switches.
2. The simulation did not show the effect of residual rudder input from the TAC after the second engine failed. The Boeing simulations described in the SIR showed that residual rudder caused a slow roll to the left after the autopilot disconnected, causing the aircraft to enter a descending spiral turn to the left.
3. Considerable manual inputs were required to establish the required rates of descent with wings level. I think it’s far more likely the high rates of descent developed as result of the aircraft entering a spiral dive.
@Tim, @Andrew:
Back in 2017, I ran simulations which reproduced the final BFO values with no pilot inputs by introducing “lateral asymmetry” which resulted in a banked steep descent after fuel exhaustion.
https://mh370.radiantphysics.com/2017/06/04/mh370-end-of-flight-with-banked-descent-and-no-pilot/
The conclusions were:
The BFO values at 00:19:29 and 00:19:37 suggest that MH370 was descending at an increasingly high rate. With the assumption that there were no pilot inputs, the descent rates suggest the aircraft was in a roll as it was descending. According to Boeing documentation, when the hydraulic pressure is supplied by the ram air turbine (RAT), the aircraft banks to the left, which may have been the cause of the rolling during the descent. Using the PMDG 777 model add-on to FSX, simulations were performed for a banked descent that matched the descent rates suggested by the BFO values. In the simulation, the aircraft impacts the water at speeds around Mach 1 and with nearly vertical nose-down pitch. The distance of the impact point from the crossing of the 7th arc is less than 5 NM.
In that article, I presented the possibility that the cause of the roll after fuel exhaustion was the RAT providing hydraulic pressure for the right flaperon but not the left, which produces a roll to the left. This was based on a training note in the Aircraft Maintenance Manual that Don Thompson found:
When the RAT is extended and hydraulics off, the airplane rolls left. Two to three units of right control wheel rotation are necessary to hold the wings level.
@Victor
Thanks, that’s something else that’s not modelled in Jeff Wise’s simulator video.
@All
fyi new OI status update available now in some MH370 FB groups, probably on MH370 Families page.
@All,
There is a straightforward and reliable method to put an upper limit on the potential savings in fuel flow by turning off the cabin pressurization, the cabin air conditioning, and optional electrical equipment.
I can separate the Boeing tables of fuel flow into two components – the thrust-producing flow and the non-thrust-producing flow required to drive external loads. First I compute the engine thrust required using the lift and drag coefficients for this aircraft type. Next, knowing the Thrust Specific Fuel Consumption (TCFC) for the Trent 892B engine (57 kg/hr/engine/kN), I compute the fuel flow needed to produce that thrust. The additional flow included in the Boeing tables drives the external loads (both electrical and pneumatic).
With my current fuel model I can predict the total Boeing flow at LRC, Holding, and M0.84 speeds within about 0.5% at one sigma by assuming this constant external-loads flow produces the equivalent power of 3.4 kN of thrust. That is, this “extra” flow beyond what is needed to provide thrust varies with pressure and temperature and Mach in the expected ways, but a simple and accurate model is that it represents a constant “equivalent thrust”, and this flow is assumed by Boeing in their flow tables to be a constant-external-power load. Thus, the external-loads flow will vary with flight level and total air temperature (because these parameters affect the engine performance), but not with aircraft weight. Therefore, at a constant altitude and temperature the external-loads flow will be unchanged by the reducing aircraft weight as fuel is consumed.
The external-loads flow has two contributors: the power needed to drive the gearboxes which turn the Integrated Drive Generators (which power the electrical equipment including the air conditioners and fans) and the power needed to provide the compressed air for cabin pressurization and pneumatic pumps.
By analyzing the Boeing flow tables, I found the “equivalent thrust” value of the external-loads flow is 3.4 kN per engine. At a cruising altitude of FL390 and an aircraft weight of 200 MT (resembling MH370 circa 19:41), the LRC Mach is 0.840 and the thrust per engine is half the drag, or 100.2/2 = 50.1 kN/engine. Under these conditions the total flow of 3,054 kg/hr/engine provides 50.1 kN of thrust and 3.4 kN of equivalent thrust for the external loads. The percentage of the external-loads flow is 6.4% per engine, or 195 kg/hr/engine.
Nearing fuel exhaustion the weight is 174 MT and the LRC Mach is 0.836 at FL390. The drag is 87.9 kN and the external loads are equivalent to the constant 3.4 kN of equivalent thrust. At this lower weight the external loads consume 7.2% of the total flow of 2,714 kg/hr/engine, which is the same 195 kg/hr/engine.
Therefore, an upper bound on the fuel flow savings that can be achieved in 9M-MRO by turning off the cabin pressurization, the air conditioning, and non-essential electrical equipment is 195 kg/hr/engine or about 6-7% of the flow. Some portions of the external loads, both electrical and pneumatic, are required to maintain flight, but I do not have any reliable estimates for these. It is therefore unclear at present exactly what fraction of the external-loads flow could be saved by extreme measures. I would estimate that a reduction by at least 50% is possible.
In UGIB (2020) we estimated the savings provided by turning the air packs off to be 1.6%. That estimate has been validated recently by Boeing. The total fuel savings required to exactly match the UGIB route and the MEFE time is 2.4%. We did not allow for reduced electrical loads in UGIB (compared to what is nominally assumed by Boeing in their fuel flow tables). It seems likely that some of the assumed electrical loads during normal operation were shut off which produced a flow reduction of at least 0.8%. Electrical load reductions are possible by turning off the exterior lights (nav, strobe, and logo), the primary cabin overhead lights, the galleys, and the air recirculation fans. These can be shut off by switches on the flight deck. At least 30 kg/hr/engine can be saved by these pilot actions, which is about 1% in fuel savings. Therefore, reasonable electrical load reductions during the SIO Route can fully resolve the small UGIB fuel shortfall and enable slightly longer-range routes to Arc 7. Larger fuel savings are possible by further reductions in non-essential electrical loads. It seems unlikely that the fuel savings produced by cabin depressurization and shutting off the air conditioning are much larger than the UGIB estimates, but additional electrical load reductions could have produced additional flow reductions, providing greater range and potentially enabling a more southerly crash location. I think a flow reduction of at least 100 kg/hr/engine is possible without hindering auto-piloted flight. Thus the range could been extended by the equivalent of at least 1.0 MT of fuel during the SIO Route after circa 19:00, compared to the range computed using the Boeing flow tables. It is possible that flow savings of more than 100 kg/hr/engine could have been achieved through extreme measures, but these must be noticeably less than 195 kg/hr/engine to allow auto-piloted flight.
@Victor Iannello
I read the 2017 paper that you linked, and it occurred to me that one possibility worth considering is a forced high-speed descent driven by pilot input, potentially leading to a partial or major in-air breakup prior to ocean impact.
In this scenario, the aircraft would be pushed well beyond its normal operating envelope in a steep descent, where dynamic pressure and structural overload could cause significant disintegration before reaching the water.
Under such conditions, there wouldn’t necessarily be a single compact “crash site” on the seabed. Instead, the wreckage could form a diffuse debris corridor, with heavy components such as engine cores and landing gear descending more vertically, while lighter structures and control surfaces scatter over a wider area.
I’m interested in how others here would view this possibility, particularly in terms of how it might affect search geometry and methodology. If the final descent involved a steep, high-energy breakup before submersion, it seems the search approach would shift away from expecting a single concentrated wreckage field and instead prioritize locating key survivable components, such as engine cores and other dense metallic structures.
@all:
A few comments on the recent discussion.
1. First, the JW simulation is not valid because they did not set up the simulation to run out of fuel. Instead, they simply cut the fuel off manually (with fuel left in the tanks). In the simulations Paul and I ran in the UAL Class D simulator on Nov 2, 2014, we set up the fuel in each tank to a small value (800 lbs, typ) and let the plane fly on autopilot until MEFE, then continued to watch without touching anything. This is the only way to determine how the plane actually responds to MEFE. Our sim’s confirmed the information provided by Boeing re time for RAT to deploy, time to APU start, TAC response, etc. (We did not attempt to simulate the 1000’s of live pilot scenarios.)
2. Our simulations included trials with the left engine running out of fuel first, right engine first, and both engines near simultaneously. All the sim’s resulted in impact near the point of MEFE with various spiral descent patterns…but none, including those with Phugoids, “glided” anywhere.
3. There have been many estimates of how far a B777 can glide. But, having flown a dozen gliders over the last 55 years… 1000’s of flights…I can confidently state that theoretical values over 100nm are extremely unlikely, even in the hands of an experienced pilot.
4. Note that the the time of MEFE and BFO data tell us for certain that the plane was not anywhere near 40,000 feet at 00:19:37. At the Dec 7, 2017 OI Meeting in London, we discussed the likely altitude at 00:19:37. Pete Foley and I decided to settle on 20,000 ft for the purpose of defining a search area. Everyone agreed. If the plane was at 20,000 ft and descending at 15,000 ft/min then, there is absolutely no way the plane recovered and glided at max L/D to some point 100 nm from the arc.
@Bobby,
I was interested to see that your “residual” method for secondary power consumption arrives in the ballpark: 6.4%-7.2% of fuel flow depending on weight/altitude. By my reckoning the IDG portion is something like 1.5%. The rest of it should be pneumatic – and the vast majority of that is the “customer bleed” for environmental control systems, implying something close to 5% (compared to the 3%-5% numbers that I have been citing). It is also materially higher than the 1.2% (air pack, excl. ram-air intake drag) in UGIB looks very low.
I have already demonstrated the feasibility of attaining the required endurance if you assume even a “conservative” 3% saving attributable to bleed air off after 17:22.
It seems to me that your new estimate lends further support to the fuel saving magnitude that I have documented, and the fuel-feasibility of a 39.5S terminus.
@airlandseaman
With all being said, it really does not make sense to search up to 70NM from the 7th arc, eve 50NM seems to much
@Paul Smithson,
You continue to ignore and misconstrue certain facts and assumptions about the external loads and their effect on the fuel flow of 9M-MRO.
First, the UGIB estimate of the impact of the air packs being off is 1.6%, and this estimate was recently validated by Boeing. Your assumption that the air packs savings is 3-5% is in conflict with both UGIB and Boeing.
Second, you have not presented any Boeing documents which present either the total electrical load or pneumatic load used by them in preparing their flow tables. Instead, you have made assumptions about these numbers, but they are not validated for 9M-MRO. They are just your guesses.
I am doubtful that the additional fuel savings obtained by turning off many of the typical electrical loads is sufficient to reach 39S. However, the lack of Boeing flow tables for ECON modes is a limitation in evaluating all speed modes for feasibility. That is exactly why I am working on a new fuel model – to allow evaluation of all ECON Modes, if that is possible to do using the available Boeing information.
Your proposed crash latitude is not a new idea. Several people, me included, looked at this early on in the summer of 2014. Back then we did not have the capacity to make reliable flow estimates. That situation is now much improved. However, there are still several make-or-break issues with such a route. First, one has to demonstrate fuel feasibility. Making a guess as to fuel savings does not constitute a demonstration of feasibility. Second, one much demonstrate a complete path from 18:22 to 00:19. This constraint has two parts. First, the path must match ALL the 18:25 to 18:40 BTOs and BFOs. Second, the 19:41 location predicted by fitting the satellite data from 19:41 to 00:11 must be reachable. To date, I have not seen a route ending circa 39S that does both of these. Until one can meet these last two constraints, the fuel sufficiency is immaterial.
Thank you very much, Mr. Iannello, for allowing me to lend my theory on your excellent blog.
I eagerly await success in finding MH370.
@Andrew, @Victor, @airlandseaman
@Andrew: Thank you for your – as always – thorough analysis.
@airlandseaman: You said at the London OI meeting, everyone agreed with the 20,000 ft altitude for the purpose of defining a search area. At this point (00:19:37 at 20,000 ft and descending at 15,000 ft/min according to BFO), how far could the plane glided in the worst case (immediate recovery efforts, best glide speed, etc.) ? Did this meeting agree on a maximum distance ?
@Victor: You wrote “an efficient, stable glide starting at about 40,000 ft could have achieved a distance of 120+ NM from the 7th arc”. I don’t understand the logic behind that statement. Could you explain ? The final 2 BFO values mean a considerable altitude loss, no ? So why does your glide start at 40,000 ft? How is a glide starting at that altitude possible despite the steep dive indicated by the BFO values ? How does this stable glide at 40,000 ft produce these BFO values ?
It seems there is some agreement that any possible glide would have to start at a much lower altitude:
* Andrew’s scenario above “puts the aircraft just under FL200 at 00:19:37”. The subsequent recovery after 00:19:37 would still cost further altitude.
* ALSM agrees with that altitude calculation: “At the Dec 7, 2017 OI Meeting in London, we discussed the likely altitude at 00:19:37. Pete Foley and I decided to settle on 20,000 ft for the purpose of defining a search area. Everyone agreed. If the plane was at 20,000 ft and descending at 15,000 ft/min then, there is absolutely no way the plane recovered and glided at max L/D to some point 100 nm from the arc.”
@Peter Norton: The 120-NM glide was meant as upper limit without assumptions about lost altitude during the initial dive. That’s why Bobby and I recommended the 70-NM glide based on more practical considerations. A glide of this distance would require a “dive-glide-dive” scenario, which seems less likely than an impact relatively close to the 7th arc, but can’t be dismissed.
https://mh370.radiantphysics.com/2023/06/12/improved-drift-model-and-search-recommendations-for-mh370/
personally I say MH370 had descended to FL100-FL200 well before Arc6, at Arc7 we are probably seeing a second descent to fly in the thick clouds at say FL050. The glide allowance was used up before Arc6, but at Arc7 there is still a couple tons of fuel to probably get 100-200nm away from Arc7. This path needs to be conducted in a southeast region (30-32s) where distance to Arc7 is shorter, in that region of Arc7 where the ATSB historical graph shows that there is greater potential distance from Arc7, which we have ruled out.
Despite some challenging weather, it’s looking like Armada 8605 will complete the scanning of all of the outer (SE) proposed area before heading to Fremantle in several days. They have reported no interesting contacts.
Peter Norton: We reached a consensus to use 20,000 ft as an assumption for the purpose of defining the nominal 7th arc. Clearly it was not at 40,000 ft at 00:19:37. And given the ROD then (15,000 ft/min), the TAS would have been very high (close to Mach 1?) and the L/D relatively low. This is why we continue to believe that the POI was very likely to have been within ~20nm of the nominal 7th arc. But we do not know for sure what happened between 00:19:37 and 00:21:06 when the 9M-MRO IFE system failed to logon. We can be fairly sure the plane was already in the water or inverted at 00:21:06. (The APU could not have run out of fuel by 00:21:06.)
The best L/D for a B777-200ER is about 18/1 at ~220knts. If the plane was at, say 20,000 ft and *flying at the best L/D Speed*, it could glide ~59 nm. But we know it was not anywhere near 220 kts. I haven’t seen a polar plot for the B777, but I would guess the L/D at 500 kts is closer to 10:1, and the data suggests the speed was even higher.
@Bobby,
If you were not so busy composing that patronising kerfuffle of self-aggrandisement, you might have noticed there was an important statement-cum-question in in my post. Does it not strike you a as a bit strange that your new estimate of secondary power loads (6.4%-7.2% of fuel flow at cruise) is more than five times your estimate of packs-off (~1.2%)? You may choose not to answer, but please don’t give me the diversionary BS.
Following this great discourse for a couple of years, as a non-scientist but intellectually curious. Thank you for sharing your insights and process.
I’m sure this has been seen by many here already, but I found this recent (Jan 21) exploration of the bathymetry of the general area quite fascinating.
https://www.youtube.com/watch?v=AsBd3admKtg
@airlandseaman + @Victor:
Thank you both for this information and clarification. This is extremely helpful.
Based on this apparent consensus, I don’t see any way in which the plane could have ended up further away from arc7 then your stated upper limits, if we can trust the BFO data.
As a result, I would agree with those of you who recommend to first
(1) fill in the data holidays [if any remain], then
(2) as ST says “Ed’s Java Anomaly as it is a small area to search and has scientific basis per Victor’s previous blog post as well” and lastly
(3) expand the search area along-track (rather than “searching wider”). I would even suggest to expand the lower end (in SW direction) despite fuel limit and timing seemingly excluding this possibility, just to make sure and given that the area along the arc in the opposite direction (NE) appears equally unlikely (if I correctly read your analysis of the BTO/BFO data and the drift models).
@Victor: BTW, I may be misreading your map but isn’t the southwestern border of the area labeled “searched by ATSB” missing ?
Interesting look at the original Fugro bathymetric data in the current search area. Video by geospatial engineer Sam Roberts: https://www.youtube.com/watch?v=AsBd3admKtg
His first video on the wider search area is here: https://www.youtube.com/watch?v=MSe3KWS_A_k
You can really appreciate the complexity and vertical scale of the seabed at this location.
@airlandseaman
It is fairly easy to generate a reasonably credible ‘hypothetical glide polar’ for any conventionally configured aircraft by using the ‘non-dimensional ordinates’ of the ‘universal polar’ as derived in Appendix 7 (pp 373 – 381) of ‘New Soaring Piot’ Third Edition 1977 (copy to hand) by Anne Welsh, Lorne Welch and Frank Irving. Back then, I found that this table translated pretty to the aircraft I was flying.
The ordinates are shown in Table A7.1 on page 374. For the benefit of anyone who may be interested in knocking up a spreadsheet and running a few ‘what if’ scenarios, I have just taken a photograph of page 374, and uploaded it here. Please ignore my notations (made a hell of a long time ago – last century – not this one).
With regards to 9M-MRO specifically, after much net research (three or four years ago) regarding drag penalties for windmilling large open fan engines, I estimated 9M-MRO’s best L/D of 17 (with both engines windmilling) at 200 KTAS.
Directed at nobody in particular:
“It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.”
Mark Twain
@Tim: Many thanks for your feedback. I agree with everything except this:
I would caution against the notion of a Sully-style “controlled landing” in the Roaring Forties. Are you aware of any successful ditching anywhere but in flat water? I am not. Crashing into high waves at even the lowest possible landing speed would still rip the plane into pieces. It would be like crashing into a concrete wall.
MH370 crashed into 13 ft waves (4 meters) according to historical oceanographic data I pulled from Copernicus in CSV format:
valid_time,latitude,longitude,mwd,mwp,swh
2014-03-07 23:00:00,-34.5,94.0,216.75789,12.105669,3.95555
2014-03-08 00:00:00,-34.5,94.0,216.80539,12.098278,3.95007
2014-03-08 01:00:00,-34.5,94.0,216.85207,12.085715,3.93216
2014-03-08 02:00:00,-34.5,94.0,216.86726,12.371469,4.14154
mwd = mean wave direction, unit: degrees
mwp = mean wave period, unit: s
swh = significant height of combined wind waves and swell, unit: m
I requested the data for the UGIB LEP (S34.53° E93.84°) and received the data for the closest available coordinates:
“The returned data will contains the closest data point on the native grid with no interpolation.”
@Victor Iannello wrote: « A glide of this distance [70 NM] would require a “dive-glide-dive” scenario, which seems less likely than an impact relatively close to the 7th arc, but can’t be dismissed. »
@Victor: For the reason stated above, I don’t believe a second dive is necessary (to explain the debris). Due to the significant swell, even a “dive-glide” scenario would end in a catastrophic crash.
Jeff Wise keeps mocking the dive-glide-dive scenario, saying it is strange that the pilot would pull the plane into a steep dive, change his mind and recover, but then change his mind again and go for a second nose dive. He says, advocates of the southern route have to resort to silly scenarios like this, since all likely resting places have already been exhausted. He would indeed have a point there, if this were not a straw man fallacy: the second dive is not required.
@Tim, this is a longstanding fear of mine. As far as I know, all data has been double-checked (comparison of the BTO/BFO data with other flights, etc.). But I am still worried. It would be a shame if for 12 years everyone were looking in the wrong area due to an error somewhere in the source data – considering that the plane is not where the data points to.
@Peter Norton – For the work so many contributors have put in and many ending with more or less the same area as the UGIB priority search area, think we can just hold on to some hope a closer analysis of what the current AUV scans bring up to identify the debris or plans to cover any remaining data holidays.
Beyond that, definitely think the Java anomaly should be a focus. The one thing that stands out for me as a reader about Ed’s work is how precise the Java Anomaly is on the 7th arc (8.36S 107.92E). Further to that is his work on the flaperon and analysis on his website. It definitely deserves a look in my humble opinion if the plane is not found in the current search.
There is also @Viking’s paper where he describes a mushroom cloud and end terminus around 10kms of 13.53° South, 107.11° East.
These are two people with totally different approaches, but have an end point estimate 5 degrees apart in latitude which should hopefully be a searchable zone.
@Paul’s recommendation would come in next as there are some good points to his recommendation – an experienced pilot who knows it is the last flight of a lifetime could push all the envelopes and do things beyond what are known limits or rather expected limits.
@all
I acknowledge that the prevailing opinion here is that a well-controlled ditching is impossible, and thus dismissed out of hand, but I respectfully disagree. Zahari was a very experienced, highly skilled airman, acknowledged as a meticulous planner, and obviously a very determined man. He should not be underestimated.
@ST
With regards to your comments re an experienced pilot (nee Zahari) you could also add the following to his ‘bag of tricks’:
(a) extensive motorized para sailing experience,
(b) extensive Radio Control PBY experience, (complete with original USN Manuals)
(c) arriving at the end of flight at sunrise,
(d) having the sea conditions (swell / waves / wind) fully visual
(e) knowing full well to align with, and ‘alight’ (not ‘ditch’) on, the top/rear of a crest, along the west north westerly axis, with the rising sun over his right shoulder, producing a shadow of the aircraft approaching on the white foam crests, at his ten o’clock, one by one, giving him better flare guidance than any RadAlt.
Armada 86-05 is announced at Fremantle as:
arrival: 29/12 at 7:00
Departure: 30/12 at 18:00
The posted “next port” is … Pago Pago in the US Samoa Islands …
@Jean-Luc
“29/12” and “30/12” wouldn’t that be December?
It is possible that OI intends to search the second area NW of the 7th arc then move directly on to a new assignment. The second search area shown on CAPTION’s Armada tracking (thank you for making that available) looks smaller than the first.
I think it’s possible (+50% chance) OI has decided to give up. It would be nice if they’d go back closer to the arc. Possibly check Contact 13 from the original search and that area in general.
@Victor: Sorry for having misread your map, I was momentarily confused.
@Paul Smithson,
The fuel flow value of 195 kg/hr/engine I derived for the non-thrust-producing flow assumed by Boeing in their fuel flow tables is the intercept value of new-engine flow at zero thrust. Over a wide range of thrusts the plot of flow versus thrust is linear. There appears to be a very small drop-off in efficiency at high thrusts.
The slope of this line is the Thrust Specific Fuel Consumption (TSFC) which is 57 kg/hr/engine per kiloNetwon of thrust for Trent 892B engines. For example, the typical cruise flow required per engine to provide 100 kN of total thrust (= drag in steady flight) is 2,850 kg/hr/engine.
The 195 kg/hr/engine delivers 3.4 kN of equivalent thrust. The power is the thrust times the speed. At 480 knots the speed is 247 m/s and 3.4 kN of equivalent thrust produces 845 kW or 1,133 HP per engine.
This non-thrust-producing flow powers three categories of loads:
1. Integrated Drive Generator (supplying electricity for aircraft systems),
2. Compressed air (providing cabin pressurization and driving Air Driven Pumps , or ADPs), and
3. Internal engine losses (turning shaft bearings and driving the oil pump).
Currently we don’t have a power budget for the auxiliary loads (electrical and pneumatic) and the internal engine losses.
The internal engine losses (Category 3) are unknown to me but are significant.
The Air Driven Pumps provide the primary power for the L and R Hydraulic Systems and the “demand” power and backup power for the Center Hydraulic System.
The total of 195 kg/hr/engine is valid at standard temperatures. It will be increased by the usual +3%/10C factor for elevated temperatures. It is not a constant fraction of the total engine fuel flow, but the percentage does not change too much for aircraft weights less than 210 MT.
It is possible to significantly offload the IDGs from the flight deck by turning off non-essential electrical loads such as external and cabin lighting, galleys, air conditioning packs, recirculation fans, etc. Turning off the air conditioning packs reduces both the electrical load on the IDG and the compressed air load.
The air conditioning packs consume about 50 kg/hr/engine and do not dominate the non-thrust-producing flow. That said, there are other opportunities for flow savings beyond UGIB by assuming additional electrical loads are shut off. Unfortunately, we don’t have a list of the electrical loads which are assumed by Boeing to be operating for the purpose of creating the flow tables. One can make educated guesses, and I would expect that up to an additional 60 kg/hr/engine might be possible. Making that assumption, there could be additional fuel savings after diversion up to about 720 kg compared to the UGIB fuel predictions for Case #9 with air packs off after diversion, no descent, and no electrical load reductions. This is an additional 1.6% of average flow on top of the 1.6% savings by turning the air packs off, for a total of 3.2% in average flow reduction compared to the Boeing flow tables.
To correct my previous comment, Andrew informs me the accessory drive box on each engine drives both the IDG and the L or R primary hydraulic pump. Therefore, only the Center Hydraulic System has an Air Driven Pump, and it services demand when required and as a backup. Thus, in normal cruise flight, the L and R primary hydraulic pumps are gearbox driven and the C pump is electrically driven by the IDG. So all the primary hydraulic pumps are powered by a portion of the 195 kg/hr/engine non-thrust-producing flow.
Looks as if Armada 86 05 has set off for port.
@Andrew
On the MH370 ACARS, is there any way the pilots could know that the half-hourly was sent at 1707?
@TBill
The MFD Communications Function on the B777 has a REVIEW menu where ACARS messages sent to and from the aircraft can be accessed. On the aircraft that I flew, only those messages sent by the crew were available for review. I don’t know of any way to check that an automatic messages has been sent.
@Andrew
OK thank you I was trying to sort out events at that time
17:06:43 Time of Last Column ACARS Engine Data
17:07:07 ?Apparent time of ACARS Transmission from aircraft
17:07:29 ?Apparent time of Download
17:07:55 Radio “maintaining level three five zero”
Correction above-
Delete the 17:07:07 line that should be 071707 (Date dd Time hrmin)
follow-up to the preceding AI171 discussion:
Whistleblower Report to the U.S. Senate Permanent Subcommittee on Investigations (PSI)
dated 12 Jan 2026
published by: Aviation Herald
@Peter
This whistleblower guy is a Boeing hater, he was on Piers Morgan for AI171, and I give him a Grade D- on aviation aptitude. If you want to fix aviation safety, he should attack Boeing/FAA for not doing more to prevent pilot suicide/mass murder. I do not like using those words, but let’s be clear.
IMHO also why we will never find MH370, the denial and sensitivity of that.
@TBill:
I don’t know who the whistleblower is, but we should always attack the content, never the person.
Regarding mass murder: Yes, that’s hard to fathom.
What additionally bothers me about the AI171 suicide theory are the chances to survive severely handicapped and disfigured when cutting off fuel immediately after take-off.
@Peter
His content is wrong unless we can tell Congress that his attack on Boeing is valid only indirectly, due to stigma, nobody wants to fess up to the true cause (preliminary at this juncture I realize).
Months ago I actually wrote to both of my Senators about AI171, saying that we should not allow lawyers to sue Boeing on false pretenses. Lawyers are hoping they can convince the jury of Boeing hidden glitches, whereas NTSB is not allowed to testify on true likely causes.
Both Senators thanked me for my comments on the wrong crash (the DCA mid-air crash with the army copter). Air India is not something Americans remember. Not on our radar screen because nobody here holds Boeing responsible.
@TBill, @Peter Norton: Eventually more FDR, CVR, and other data will be released on AI171, and we’ll be more certain about the sequence of events and who is at fault. In that absence of data, others will jump in, inject their biases, and perhaps profit from the uncertainty. Boeing is probably wrongly accused, but with its team of lawyers and deep pockets, it will have ample opportunity to make its case at the appropriate time. We just have to patiently watch it play out.
@All
When working on some additional calculations, I found an issue in the scripts for path evaluation that I used in the concept technical note I shared on Dec. 18th. In the implementation of the formula for the path probabilities (eq. 2) in the scripts, the BTOR for arc 5 and the BFOR for arc 6 were not properly taken into account. It actually explains why I found rather large values for the BTOR5 value when optimizing the overall probability. After fixing the issue, the main results and conclusions (like dominant control settings and indicated latitude range) in essence didn’t change, but I had to redo all figures. I carefully double-checked all scripts and I took the opportunity to shorten and rewrite especially section 3.2.
There is one important change, related to Table 3 and the discussion around it: as the probability weighted mean BTOR5 value is reduced, the basic statistics of the “straight path” fit look better and I concluded there is no need to consider omitting the 19:41 (arc2) BTOR value.
https://www.dropbox.com/scl/fi/n7x3dz6p059s709a6j10h/Monte_Carlo_v25_8.pdf?rlkey=7cen75rzge04nc21obciipgqy&st=0utmfaas&dl=0
This article that focuses on possible debris sighted by a RNZAF P3-Orion aircraft, puts the search location much further north: https://mh370.radiantphysics.com/2017/11/25/possible-mh370-debris-seen-in-aerial-search-in-spring-2014/
What happened to this line of thought? The area was covered by OI in 2018, but is it possible that search boundary was too narrow? And what tech did OI have in posession at that time – would it be worth another look in 2026 with their current hardware?
@Niels, tidy update.
Any comments re Fig 4b v25_8 vs Fig 6b v24_6 ?.
Well, folks, what’s your guess? Is OI going to return to the 34-36S search area?
I’m not sure what OI intends, but, of all the areas searched so far, in my opinion, the remaining area of the proposed search in the 2025/2026 campaign has the least chance of success. (I am referring to the rectangle on the inside of the 7th arc.) Regardless of whether the flight was piloted or not, the idea that the plane turned in the opposite direction after the LEP makes the least sense to me.
@Theo asked: What happened to this line of thought? The area was covered by OI in 2018, but is it possible that search boundary was too narrow? And what tech did OI have in posession at that time – would it be worth another look in 2026 with their current hardware?
Yes, that area was searched close to the 7th arc by OI in 2018. The scanning capabilities of the UAVs at that time were not very different than today. But also, the CSIRO drift models and path reconstructions assuming automated flight (no pilot inputs) predict crossing the 7th arc at more southerly latitudes.
That said, we can’t say with certainty that the impact did not occur at that latitude but at a further distance from the 7th arc.
@Peter Norton @Victor
How sure are you that the Whistleblower Report is correct?
@Search
Last few days on X and FB groups, there was some thought that 86-05 may move on to a new job at least temporarily. There is some thought a bigger vessel might be used next.
Even though I do not feel MH370 is in current search area, if I take as the goal to negate a ghost flight scenario to 34s region, I am OK with inside Arc7.
@Viking: I’ll let others answer as this is beyond my expertise. I saw the report on avherald and given its recent publication date I posted the link just in case someone is interested.
@TBill
If you look at it from an “actual time on task / total mission time” ratio point of view, it is pretty clear that both the A78 and A86 are not really as suitable for SIO weather conditions that OI might have hoped.
A7806 = 23rd February 2025 to 28th February 2025 was only 5 days on task.
A7806 = 11th March 2025 to 28th March 2025 was only 17 days on task.
A8605 = 30th December 2025 to 23rd January 2026 was only 24 days on task (but 2+ days hooved to).
Therefore, the Armadas have only achieved a total of 46 (44 ?) search days on task in 11 months, or 41 (39 ?) search days on task since the contract was signed, or only 24 (22 ?) days on task since the announcement of the 55 days of intermittent searching.
I think that OI have reluctantly decided to face reality, because they must now realize that neither A78’s or A86’s are suitable for SIO weather conditions, so I think it is highly unlikely that A8605 will return to the 34S-36S box, and instead will be send to Pago Pago.
If we assume that 33 of the 55 search days on task remain, then presumably they have to be used between now and October (18 month limit on the contract), which is enough time for OI to bring in Island Pride, or even lease another similar vessel.
@George G
Thank you; good point indeed: “Any comments re Fig 4b v25_8 vs Fig 6b v24_6 ?”
The main difference is that in the updated version a wider range of flight levels was used to created fig. 4b compared to fig. 6b in the concept version.
I think that is the reason for the slightly more gradual drop-off towards the south. Fig. 4b is also more noisy as it is based on the “old” set of 20 million flight paths which has fewer paths per flight level than the set used for fig. 6b. In fig. 4b towards the right this results in the strange peak: in the normalization the ratio of two really small numbers is calculated for the bins near S26 deg, resulting in significant numerical inaccuracy.
@Niels, Ta.
Do you know any case of a “well-controlled ditching” in 13 ft waves ?
Or anywhere but in flat water ?
Ok, so ships are not suitable for SIO weather conditions, but well-piloted planes are …
(sorry, could not resist).
We know from OI reports that weather conditions sometimes hampered the operations. But maybe the vessels were also needed in other missions ? Don Thompson would know more about this.
@Victor Iannello
Victor, are the “olive-green areas with white outlines” that appear on maps in the post “New MH370 Search Continues” drawn to scale?
I am preparing something more related to data gaps and LPDs.
@Marijan: Yes, everything should be drawn to scale. The white outline green areas of low-quality data are from files provided by Geoscience Australia. Meanwhile, the mixed-brown shaded areas in the figures are from the ATSB scanning data, also provided by Geoscience Australia.
@ Peter Norton
The issue with a ship for launching or recovering anything ‘over the stern’ is primarily the ‘pitching’ (near stationary with ships head “into the waves/swell”). You have to consider the ‘rise and fall’ of the stern, both in terms of height (metres) and the period (cycle time). Obviously, there are ‘safe operating limits’ for any ‘system’. WE don’t know what they are for OI’s setup on the A78’s and A86’s, but obviously those limits have been exceeded a number of times, forcing them to suspend operations, sometimes for extended periods. I was merely making the observation, that OI did not have these problems to anywhere near the same degree with Seabed Constructor, even when it was operating ‘out of season’.
With regard to presenting an aircraft to the sea, you align “along (parallel to) the waves/swell”, NOT into them. Flying boats used to do it all the time, as a matter of routine, usually in sheltered waters, granted, (sometimes even at night – use of ‘flare floats’ etc – very dodgy) but often in the open oceans as well, when real needs arose. Tricky to do it well, and safely, yes, granted, no argument, but do-able, not impossible, particularly when you know it is a one-time deal, with no need to consider taking off again, or worrying about filling out the insurance claim form.
Regarding the notion that MH370 ended with a controlled ditching, there have been 16 ditchings or controlled-flights-into-water (CFIW) involving commercial jets since 1963.
1. Aeroflot Tupolev Tu-124, ditching, Neva River, 21 August 1963
2. Japan Airlines Flight 2 (Douglas DC-8), CFIW, San Francisco Bay, 22 November 1968
3. ALM Antillean Airlines Flight 980 (Douglas DC-9), ditching, Caribbean Sea, 2 May 1970
4. Aeroflot Tupolev Tu-134, ditching, Ikshinskoye Reservoir, Moscow, 17 July 1972
5. National Airlines Flight 193 (Boeing 727-235), CFIW, Escambia Bay, Pensacola, Florida, 8 May 1978
6. Tarom Romanian Airlines Tupolev Tu-154B-1, CFIW, Levrier Bay, Nouadhibou, Mauritania, 7 August 1980
7. Sudan Air Boeing 707-348C, CFIW on the Nile River near Khartoum, 10 October 1980
8. Japan Air Lines Flight JL350 (McDonnell Douglas DC-8-61), flown into Tokyo Bay on approach to Tokyo-Haneda Airport, Japan, 9 February 1982
9. Faucett Perú Boeing 727-247 disappearance, ditching, Atlantic Ocean approximately 290 km southeast of Cape Race, Newfoundland, Canada, 11 September 1990
10. Ethiopian Airlines Flight 961 (Boeing 767-200ER), ditching, Grande Comore, Comoros Islands, 23 November 1996
11. Trans Arabian Air Transport Boeing 707-351C ST-APY, CFIW, Lake Victoria, Tanzania, 3 February 2000
12. Garuda Indonesia Flight 421 (Boeing 737-300), ditching, Bengawan Solo River, Java, Indonesia, 16 January 2002
13. US Airways Flight 1549 (Airbus A320‐214), ditching, Hudson River, New Jersey, USA, 15 January 2009
14. Lion Air Flight JT904 (Boeing 737-800), CFIW, Bali Sea west of Denpasar, Indonesia, 13 April 2013
15. Air Niugini flight PX73 (Boeing 737-800), CFIW, Chuuk Lagoon, Micronesia, 28 September 2018
16. Transair Flight 810 (Boeing 737-275C), ditching, Mamala Bay, Honolulu, Hawaii, 2 July 2021
Of those, we have at least some photographic records for 14, and we have decent accidents reports for six.
There is also a good paper on the topic – Review of Transport Aircraft Ditching Accidents (Lindenau and Rung, 2009) – that describes the transition from aerodynamic to hydrodynamic to hydrostatic forces encountered in such an event, and the manner in which they interact with a typical airframe.
Two things to note about the data set:
• While it is a small data set, it is essentially complete; ditching/CFIW events are rare but generally well reported. That said, we do need to be cautious with regards to over-generalising.
• Most of the ditching/CFIW events took place in relatively calm conditions (rivers, lakes or enclosed or otherwise sheltered waters). There are only three “open ocean” ditchings; ALM Antillean Airlines Flight 980, Faucett Perú Boeing 727-247, Transair Flight 810.
Of the “open ocean” events, we have good records for only the most recent, the Transair flight 810 ditching. ALM Antillean Airlines flight 980 sank in reasonably deep waters soon after ditching and was never recovered, and the Faucett Perú Boeing 727 disappeared without a trace despite a relatively intensive air (six aircraft) and sea (six vessels) search.
Despite a wide range of touch down speeds (from around 100 – 180 knots) and aircraft configurations (from “clean” (ie no flaps) to landing flaps/gear down) the data regarding the damage sustained in those events in general, and the aircraft components that were liberated in particular, is pretty consistent.
Counter-intuitively, trailing edge devices (TEDs) – flaps and ailerons – do not commonly separate in ditching/CFIW events. Of the 14 events for which we have some form of record, only four evidence TED separations.
Perhaps unsurprisingly when you consider the design and construction of their attachment/movement devices, flaps (even when deployed in landing configuration) essentially never separate fully, even after suffering considerable damage. JAL flight 2 from 1968, a DC-8 CFIW event into San Francisco Bay, and Air Niugini flight PX73, a B737-800 CFIW event into Chuuk Lagoon, Micronesia back in 2018 appear to be the exceptions where there is some evidence of a flap separating in each event.
Notably, there was no evidence of TED separations for the open sea ditching of Transair Flight 810, an event which caused the considerable break up of the B737 involved.
I could find no evidence of seconday control devices, such as spoilers, ever separating in a ditching/CFIW event.
Similarly, I could find no evidence of underwing fairings (flap-track/undercarriage fairings) ever separating, although there were rare instances where flap-track fairings appeared to have been damaged.
Likewise, as might be expected given the lack of TED and secondary control device separations, I could find no evidence of trailing edge fairing separations.
Wing-to-body fairing damage/separations were also unlikely, appearing to occur in about 20 percent of events.
Hull break-ups significant enough to liberate internal components occurred a bit less than half the time. Notably, based on survivor reporting from the ALM Antillean Airlines flight 980 open ocean ditching, there was no significant hull break up for that event.
Of particular note, no ditching/CFIW event, regardless of the severity of the impact and subsequent general level of damage (eg Ethiopian Airlines flight 961), ever occasioned damage to the vertical stabiliser.
Amongst the recovered MH370 wreckage items we have:
• four TEDs (right flaperon, piece of left outboard flap, piece of right outboard flap, piece of right aileron) – seen in less than 33 percent of ditching/CFIW events,
• two trailing edge panels – never seen in ditching/CFIW events,
• a piece of an inboard spoiler – never seen in ditching/CFIW events,
• pieces of flap-track fairings from either wing – rarely seen in ditching/CFIW events,
• pieces of wing-to-body fairings – seen in less than 25 percent of ditching/CFIW events,
• a good number of internal components (piece of the Door R1 stowage closet, a seat-back IFE frame, pieces of cabin floor and internal panelling) – likely seen in about 40 percent of ditching/CFIW events, and
• a piece of a panel from the right-side of the vertical stabiliser – never seen in ditching/CFIW events.
I think that it is fair to say that the recovered wreckage is broadly not consistent with wreckage that is typically liberated in a controlled ditching event.
@Victor Iannello
Thank you, I will be back soon
@Mick Gilbert,
Informative gathering of pieces of research/history.
@Mick: Thanks for that detailed summary of the 16 ditchings and controlled-flights-into-water (CFIW) involving commercial jets since 1963. From the first photos I saw of the flaperon, I suspected a partial in-flight breakup. Much of the debris since then is consistent with a partial in-flight breakup. Larry Vance’s rush to posit the ditching theory was flawed from day one.
@George G
Thanks George.
@airlandseaman
Thanks Mike.
I was frankly surprised at how infrequently trailing edge devices actually separate in ditching/CFIW events. And I was likewise surprised at the apparent robustness of faired items like flap track canoes.
Regarding LV, the less said the better, I think.
@all
Hello everyone
Couple of questions about the non-existent IFE log-on at Arc 7.
The 2023 IG paper gives three possibilities for the missing transmission: the plane crashed before it could be sent, the SATCOM antenna(s) couldn’t see the satellite or the APU had ran out of fuel. I was wondering if there are other possible explanations for why the log-on didn’t come.
1) CAPTION claim in their paper that the whole SATCOM would eventually be shed by the ELMS if both engines were out. Is this true?
2) What about the IFE system itself: would it be powered up in the first place if electricity was only being generated by the RAT and the APU? In other words, should we even expect the IFE to log on to the network at the seventh arc, given that the main engines had ran out of fuel?
3) If the IFE/PASS SEATS switch had been turned off sometime between 18:28 and the aircraft running out of fuel, would it also remain in the OFF position after MEFE and the RAT/APU restoring power to (some of) the plane’s systems?
@Mick Gilbert:
Outstanding research. Very interesting and helpful. Thank you.
Any idea how this alleged “short stop” was determined ?
There was no video footage and since the aircraft was “relatively intact”, it probably could not be inferred by material compression.
@ventus45:
Ok, so we have 3 open ocean ditchings: once the hull survived, once it broke apart, and once with unknown outcome (disappeared despite the calm seas).
1970-05-02: “aircraft relatively intact after the water landing” [wiki]
1990-09-11: Disappearance. “The weather at the time of the disappearance was described as good and the seas were calm.” [wiki]
2021-07-02: aircraft broke in half: [image]
However, in the 1970 ditching the waves were much smaller (0.8m or 1.4 m) than for MH370 (4 m):
coordinates:
according to Wikipedia: 18°N 64°W
according to NTSB report: 17°53’N, 69°14’W
I looked up the significant wave height for both locations in the Copernicus Climate Data Store.
Here are the results for the closest available coordinates:
valid_time lat lon mwd mwp swh
1970-05-02 19:00 18.0 -64.0 112° 5.4s 1.42m
1970-05-02 20:00 18.0 -64.0 111° 5.5s 1.39m
1970-05-02 19:00 18.0 -69.0 119° 5.2s 0.78m
1970-05-02 20:00 18.0 -69.0 119° 5.2s 0.78m
mwd = mean wave direction
mwp = mean wave period
swh = significant height of combined wind waves and swell
@Jean-Luc Marchand:
I was unable to reach you via your contact form, so I am trying here.
I wanted to thank you for making available the Armada tracking link:
https://www.mh370-caption.net/index.php/armada-tracking
This is really helpful. Thanks a lot.
Unfortunately, in the options menu all the ATSB terrain data is missing:
– ATSB Bathymetry
– ATSB Deep Tow Side Scan
– ATSB AUV Side Scan
– ATSB Side Scan
I only see a gray map background.
Could you make this data available again ?
@ventus45
re seaplanes: I am not an aircraft designer, but the seaplanes on https://en.wikipedia.org/wiki/Seaplane look quite different from a passenger airliner, particularly with regards to the engine position. And all the photographs of seaplanes I can find show them on flat water. Can you find a single image of a seaplane in rough seas with 4m / 13ft waves like MH370 crashed into ?
Here is the “FAA SEAPLANE OPERATIONS HANDBOOK”. Chapter 8 on “OPERATIONS IN OPEN SEAS” says:
Special training is required for a seaplane, which the pilots of MH370 did not have, as far as I am aware. You seem to think of swell like highway lanes or airport runways, but out on the open ocean, the pattern is usually more complex and chaotic (as the document above says). You can’t expect a long, flat landing strip.
Ventus45: I agree with your theory. I don’t know exactly when his (possible) dementia first developed, but symptoms are subtle and gradual. Family members, in hindsight, can look back and recognize the small changes in behavior and personality that led up to the “threshold event.”
We do know Captain Shah loved to fly. I believe he intended to park his gigantic aircraft underneath the water as gently as possible, but was foiled by the massive waves. What a coup it would’ve been to outperform the renowned Sully.
Shah at that point was devoid of all empathy, but grandiosity remained.
@ventus45:
I searched for the wave height limit of flying boats.
The ShinMaywa US-2 “was designed with the capability to land in waves up to 3 m (9.8 ft) in height”.
There is a question over at aviation.stackexchange:
“What flying boat has the highest maximum takeoff/landing sea state?”
The accepted answer refers to the Grumman Albatross which is “certified for class 6 seas, which can be swells up to 10 ft. or waves up to 6 ft.”
That’s still less than the 4m / 13ft waves at the 7th arc on 8 March 2014.
And 9M-MRO was no seaplane.
Neither did the pilots have the special training for sea landings.
@Julian
Re: IFE at end – taking crack at
I personally favor: active pilot cut power to SATCOM.
In the ghost flight event, possibilities:
(1) Immediate Crash
(2) Loss of Power/APU
(3) Antenna issue
(4) Load shedding poss
(5) IFE Switch Cockpit* was Off’ed (not)- IG says does not prevent IFE boot-up
(6) IFE Circuit breaker MEC Bay was Off’ed
*IFE Switch Cockpit- this model did not have the switch as orig design, was added later
If APU was on, presumably SATCOM would have capability for power. RAT no.
Mick,
Regarding the vertical stabilizer, one can add Air France 447 to the list and possible Air India 171. For the latter the VS broke off and fell to the ground. Images of what might appear to be damage on the leading edge I think are just due to foreground debris.
@ Mick Gilbert.
Excellent research on the jet ditching, much appreciated for compiling those details.
Number 10, the Ethiopian Airlines Flight 961 (Boeing 767-200ER), ditching, Grande Comore, Comoros Islands, 23 November 1996 is most similar to what I envision from an MH370 viewpoint, ‘from just after initial impact’.
Although the initial impact and deceleration for MH370 would have been wings level and with a very low sink rate (the single most important ‘initial contact parameter’ after wings level) the reality is that in those seas, within a second or two, a wing would have “dug-in”, and produced severe yaw and/or roll, resulting in structural damage similar to Flt-961.
On the other hand, number 13, US Airways Flight 1549 (Airbus A320‐214), ditching, Hudson River, New Jersey, USA, 15 January 2009 was wings level on impacting flat water (so it did not yaw or roll) but it impacted with a higher than desirable sink rate.
Due to various factors, the crew could not, and did not, “hold off” in the last few seconds to minimize the sink rate. (Appendix D of NTSB/AAR-10/03 PB2010-910403 is worth reading). In any case, the higher than recommended sink rate meant that the force of the impact twisted the airframe, the cargo door seals began to leak, there was damage at the tail, and it began to settle into the water tail low.
The ‘water run’ (as opposed to ground run) was not recorded by the FDR, but depending on which videos you watch, it appears that it took about 5 seconds from impact to stop. Since it alighted on the calm water at approximately 125 knots or approximately 64m/s, if it stopped in 4 seconds, the average deceleration was 16.0m/s^2, so the ‘water run’ would be about 128m. If it stopped in 5 seconds, the average deceleration was 12.8m/s^2, so the ‘water run’ would be about 160m. If it stopped in 6 seconds, the average deceleration was 10.7m/s^2, so the ‘water run’ would be about 191m. These deceleration rates are obviously large, about half that of landing (trapping) on an aircraft carrier.
Airliner Vref landing speeds are based on normal landing criteria, where you will be descending at about 500 to 600 feet per minute coming over the threshold, and beginning the flare, with an intention to firmly touch down only a few seconds later at anything form a few – to say 50 feet per minute vertical speed, as you ‘retard’.
That is an entirely inappropriate set up for a ditching. I emphasize, that when considering ditching, it is the final few seconds that count, and the single most important ‘initial contact parameter’ (after wings level of course) is achieving ‘the lowest possible sink rate at impact’.
It’s all about ‘energy management’. You must have ‘a reserve’ of ‘speed to bleed’, because you need both the time, and the energy, to ‘stabilize (level) in ground effect’, as low as possible, to be actually able to achieve ‘the lowest possible sink rate at impact’.
Therefore, for an airline pilot, as you enter the flare regime, you ‘absolutely must be carrying some excess speed’ above whatever your Vref speed is for your actual flap configuration.
How much extra you carry is up to you, it depends on whatever your aircraft’s ‘power off bleed rate’ is, in ‘that’ configuration (in knots / second), but I would say that to a first approximation, you need enough excess speed to be able to safely and comfortably ‘hold off’ for about 5 seconds, to be able to stabilize ‘in the float’ and then ‘fly her on’ – ‘nice and gentle’ like.
Now, turning to question of the recovered debris.
Frankly, I don’t see any significant difference between the characteristics of the MH370 recovered debris and the debris recovered from Ethiopian Airlines Flight 961. I expect MH370 to have yawed post impact in a similar fashion to 961, with the fuselage breaking into perhaps three or four major sections due to side load forces. Remember, in most cases, it is not the circular structural frames that deform and break, they are very strong and stable structures, it is the individual longerons and stringers that are the weaker structural members, they fail first, and then ‘unzip’ between the circular frames.
As for trailing edge devices, I actually agree with you, but then again, as I explained to DennisW in this forum long ago, I have a different view of where the prime item (and its mate) came from.
@ Peter Norton
Firstly, thank you for obtaining the sea state data, that is interesting.
Secondly, the term ‘seaplane’ is used far too loosely, particularly in American publications.
There is a distinct difference between what I would call ‘float planes’, (that have an aircraft structural configuration but with separate floats instead of conventional landing gear and are generally small and light), and what I would call ‘real sea planes’, that are in fact ‘flying boats’, which have a deep fuselage with integral boat hull and are generally large and relatively heavy.
I am not interested in ‘dinky toys’ like DHC-2’s, DHC-3’s, DHC-6’s, or anything similar. I fail to see the relevance of discussing them or the techniques of operating them. I only consider genuine boat hulled ‘FLYING BOATS’, like (for example) the old WW2 era Short Sunderland, Mariner, Marlin, the Mars, the various Dornier’s etc, and of course the venerable PBY. We can also add in the more modern (mid size) Grumman Albatross, the Japanese US-2, the Chinese AG600, and of course numerous Russian ‘cold war era flying boats’.
The point is, they all do have ‘open ocean’ capabilities, and there are of course standardized operating limits for each type, crews were trained to operate them, and did, and still do.
The principles of open ocean operations of large aircraft (flying boats) are what is at issue here.
Sea states are an issue and there are techniques for dealing with them, and in fact, long ago, I remember reading RAF advice on how to ditch all the WW2 bombers, Lancaster’s, Wellingtons etc. Back in the day, the possibility of having to ditch in an emergency was a very real one, post war in the age of the four engine piston airliners. It was a know risk, procedures were developed, and crews were trained. As I remember, even a Boeing Stratocruiser was successfully ditched in the middle of the Pacific Ocean.
You and many others seem to have been so seduced by the safety and reliability of modern jet airliners that you dismiss the relevance of the possibly of a deliberately planned ditching of MH370 almost as an absurdity.
Further, you ASSERT:
“Special training is required for a seaplane, which the pilots of MH370 did not have”.
Seriously, even when it is a well known and documented fact (with a picture way back) that among Z’s favorite toys, was a radio controlled scale model PBY that he loved flying.
I suppose you would say: ‘Totally irrelevant your honor, just a toy, a micro dinky toy at that”.
@Peter Norton
G’day Peter, thanks for that.
The notion that ALM Flight 980 decelerated from around 90 knots to zero in less than the length of the aircraft (a McDonnell Douglas DC-9-33CF is only 36.4 metres long) is not addressed in the official NTSB report (NTSB-AAR-71-8). Back of the napkin, that sort of deceleration would give rise to 4G – 6G forces, which I would think might rate a mention in terms of the possible impact on passenger survivability.
The NTSB report does note the following:
“There is no doubt that the ditching was accomplished under extremely adverse conditions. The captain demonstrated exceptional airman-ship in the control of the aircraft, …” (p.11)
Again, if the touchdown attitude and subsequent deceleration were other than “good”, I would not expect that acknowledgement of the Captain’s handling of the ditching.
Finally, I note that the paragraph in the Wikipedia article that mentions that deceleration is annotated “[citation needed]. I would treat it sceptically.
@sk999
G’day Steve,
Yes, while technically neither AF447 nor AI171 were ditching/CFIW events, they both evidence the relative robustness of the vertical stabiliser, and the fact that its position tends to keep it out of the impact zone for those sorts of accidents.
I made the point quite a few years back, that wreckage item 22 – the torn out section of a panel from the right side of the vertical stabilizer found at Linga Linga beach, Mozambique – was probably the most instructive in terms of determining the nature of the end-of-flight sequence.
That item suggests significant damage having been caused to the vertical stabilizer. And damage to the vertical stabilizer is entirely inconsistent with a ditching style sequence, even one that goes horribly wrong such as Ethiopian Airlines flight 961.
@Mick Gilbert: many thanks for your expert analysis
@ventus45:
In your theory, what plans did the pilot have for his own life, given that a successful ditching would be counterproductive to a quick and painless end ? What was his endgame ?
But back to the technical side of it. I appreciate the fervent defense of your ditching theory. Your Boeing 377 Stratocruiser broke apart despite ditching on a very flat water surface (as opposed to the 13 ft waves at arc7): “A wing hit a swell, spinning the airplane to the left. The tail broke off..
I’ll defer to others. I would love to hear from someone with more expertise about the feasibility of landing a B777 in 4m/13ft waves.
@Mick Gilbert and all other fellow researchers:
To get a better sense of the type of debris field we might expect for MH370, I am collecting sonar imagery of other aircraft crash sites here:
https://aviation.stackexchange.com/a/114919
So far, I was able to locate sonar images for AirFrance 447, EgyptAir 990, SilkAir 185, and the Rockwell 1121A Jet Commander mentioned above.
I’d be grateful for any additional image you can contribute.
For a comparison of the MH370 debris from the vertical stabilizer, a similar piece from MH17 (9M-MRD), along with the main part of the stabilizer to see where it came from. It is always best to compare the debris washing up on the western side of the SIO to MH17 where possible, most of it would have been detached the same way, while in flight.
https://www.dropbox.com/scl/fi/kjkr58tt6sbhejcxftou5/MRD-tail-bit.jpg?rlkey=wr7q49j4j4y5lrp5wo0jonfra&st=n28sth9v&dl=0
https://www.dropbox.com/scl/fi/so3w97k2cmdab92snb5sk/9M-MRD-RS-tail.jpg?rlkey=cqctz9gznrm8wj62q30xl7a8v&st=nw3gbomz&dl=0
Apologies if the links are not working as intended.
Air India 787-8:”According to reports, the crew encountered an issue during ground engine start. The left fuel control switch failed to lock securely in the RUN position on two separate attempts. It then shifted toward the CUTOFF position when slight pressure was applied.”
https://aviationsourcenews.com/air-india-grounds-787-8-dreamliner-after-fuel-switch-anomaly/
Holy Moly.
To get a better understanding of the location, nature, and extent of the missing or low-quality sonar data from the ATSB-led search in the area 35-36S (where their concentration is relatively high compared to other areas), I combined the following three maps from various sources:
1. The map posted in this blog under the article “New MH370 Search Continues” which will be referred to as the V-map, where missing and low-quality data are jointly marked as dark green areas outlined in white
2. The map obtained as a screenshot from a Geoscience Australia presentation available on YouTube, contains information about the nature of the missing and low-quality data. This map is referred to as the G-map
3. Map obtained as a screenshot from the CAPTION Armada tracking webpage with the “GEBCO overlay 2” option enabled, referred to as the C-map
Original, unaltered image maps from the mentioned sources are uploaded here:
https://drive.google.com/file/d/137PmrjIQOqzd33HNIHfTmgUZGgtGXM39/view?usp=sharing
The first image is composed of the map published in the article from this blog (V-map) and a screenshot from Geoscience Australia (G-map). The section of the G-map located around 35–36°S was cropped, rotated, made partially transparent, and aligned with the V-map. Additionally, I used geographic latitude to create the distance scale, with an estimated error margin of approximately 10%.
The information that can be obtained from this map combination includes the approximate size and type of areas with missing or low-quality data, namely “data gaps” and “LPDs.” By using the distance scale and different colorations (blue for data gaps – shadow zone type and green for LPDs), it was possible to identify at least 14 locations, each measuring 1km or more. These locations are marked with arrows and labeled H1 through H14. The small black squares next to the arrows represent areas of 2km by 2km, making comparison easier. There is the possibility of identifying additional similarly sized sites, just the most obvious were marked. The map is uploaded here:
https://drive.google.com/file/d/1QyBsmSXkFZ-r1QGnP1zVmklM8xU_Eu5N/view?usp=sharing
The second map combines the V-map with the C-map. The purpose of this map is to emphasize the topology of the sea bottom in the area and how it relates to data gaps and LPDs. From this map it can be seen that the vast majority of these areas are located within the “valleys.” There are two long valleys stretching in the north-south direction interconnected by several smaller ones. The map is uploaded here:
https://drive.google.com/file/d/1MsppYTfIx59KU3ePgyVcv5Ari5scvL5T/view?usp=sharing
Finally, the third map results from combining the G-map and the C-map. This map gives a more general overview of the analyzed area and the adjacent areas. This map also confirms the previous one regarding the location of missing and low-quality sonar data – they are situated inside the valleys. The map is uploaded here:
https://drive.google.com/file/d/1JW94JjZP84QmJm7aq-WQmQVb5GiAcRMg/view?usp=sharing
The question that remains is:
Why does OI not adhere to the 200m x 200m criterion as the maximum allowable data gap which was used by ATSB for the area between 36°S and 39°S? Why does the same criterion not apply to the search zone north of 36°S, despite numerous scientific evidence – including ocean drift analysis, fuel calculations, and BTO and BFO data – indicating that the plane ended its last journey there, close to the seventh arc?
@Paul Smithson
RE: “The left fuel control switch failed to lock securely in the RUN position on two separate attempts.”
You’d have to wonder why they continued, instead of returning to the gate and having it checked by the engineers.
Air India B787 (cont.).
From BBC today ‘The Directorate General of Civil Aviation (DGCA) said on Tuesday that when the switches were operated according to Boeing’s recommended procedure, they were found to be “satisfactory”, staying at run instead of moving to cut-off (which can stop fuel supply to the engines).
It added, however, that incorrectly handling the switch caused it to “move easily from run to cut-off”.’
Implication: fuel cut-off malfunction vulnerable to human factors. Which sounds like a very critical design issue!
@all
The page of Armada 86 05 has been updated…
https://www.mh370-caption.net/index.php/armada-tracking/
…as well as the page providing the GoogleEarth’s kml files of both Armada 78 06 and 86 05
https://www.mh370-caption.net/index.php/2025-armada-78-06-paths/
@Peter Norton
We have improved the maps fetching code, but please note that the ATSB maps are provided by the Geoscience Australia server which appears to be very slow.
@PaulS
One thing AI171 teaches is the need for cockpit video, if for no other reason to be able to “show” the doubting public, and protect Boeing from false accusations. I of course believe deterrent is the more important justification to help prevent. (“Show” in quotes because the recordings are confidential.)
In the AI171 media coverage it was said Boeing does apparently offer cockpit video capability and that some countries may already require it. I would think USA/FAA/Boeing should recommend cockpit video, encourage it, but if necessary force the customer to reject it. Although USA itself had a recent incident (jump-seat pilot fuel cut off), it could possibly be true that the USA has the rogue pilot situation more under control, due to procedural and cultural reasons etc. So let USA be safety laggards, as usually seems to be the case now.
@Jean-Luc Marchand:
It seems to work now. Many thanks for the update!
@Kenyon:
I assume, you are the author of:
– Kenyon et al., Ground truthing 6.5-kHz sidescan sonographs: what are we really imaging?, J. Geophys. Res., 96, 5955-5974, 1991.
– Kenyon, N.H., Speculations on the Geological Causes of Backscatter Variation on GLORIA Sonographs from the Mississippi and De Soto Fans, Gulf of Mexico, Geo-Mar. Lett., 12, 24-32, 1992.
?
@Paul Smithson
The DGCA’s statement in full:
Rejoinder on News Item relating to purported malfunction of Fuel Cut Off Switch on M/s Air India Boeing B787-8 aircraft VT-ANX
The fuel control switch locking feature has been in use for over 40 years on many other Boeing types.
@ventus45
G’day ventus,
Ethiopian Airlines flight 961 is the reason I eschew the “high-speed/high energy v lower speed/lower energy” approach to categorising impacts, particularly with regards to damage and liberated wreckage. In terms of ditching/CFIW events, ET961 is essentially the exception that proves the rule.
In Review of Transport Aircraft Ditching Accidents, Lindenau and Rung describe the ditching/CFIW event as a transition from aerodynamic to hydrodynamic to hydrostatic forces. They note, uncontroversially, that it is the application of hydrodynamic forces, to structures not specifically designed and constructed with those forces in mind, where damage and survivability are determined.
Lindenau and Rung also discuss some of the controllable factors that obtain to the transition from aerodynamic to hydrodynamic forces; airspeed, attitude (both pitch and roll), and rate of descent.
Parenthetically, it should be noted that controllable thrust plays a critical role in establishing and maintaining all three of those factors (in the case of attitude, pitch not roll), as does flap configuration. Given the absolutely critical-to-outcome nature of controllable thrust in particular, it is difficult to conceive of a circumstance where that factor would be deliberately eliminated by flying to fuel exhaustion.
Going back to Lindenau and Rung, while they only examine one CFIW event – Trans Arabian Air Transport Boeing 707 ST-APY, Lake Victoria, Tanzania (my ref. 11 above) – they do note that the generally highly successful outcome was almost certainly attributable to the aircraft being configured and flown for landing. We see similar outcomes for the CFIW of Lion Air Flight JT904 into the Bali Sea west of Denpasar, Indonesia, and Air Niugini flight PX73 into Chuuk Lagoon, Micronesia (my refs. 14 and 15).
Returning to comparing ET961 to the MH370 wreckage, there are a couple of inconsistencies, not the least of which being what must have been very significant damage to the vertical stabiliser in MH370’s case, something that has never been seen across any of the recorded ditching/CFIW events.
When thinking about the dynamics of a ditching event, there will be a finite amount of energy that can be translated to the forces being applied to aircraft structures for any given case. That energy is dissipated over the duration of touch down/alighting phase. Where and when that energy is translated into impact forces will be largely driven by the aircraft’s attitude, but also by the basic structure of the aircraft. The underslung, wide diameter engines, dihedral, and wing sweep clearly play a role.
There will be a requisite amount of force needed to liberate any particular component. Thus, we can think of how the total energy budget might be expended, in terms of where on the aircraft, and when during the touch down/alighting phase, bearing in mind that energy will be dissipating over the course of the touch down/alighting phase.
For MH370, we have three adjacent trailing edge devices from one wing; right flaperon, right outboard flap, and right aileron. Given the wing’s dihedral and sweep, it is challenging to determine an orientation that directs hydrodynamic forces sufficient to liberate each item onto all three. It is worth noting here that there has been no ditching/CFIW event where multiple adjacent trailing edge devices have been liberated.
Further complicating this is the need to also orient the aircraft such that components from the left wing – the outboard flap and flap-track fairing – can be exposed to hydrodynamic forces sufficient to liberate each of them. It is worth noting here that I could find no evidence of underwing fairings (flap-track/undercarriage fairings) ever separating in a ditching/CFIW event, and with MH370 we have flap-track fairing wreckage from both wings.
And you also need to find an orientation that allows hydrodynamic forces to be applied to the vertical stabiliser, sufficient to tear a section of panelling out of it. As repeatedly noted, there is no ditching/CFIW event precedent for this.
I would contend that it is essentially impossible to reconcile very significant damage to the vertical stabiliser with a ditching that also liberated all the other wreckage items.
@ Mick Gilbert
I agree with most of that, but not all.
We have to remember that this ditching was in a churning sea. Contrary to your position, it is my view that the vertical stabilizer damage (being apparently unique to MH370) actually supports a ditching, for the following reason.
Assume a successful deliberate ditching occurred, where large segments initially floated (perhaps for some minutes).
In that time period immediately after ditching, parts liberated in the impact would be thrown around by the churning waves, and very likely impact other parts of the (still floating) aircraft, specifically the empennage section, and/or, if the fin (vertical stabilizer) had separated whole (like AF447) it would be highly likely that it would be impacted by other parts, and damaged, thus releasing the recovered part.
Similarly, the wings, if one or the other broke away from the center section, that would certainly liberate wing/fuse fairings, and, since the wings are empty fuel tanks, (if the wing box itself is not severely ruptured) they are likely to float for some time, and being large and heavy, would severely impact and damage any remaining sections of the fuselage that are still floating, puncturing the shell and liberating internal parts, as well as those impacts also breaking off TE parts.
Thus, I still hold to a ditching, because MH370 is in many ways different to the other ditching’s. Apples with apples, etc.
@ventus45, @Mick,
For some time I have quietly wondered the following:
If a wing devoid of fuel separated from the aircraft “proper” just how long might it have taken to reach the ocean floor ?. If the sinking or floating wing was also devoid of its engine, would it take more time.
Specific reason for the wondering: Would this separated wing be easily seen by sonar, after 12 years ?. I would expect so. ?
@George G. I don’t think it would have been possible for a “whole wing” to separate intact. Under high G loads, it would shatter… like glass. (https://www.reddit.com/r/Damnthatsinteresting/comments/wy7mwo/in_1995_boeing_tested_a_777_wing_to_destruction/)
Armada 7805 is heading to Pago Pago. Caption’s Armada tracking notes that Armada 7804 is scheduled to arrive in Mauritius on February 27 2026.
@airlandseaman,
Mike, Thank you.
@airlandseaman. And before that happened the engines would be ripped off. The engine mounts are designed to shear off, presumably with minimal damage to the wing.
@ALSM.Had there been control surfaces and wings attached I envisage that they, closing panels and fairings might have been shed and, as they seem to have been, riven.
Engines and undercarriage?
Shock could also affect the other wing.
Wide dispersion.
Perhaps hitting the water in a steep dive at extreme speed, like the Silk Air Flight 185 737, would effect much the same, albeit pitch control in that required a pilot.
There are no beyond-the-envelope flight test data for either aircraft’s simulators, the ATSB making this clear re the Boeing 777 simulations.
Boeing did an analysis to help fill that gap for the 737 though not the 777.
In the 737 instance there was empennage control surface shedding from flutter. None such has been identified amongst MH370’s recovered items. Though hardly conclusive, perhaps pitch beneath flutter speed might be suggested.
@Paul Smithson
The plot thickens:
Reuters: UK asks Air India to explain Boeing Dreamliner fuel-switch incident
Not surprising, given they departed with a potentially faulty fuel control switch.
@George G
G’day George, there is only one precedent for a wing becoming completely detaching from the fuselage consequent to a ditching.
Japan Air Lines Flight JL350, was a McDonnell Douglas DC-8-61 where the Captain (suffering from a mental condition, then undiagnosed) attempted to deliberately fly into Tokyo Bay on approach to Tokyo-Haneda Airport, Japan back on 9 February 1982. In that incident, the starboard wing suffered severe damage just outboard of the wing root sufficient to separate the wing from aircraft. The post event photos show the starboard wing still adjacent to the wing root but rotated forward (perhaps answering the seldom asked question, what would a DC-8 look like with a straight wing?).
It is worth noting that despite that damage, JL350 remains broadly consistent with the general theme for limited component separations. There were no trailing edge device separations, no s3d9jdiqt6
@George G
A salutary lesson in trying to post from a phone with fat fingers.
Continuing from above,
There were no trailing edge device separations, no secondary device separations, and no damage to the vertical stabiliser.
Andrew: There was no faulty fuel control switch on AI171.
@Julie
I was referring to the AI132 incident in London on 1 February, not AI171. The AI132 pilot-in-command reported that the “Left fuel control switch slips from RUN to CUTOFF when pushed down slightly. It does not lock in its position.” According to reports, that occurred on two start attempts, and on the third attempt the switch stayed in RUN. It appears the issue wasn’t investigated further until after the aircraft arrived in Bangalore, where it was grounded. The UK Civil Aviation Authority has asked Air India for a “please explain”.
@Mick Gilbert,
Phone and fingers, and thumb hitting things you didn’t know the phone had make me rarely attempt to send long E-Mails or similar from the device.
Your “no trailing edge device separations, no secondary device separations, and no damage to the vertical stabiliser” led me to again review the listing in the Malaysian SIR (Table 2.6A) and the associated Figure 2.6A.
Item 1:Right Flaperon,
Item 10: Left Outboard Aft Flap Section,
Item 19: Right Outboard Flap,
All confirmed from 9M-MRO. … chewing lip motion ?
@ventus45
And yet MH17 had almost the exact same piece from the vertical stabilizer detach from it, MH17 was not a ditching, and the damage occurred before MH17 crashed, the leading edge and the piece that came from where the MH370 debris came from ended up no where near the horizontal stabilizer. It would have occurred in both MH17 and MH370 in almost exactly the same way, nothing to do with a ditching, nothing to do with a crash, not even anything to do with what happened after MH370 ran out of fuel.
I know which flight you were referring to, Andrew. I think they are desperately trying to protect a fellow pilot who committed murder/suicide.
@ Annette Mansfield
When an aircraft breaks up, whether it be in the air, crashing on the ground during a landing gone wrong, or breaking up on hitting the sea in a ditching, makes very little if any difference so far as impact damage to the empennage from debris liberated from the front parts of the aircraft are concerned.
In all cases the empennage is still moving forward with the contiguous mass of the aircraft at the impact velocity minus a few knows, whilst the liberated items of debris are decelerating rapidly, and thus moving back towards the tail, even though everything is still going forward.
These liberated debris items will impact the empennage area at shallow angles (more like a side swipe than a perpendicular impact) and the relative impact velocity is low.
Therefore, the damage imparted by the debris will be very localized and very slight, perhaps only sufficient to break away a small piece of a frangible panel.
In the case of MH17 that you mention, any such damage occurring at altitude will result in a small light part that will immediately decelerate to near zero forward velocity due to drag and begin to fall down near vertically, quickly reaching it’s low ‘falling leaf’ terminal velocity, with any further lateral velocities and subsequent displacement from the nadir at impact and liberation being solely a function of the atmospheric wind vectors encountered on the way down.
However, the major part of the structure that it came from was still travelling forward at over 400 knots for some time.
It is thus hardly surprising that the two ended up many kilometers apart.
In the case of MH370, ‘No Step’ is an example, a small piece, low mass, frangible, thus easily ‘knocked off’ with relatively little force.
It was also buoyant, floated, drifted, and was found, many thousands of kilometers from the nadir of it’s liberation.
@George G
G’day George,
A bit more stimulus for said chewing of lip, perhaps.
In the course of gathering data on ditching/CFIW events, it quickly became evident that a critical examination of the somewhat common assertion that the damage to the trailing edge of the flaperon was possibly caused by water impacting it in the course of a ditching was necessary.
I think that that possibility was first advanced in the French Direction Générale de l’Armement – techniques aéronautiques (DGA TA) examination report on the flaperon from back in 2015-2016.
In section 10.3. CONCLUSION – SEPARATION AND DAMAGE SCENARIO, of that examination report, the DGA TA “hypothesize [their bolding] a scenario that could have led to the separation and to the damage noted on the flaperon.”
Here it is important to note that by the DGA TA’s own admission, no specific research was conducted into the practicalities of their hypothesised sequence of events, nor apparently was any evidence gathered that might lend weight to or contradict the hypothesis.
Section 10.3 concludes, with the following,
“Beyond the possible in-flight loss, which is not entirely excluded, this scenario can only be considered as a hypothesis …”.
They go on to note that “the flaperon is aft of the engine which leaves some doubt as to its loading during aeroplane impact with the water“.
Had they have undertaken even a cursory review of ditching/CFIW events (there had been 14 at the time, for which there were records or photographs of 12), they would also have noted that the damage/wreckage data did not support their hypothesis. As I have pointed out, trailing edge device separations in such events are not common.
They might also have noted that many trailing edge devices (TEDs), particularly the inboard flaps, on a variety of other commercial passenger jets also use a similar structural design to the B777 flaperon, where a composite (or lightweight) trailing edge component is attached to the rear of the main structural assembly to create the overall aerodynamic shape of the device.
In the majority of the recorded ditching/CFIW events (10 out of 16), flaps were deployed. Flap settings ranged from about 8° (Transair Flight 810, B737-200) to around 50° (Japan Air Lines Flight JL350, DC-8-61). There does not appear to be an instance where hydrodynamic impact forces broke the trailing edge component off the main structure, even when the flaps had been deployed fully.
And, of course, what the DGA TA did not know when they were examining the flaperon was that other significant trailing edge wreckage would soon be in evidence, specifically:
• a piece of left outboard flap (May 2016),
• a piece of right outboard flap (June 2016), and
• a piece of right aileron (December 2016).
The ditching/CFIW data suggests that the likelihood of any one of those trailing edge devices being liberated in the DGA TA’s hypothesised ditching scenario would be low. Having all four TEDs liberated, in particular three adjacent devices from one wing (right flaperon, right outboard flap, and right aileron), has never been seen.
There is an important lesson here, I think. When most people read, or subsequently heard about the DGA TA’s hypothesised ditching scenario, I suspect that they likely quickly formed a mental picture of what was being described. That mental picture would have been congruent with most people’s practical experiences, specifically that water creates a lot of drag forces on flat surfaces. And in a little more than the blink of an eye, in the complete absence of any form of testing, or even historical analysis, hypothesis becomes part of MH370 lore.
The fact that the hypothesis is largely, if not entirely, contradicted by the data does not appear to factor into many people’s view on the topic. It seems that in a shoot out between “conceivable” and “wrong”, “conceivable” gets a considerable leg-up.
For me, it just reinforces why Step 2 in any analysis is GET THE DATA.
@Mick Gilbert,
@airlandseaman,
Perhaps we ought to consider a break-up of a main lifting surface whilst in high speed descent.
If such break-up included break-up outboard of the engine position then release of the flaperon might have been likely.
The major effect regards searching may be the size and nature of the final debris field.
@Geroge G,
Why discount flutter as the reason for the departure of the flaperon?
The damage to the flaperon is entirely consistent with aerodynamic flutter [note: NOT aeroelastic flutter]. The weakest part of the flaperon is along the line of fasteners which secure the flaperon trailing edge in place. This is exactly where the the fracture occurred.
Do not confuse with aeroelastic flutter, which is an entirely different mechanism, but one often shown in video clips as a representation of flutter.
@Brian Anderson, Refer your comment#39981 of February 8, 2026 at 5:48 pm
I am not discounting flutter.
In particular, Refer to the Malaysian SIR Issued on 02 July 2018.
Refer to the photo of Item 10 of Table 1.12A – Items of Debris, on Page 144.
“This part is confirmed from MH370”
On Page 144 it is listed as “Left Outboard Flap”.
Refer to Page 392 of the same report.
Here, in Table 2.6A, Item 10 is listed as “Left Outboard Aft Flap Section”.
Refer to the photo of Item 19 of Table 1.12A – Items of Debris, on Page 149.
“The part is confirmed from MH370”
On Page 149 it is listed as “Right Outboard Flap”.
Item 10 as shown on Page 144, apparently only a section from the Left Outboard Flap may be similar to what is missing from the outer aft end of the Right Outboard Flap, Item 19.
I am not discounting flutter.
On the 4th February I wrote: “For some time I have quietly wondered …” and continuing on, I wrote on 8th February: “The major effect regards searching may be the size and nature of the final debris field.”
If aerodynamic forces got energetic and repetitive enough to trigger flight control surface break-up … then this might explain Mick Gilbert’s basic question. Given time, if time still available, it might also have led to bigger flight surfaces breaking up. The size and nature of the final debris field would be affected.
I may be wrong , but ?
An other mystery?
As you know, when their transponder is functioning normally, all aircraft can be seen on the website: https://www.flightradar24.com.
For ships, according to the International Convention for the Safety of Life at Sea (SOLAS), there is an obligation to have an AIS transponder, in other words, an Automatic Identification System, and it is illegal to disable it.
When their AIS transponder is functioning normally, all vessels can be seen on the website: https://www.marinetraffic.com.
As planned, the Ocean Infinity Company’s ARMADA 86 05 has resumed its search, in accordance with the contract signed with the Malaysian government: $70 million, payable only if OI finds the wreck.
The 3 January weeks searches were unsuccessful, but OI made no communication.
ARMADA 86 05 returned to Perth to resupply the ship and its crew, exhausted after three weeks at sea, tossed about by the waves, especially with a slow-moving vessel.
But then, surprise! Without any fanfare, ARMADA 86 05 left Perth… supposedly for the US Samoa Islands and switched off its AIS transponder.
https://www.marinetraffic.com/en/ais/details/ships/shipid:9754405
While this topic fascinates many people and medias, no one seems to have noticed this.
Any idea?
EX CAPTIO member
@Jean-Marc: Welcome to the blog.
As Malaysia was providing regular updates, something has changed coinciding with a decision to not continue the search in February. We can only speculate as to what changed.
Regarding the lack of AIS data, I wonder if Armada 86-05 is operating its transmitter in a low-power mode which still provides collision avoidance protection but makes it difficult for satellites to receive the VHF signal and for marine tracking sites to collect the data.
But why would Armada 86-05 elect to do this?
Why OI did not communicate?
@Jean-Marc: What’s your theory?
@Brian Anderson. You ask, “Why discount flutter as the reason for the departure of the flaperon?”
Even should the trailing edge have separated due to aerodynamic flutter at the whole’s natural divergent frequency, what would account for the separation of the rest when its aerodynamics, mass and mass distribution change will have led to quite a different flutter frequency?
Also can you be confident that it would not be a hinge that failed?
The outboard hinge might be doubly loaded were the inboard flaperon actuator in by-pass, as seems likely. Yet judging by the acute angle between the actuators and its leading edge at failure, the inboard end of the flaperon led the outboard aft at separation.
There was damage to its top inboard edge. Conceivably it had been struck downwards by the inboard flap or its mechanism, that strike possible accounting for the trailing edge separation.
Then there is torsion damage. Its leading edge outer end received a crushing swinging uppercut after separation of its actuator, the torque resulting in several shear fractures along its leading edge, deformation of its upper skin and a long crack.
While post outboard actuator separation, it may be related.
The source with all of the robustness, shape and proximity to do that is not obvious.
@Brian. Please add, “amplitude” before “divergent”, second line.
@Brian. As to just the trailing edge separating in flutter, there may be evidence of that. The adjacent outer flap’s seal pan has damage that the ATSB has attributed to a flaperon strike at the time both flaperon and that part of the flap separated from the wing, both then having been housed.
I have queried that, observing that had that been the case there also would have been damage to the flap’s seal pan farther forward, from an extension of the flaperon forward spar.
A separation of just the trailing edge, both housed would explain it though that would have important implication. The ATSB’s conclusion that flaps probably were housed, the flaperon in its neutral position when they separated at aircraft break-up, has been the basis for many concluding that there was no controlled ditching.
From the ATSB’s account that would leave the evidence still supporting its conclusion to the evident internal flap damage from its support track. However, conceivably that could have been resulted from the collapse of the flap deployment mechanism during a ditching.
Hence the ATSB conclusion that there was no flap deployment could be in question and with it, that as evidence that there was no ditching.
Another possible matter needs raising in case it should become relevant. I understand that the flap inboard end’s seal pan was made of fibreglass to protect the flaperon since were there a flap skew the flaperon could be hit. The outer flap has two deployment mechanisms and apparently it is possible that were could be a failure between them of the torsion tube driving both. The flap then could skew, even though there is a skew detector.
Most likely it did skew in the sense that both mechanisms would not fail at once, ditching or not. Possibly that led to the outer flap’s trailing edge separation that is evident, even the flap breakage at its outboard operating mechanism.
However so far as I can that is not pertinent here.
@Jean-Marc: What’s your theory?
Victor , I have been a passive reader of this blog , even before you took over.
So far , I didn’t dare to intervene in this blog because my technical knowledge is not at the level of the participants. Nevertheless , I learned a lot.
I admire your “motto” which is to go deep and to stick to the technical issues.
As it was during my seven years parenthesis with CAPIO , under the (britton like) stubborn Jean-Luc Marchand , who thinks he is the only one to be rigth.
If I may, as far as the OI silence is concerned and as for the trajectory, the technical answer is not sufficient.
If I understood , correctly, the Inmarsat BFO and BTO study;
BFO = ΔFup + ΔFdown + δfcomp + δfsat + δfAFC + δfbias
Of the six components of the BFO, two depend on the aircraft’s position and velocity.
ΔFup depends on the x, y, z, vx, vy, vz coordinates of the aircraft and the satellite, which is also moving.
δfAFC depends on the x, y, z, vx, vy, vz coordinates of the aircraft and a fixed, fictitious satellite. Anyone can use an enhanced version of a spreadsheet (SK999-Satellite Model – Yap Fook Fah) developed by Yap Fook Fah, Associate Professor of Engineering at Nanyang Technological University, Singapore.
This spreadsheet calculates the BTO and BFO by inputting the six components of an aircraft’s position and velocity at a given time (arc transit time). If the calculated BFO value differs by less than 7 Hz from that measured by INMARSAT (uncertainty resulting from INMARSAT analysis of recordings from other flights), the aircraft’s position and speed are considered consistent with the recorded data.
Therefore, BTOs and BFOs do not allow for the calculation of the aircraft’s position and speed, but they do allow for determining whether a trajectory based on operational and technical considerations is plausible.
However, the margins of uncertainty in the BFO values and the multitude of possible heading-speed combinations mean that the hypothesis adopted by the Australian Transport Safety Board (ATSB) and Inmarsat experts to guide the search is not the only possible one.
Moreover, as stated in the Inmarsat report, there are thousands of possible plausible trajectories.
Amoung all these tarjectories, I believe that the motivation should be factored in, on top of operationnal considerations but in discarding conspiration theories with no link whatsover with facts and data.
From what I read on the very respected Zaharie Ahmad Shah, I think it to simple to explain the direct far south trajectory by him being a mass murder with a very sophisticated (never ever occured in air transport) suicide.
This my humble answer to your question.
Jean-Marc
By the way, the debris and , to start with , the flaperon retro current studies and bernacles analysis are not precise enough to ” go back” to the wreck
Jean-Marc Garot wrote ‘But then, surprise! Without any fanfare, ARMADA 86 05 left Perth… supposedly for the US Samoa Islands and switched off its AIS transponder.‘
Satellite relay of vessels’ AIS transmissions is not perfect and short gaps of, typically, less than 2hrs may be evident. At this time there is a process of consolidation underway in the distribution of satellite relayed AIS data. I have access to AIS data that are being distributed via a provider’s ‘legacy’ API and via a ‘new’ API. Neither the feeds via the ‘legacy’ nor ‘new’ API show any evidence of Armada 86 05 ceasing broadcasts from its AIS transponder.
Subsequent to its departure from the Gage Roads anchorage at 2026-02-04T0400Z the position log from the ‘legacy’ API records 1275 discrete continuous positions, while the ‘new’ API records 9271 discrete continuous positions. No appreciable gap, I don’t perceive any new mystery.
@Jean-Marc Garot: When I asked about your theory, I was only referring to OI’s and Malaysia’s silence on the status of the search.
I apologize
https://www.marinetraffic.com/en/ais/details/ships/shipid:9754405
I go back to passive reader
@Jean-Marc Garot
Thank you for posting. I use DrB’s BFO/BTO spreadsheet as I felt I had found some small discontinuities around Arc2 in the YapFF version.
I submit the pilot flew south and then merged onto the home sim path, whereas the home sim path was probably 180S CMH, which is similar but slightly different than the assumed LNAV trajectory to NZPG, instead heading towards the Magnetic South Pole, with enough fuel loaded (in the sim case) to get all the way there, I seem to find.
I would like to try to convince Jean Luc/CAPTIO/CAPTION that this is the evidence if we can adopt the planned/deliberate case, which in that regard, I am philosophically aligned with Jean Luc and yourself. But so far I am not in agreement with CAPTIO/CAPTION flight path assumptions originally to Xmas and now to ~35s region.
@VictorI. Below is my third foray into aspects we might have missed or misconstrued, in case the current search fails. It is about the final descent.
What currently many of us accept doesn’t seem to fit.
https://docs.google.com/document/d/1WHV2DzaeWg-GULSYph5JWIWAIWxpf5ET/edit?usp=drive_link&ouid=105452605762640315384&rtpof=true&sd=true
@David F: Thank you for that summary. In the past, others have also recommended paths that allow a descent in daylight.
At this point, my recommendation would be to finish the search that was originally proposed before jumping to other areas. Without a methodical and thorough search strategy, it becomes difficult to dismiss 7th arc latitude crossings as “searched with nothing found”.
@VictorI. Understood.
On “methodical and thorough” it would be interesting to know what dispersion OI now supposes, whether that supposed in earlier searches met this measure and, depending on how it sees the extent of its potential futher and future searching, what sigmas’ coverage is needed to meet the same.
Plus of course where spot searching fits in as a priority and Malaysian views…..
Tough one.
Their new obsessions can widely vary. Some start taking pictures of sunsets; others like sunrises. They could fixate on certain people. It could be anything. One woman made sure her trash cans were rigidly aligned, hosed out after each pick-up, and railed against neighbors who didn’t roll them back immediately after sanitation pick-up (her neighbors were working and not at home). Others may collect bottles of shampoo. One man started spending all day at the movies. Another started collecting cases and cases of red wine, although he had no previous interest in wine. They may also start spending wildly, and be inappropriate toward women and also in public. They may lose their sense of humor, become hyper critical of others, easily irritated, agitated, and unpleasant to be around.
One thing that stands out is a change in personality. They may become restless, be constantly on the move and stay out all night. In contrast, they may retreat to a room. It is imperative that Captain Zaharie Shah’s family be asked if he changed beginning 2 to 4 years before MH 370 went down.
@Julie
Thank you for sharing your machination. I don’t wish to sound flippant, but there seems to be numerous hypotheses in your document that controvert objective observations subjected to scrutiny on this site and others over the past decade plus. How closely have you been following the technical discussions which set boundaries between what is possible and what is fantasy?
@Marijan
Regarding your Febr. 2nd post with the interesting maps you constructed:
If I estimated correctly the data gaps (H/L) indicated by you amount to about 3-4% of the larger area they’re located in. Which is a bit less than indicate by ATSB; perhaps because you left out some of the smaller gaps (< 1km)?
Would you know how the 3-4% compares to a similar analysis in the larger area indicated by "OI fill in Feb 2025"?
@Niels
Yes, you are right, ATSB (or more accurately, Geoscience Australia) calculated the total area covered in data gaps/LPDs using a 100 m x 100 m metric. This means that every area of that size or larger was marked on the map and included in the coverage statistics. Instead of marking every spot of that size or larger, I basically handpicked the largest data gaps/LPDs and marked them on the map to illustrate that there are still areas that may contain not only the main debris field but also secondary debris field, if any (check for example EgyptAir 990 from the list Peter Norton created, Feb 1st post). The method used to estimate the gap/LPD size was rather rudimentary and does not allow for precise measurements, but I am working to improve it soon.
@All,
What are your thoughts on the ADS-B trial that Capt Peter Turner has conducted? He has found that the transponder ‘ALT OFF’ position has no affect on the ADS-B readout, only on the mode C.
@Tim: Exactly what experiment was run? What Capt Peter Turner (notorious for incorrect statements under various aliases) claims is counter to CFR14 § 91.217: Data correspondence between automatically reported pressure altitude data and the pilot’s altitude reference.
(b) No person may operate any automatic pressure altitude reporting equipment associated with a radar beacon transponder or with ADS-B Out equipment unless the pressure altitude reported for ADS-B Out and Mode C/S is derived from the same source for aircraft equipped with both a transponder and ADS-B Out.
@Tim
Both geometric altitude and barometric altitude are broadcast using ADS-B.
The over-the-air transmission from the aircraft transponder is received by a ground processing system and forwarded to the central air traffic processing system in the form of a Surveillance Data Exchange message.
In the specific case of 9M-MRO at 17:20:33/34 2014-03-07 the final two surveillance data exchange messages contained a valid geometric altitude but no barometric altitude information.
The same log of surveillance data recorded messages originating from a MAS B737-8HG on the ground at Kota Bharu at circa 15:40Z showing a valid geo altitude but no barometric alt information (as might be expected for an aircraft on the ground with ALT RPTG OFF selected).
Geometric altitude is sourced from a GNSS receiver, the MMR on a B777, while barometric altitude is sourced from the air data system, the ADIRS on a B777.
If it was needed, Turner provides another example of a pilot not exhibiting expert knowledge of avionics systems.
Don said: Both geometric altitude and barometric altitude are broadcast using ADS-B.
I believe that barometric altitude is required and geometric altitude is optional. The reported barometric altitude from ADS-B has to be the same as the Mode C altitude.
In any event, I would like to see Capt Peter Turner’s experiment. Under various aliases (e.g., Paul Onions, Pete Noetic, Capt Peter Turner), he presents false information, which is why I banned him from this site.
Re: Xponder at IGARI
Jean Luc/CAPTION claims (on “Sit Down with Sid”) to have independently proven the XPonder knob was manually turned. I am not sure technically why Jean Luc feels his new analysis is any stronger than @sk999’s “smoking gun” essay which goes back quite a few years.
In any case I assume the “remarkable accident” believers can simply revert to pilots intentionally cut the Xponder for “valid reasons” to control said remarkable failure.
@Victor,
(The tribulations in attempting to be brief)
Two systems and their different mechanisms for coding the data. The aircraft transponder involves ADS-B/1090MHz ES (extended squitter) broadcast by the aircraft as currently defined by RTCA DO-260B while the ground based surveillance systems receive the broadcasts and re-code the data for distribution according to ASTERIX CAT-021.
The transponder will normally insert barometric data into the Airborne Position squitter while inserting the difference between baro alt and GNSS-HEA (height above ellipsoid) into the Airborne Velocity squitter.
If ALT RPTG OFF is selected the transponder changes the TYPE code of the Airborne Position squitter to indicate insertion of GNSS-HAE altitude. In this case, no difference information is carried in the Airborne Velocity squitter.
The ground based ADS-B receiver reports GNSS height and a barometric derived Flight Level using ASTERIX (if available).
ADS-B ES serves airborne collision avoidance purposes, ground based surface target surveillance, and ground based airborne target surveillance: a broader set of capabilities than reflected in the ASTERIX CAT-021 surveillance data exchange messages.
@Don Thompson: I don’t doubt that if barometric altitude is not available, then the GNSS altitude MAY be substituted with a flag (change in position type code). However, the barometric altitude, which is used by ATC for traffic separation, would not be available, as reflected in the FR24 and ATC ADSB data we have for MH370. I believe that “Capt Peter Turner”, or whatever he is calling himself these days, is wrong to claim that “the transponder ‘ALT OFF’ position has no affect [effect] on the ADS-B readout, only on the mode C”, as Tim is relaying.
John:
I wasn’t engaging in machinations, I was merely sharing a theory.
After 12 years and not finding MH370, why not?
I am also waiting for the final reports of AI171, MU5735, and others. Do you see a common denominator?
Your technical expertise is vast and impressive, but you are missing something: a Pattern.
@Victor, @Don,
Just doing a little of my own research. Typing “ ads-b does altitude data transmit in the alt off position” into google search. AI thinks ads-b altitude will continue to be transmitted in the ‘Alt off’ position.
Therefore, it’s increasingly likely that the evidence for the transponder switch being manipulated is false.
@Tim
Not so fast. This is what a different AI chatbot has to say on the matter:
https://grok.com/c/f4f84597-e218-4d02-851b-f0d403acead1?rid=4da510e1-2a7e-45dc-88a2-a36adaf27bfa
What was this experiment that Turner supposedly ran?
@Tim
This is the response when I used the same words you used with Google:
https://grok.com/c/6dfce280-10ee-433e-92f9-8224f0e56af7?rid=916593bd-2d11-4d7b-a7b1-ebbd899d32c4
@Andrew,
Thanks, perhaps it’s a case of some transponders do and some don’t transmit ads-b altitude in the ‘alt off’ position.
Therefore any valid trial would have to be conducted with the same transponder and set up that mh370 had onboard.
@Tim: You either did not read or did not understand previous comments.
If ALT RPTG is off, there is no barometric altitude reported. Period. It’s possible that in this circumstance, the type code for the data is changed and the geometric altitude is substituted. But it is NEVER presented as barometric altitude, and other aircraft and ground tracking stations would never present and process the geometric altitude as barometric altitude.
You are a pilot, so you understand the safety risk that is presented if the geometric altitude were substituted for barometric altitude in ADS-B out data packets without specifically flagging this substitution.
Tim wrote ‘Just doing a little of my own research.
The term ‘garbage in, garbage out’ applies to internet search engines, and with the onset of ‘AI’ the likelihood for garbage out is even higher (search engine businesses actively encourage users to consume their dubious ‘AI’ responses rather than have those users click through to independently hosted web pages where the same internet search/advertising businesses may be liable to pay the web page publisher for placing their ads).
Transponders function according to ARINC Characteristic 718; ADS-B transmissions conform to Minimum Operational Performance Standards for 1090 MHz ADS-B and TIS-B, RTCA DO260B, 2009; air traffic surveillance networks exchange data in accordance with All Purpose Structured EUROCONTROL Surveillance Information Exchange – ASTERIX.
Concerning the idea that ‘some transponders do and some don’t transmit ads-b altitude in the ‘alt off’ position: aside from my references to standards above: a) the transponder mode select state being discussed is ‘ALT RPTG OFF’ – altitude reporting off; and b), there is no ‘ads-b altitude‘ – ADS-B squitters normally encode barometric altitude and geometric altitude enabling the ground systems to extract both values. As Victor emphasises, barometric altitude is the key parameter – altitude in the air traffic management domain remains barometric.
As I set out above, when ‘ALT RPTG OFF is selected, the transponder changes the TYPE code of the Airborne Position squitter to indicate insertion of GNSS-HAE altitude. In this case, no difference information is carried in the Airborne Velocity squitter.‘ For further clarity, in this state the aircraft transponder is transmitting only geometric altitude.
DCA’s recorded surveillance data exchange traffic shows messages from 9M-MRO carrying both geometric and barometric altitude parameters until the final two messages at 17:20:34.15 & 17:20:34.55 in which the barometric altitude is absent (while valid, GNSS derived, geometric altitude continued to be encoded).
J-LM/CAPTION set out in a recent report that the surveillance data exchange traffic ‘Quality Indicators’ remained unchanged throughout the recording thus barometric altitude continued to be available to the transponder (but not inserted/encoded in the final two squitters).
I have set out here that the observed, recorded, surveillance data exchange traffic is in accordance with the prescribed standards for the operation of the ATC transponder, its control panel, and DO-260B and the circumstance where the mode select switch is rotated through the ALT RPTG OFF detent to STANDBY.
If it’s possible to extract Turner’s treatment of the topic from whichever rabbit hole it has been posted, someone please relay an unrestricted reference.
@Julie
You wrote:- “It is imperative that Captain Zaharie Shah’s family be asked if he changed beginning 2 to 4 years before MH 370 went down.”
I think that rather than “vanilla” dementia there may have been progressive neurological damage initiated by some kind of physical injury.
My best bet would be a head injury resulting from his severe hang gliding accident.
@Julie: There was an article in The Australian written by Hedley Thomas in September 2016 about the close relationship that Zaharie developed with Fatima (Tim) Pardi and her children. Although it is behind a paywall, you can read the full article here:
https://www.reddit.com/r/MH370/comments/51g2al/new_report_mh370_pilots_friendship_with_mystery/
In the weeks prior to the disappearance, Zaharie separated from Tim Pardi and her family. According to Pardi:
The two saw each other less frequently from January 2014 because of a “personal matter” she would not elaborate on. Captain Zaharie continued to see her children after she urged him not to “let the children become victims of this separation”.
There was also a discussion before the flight that Pardi would not discuss:
But Fatima Pardi would not reveal the subject of their last WhatsApp discussion before the flight.
“That last conversation was just between me and him. I don’t want to talk about it,” she said.
Zaharie’s relationship and separation from Pardi and her family is not a smoking gun proving his involvement in the disappearance. However, this account does refute reports that there was no change in behavior before the disappearance. The circumstances around this relationship were also never disclosed by Malaysian authorities.
I rarely post on here as I admittedly am not a physics expert (to say the least.)
I was curious if all plausible areas for the debris sites have been searched based on the assumption of an autopiloted flight after the final turn south? In other words, is the search at this point restricted only to locations consisted with a manually piloted flight / glide?
Relatedly, is anyone still contending that it is likely / plausible the flight was NOT manually piloted until its end (or until minutes before its end)? Thank you very much.
@BRS: My opinion is we have to have patience about recommending new search areas until the present search areas is completely scanned. That said, there is no guarantee that the plane was flown on autopilot with no pilot inputs after 19:41. The problem is it becomes difficult to localize a search area because the probability map would be very diffuse.
@DonT
“If it’s possible to extract Turner’s treatment of the topic from whichever rabbit hole it has been posted, someone please relay an unrestricted reference.”
I believe Peter Turner has been giving his presentations in Australia over the last 12-months, but until recently the presentation content was not widely available. Recently however the presentation was posted on YouTube (below).
At the 1-hr 31-min mark, I cannot quite capture Turner’s exact words, but he states he tested the accusation of manual movement of Xponder to the ALT REPORTING OFF “in the aircraft” and he found that this action does not result in zero altitude reported as claimed, rather he finds that the prior last normal altitude reading is carried forward as the reported altitude. Harder to understand, but Turner seems to add words to the effect that another pilot also confirmed his observation on a B777.
https://www.youtube.com/watch?v=RxkL7GG3k-A
PS- @For those a little less sure about what we are talking about, the last ADSB transponder data reported at IGARI, which was only made available in more recent years (on this blog for example), shows a final several readings suddenly at Zero(0) altitude. This 0 reading has been widely interpreted by many as a pilot manually turning the transponder to OFF (STNDBY), because we can see that the switch was briefly rested in the wrong detent position that stops the altitude reporting.
Peter Turner claims this interpretation is false, and is falsely being used to blame the pilot(s). Turner claims to have conducted actual aircraft tests showing that manual movement of the transponder knob cannot cause a zero altitude reading. In Turner’s opinion the 0 reading could only be caused by one failure: an explosion of an oxygen tank in the MEC Bay, which Turner feels was likely caused by MAS maintenance using off-spec soap solution when they leak test the O2 tank tubing (via chloride stress corrosion cracking of stainless steel).
@TBill
Turner’s exact words: “So, I actually tried it in the aircraft. I got someone to try it in a Boeing 737 as well. When you turn altitude reporting off, it doesn’t report zero altitude. It actually coasts the previous altitude.”
He doesn’t explain how he conducted that experiment, but I assume he used a tracking application such as FR24, or he may have coordinated something with ATC. Either way, my guess is the altitude coasting he observed was caused by the tracking application or ATC software, not by the raw ADS-B data. As Don already explained, the last two ADS-B data messages contained no barometric altitude information.
If my understanding is correct, Turner’s experiment proves nothing.
@Andrew
I’ve leak tested thousands of miles of stainless steel tubing in the industry Lab setting, and its hard to imagine tap water exposure causing SS tube fails. But everyone uses SNOOP pro soap solution or better yet trace helium leak test. And further if the SS tube did fail, its hard to imagine this doing anything other than being a fast leaker.
@TBill: I had no idea that Turner is claiming that SCC caused the rupture of the O2 tank and/or tubing. Unless the leak detection methodology applied water with high concentrations of salt or chlorine, it is hard to imagine that SCC would significantly occur at room temperature, even with significant stress and the presence of oxygen.
First the claim that ALT RPGT OFF does not inhibit the reporting of barometric altitude in ADS-B packets, and then the claim that the oxygen system failed from SCC.
Turner is good at creating scenarios that are provably false with even a cursory investigation, yet fool a lot of people. (Does that sound familiar?)
@Victor
As you know, lots of Jedi mind trickery out there, and lots of folks wanting to believe the Jedi mind tricks.
@TBill
@Victor
I believe this theory came about from anecdotal reports of aircraft technicians using non-approved leak detection fluids (LDF) with oxygen system components. To be fair, various publications cite the risks of using non-approved LDF with copper-based alloys, often used in cylinder valves, due to corrosion. For example, EIGA Doc 78 (Leak Detection Fluids Use With Gas Cylinder Packages) states the following:
One of the possible constituents of LDFs is the ammonium radical, that in association with oxygen and water is frequently responsible for SCC of copper-based alloys such as brass and aluminium silicon bronze. These copper-based alloys are often used for manufacturing cylinder valves…The corrosive effects from certain LDFs coupled with the stresses to which a cylinder valve is subjected due to mechanically applied forces, residual stresses, and gas pressure can be sufficient to result in SCC.
It is difficult to specify exactly and with a high level of confidence the minimum amount of ammonia needed
to trigger SCC. However, it is known that moisture films on brass surfaces are capable of dissolving
residual ammonia from atmospheric contamination, resulting in SCC.
That said, I think there are other reasons Turner’s scenario can be discounted.
@Andrew: TBill referred to stress corrosion cracking of stainless steel tubing, not copper alloys (like bronze and brass). In response to him, I was referring to stainless steel. Wouldn’t stainless steel be the preferred material for the crew oxygen delivery system on the B777?
@Victor
I believe the tubing is stainless steel, but the valve body of the Safran 801307-00 oxygen cylinder and valve assembly is described as being made of “forged brass”.
My understanding of Turner’s scenario is that he believes the valve assembly failed due to corrosion following the use of non-approved leak detection fluids.
@Victor
My apologies, the Safran 806835-01 is the composite version of the crew oxygen cylinder typically used on the B777. It is described as having a “chrome-plated brass valve assembly”.
@Andrew: I imagine the chrome plating offers some but not perfect protection for the brass valve; however, I could find no cases where stress corrosion cracking caused this valve to fail on an aircraft. I would put this risk as “theoretically possible”.
Brass is definitely issue with ammonia cracking even at room temp. I started out in a corrosion R&D group and the metallurgist would collect cracked brass nuts/bolts to demonstrate it. But to my knowledge we never had any problem in a (dirty) industry setting with thousands of gas cylinders. I cannot speak for aviation industry material choice or oxygen systems. I assumed Turner tried to find an angle that had some reporting, is partially why I brought it up. Turner’s focus seemed to me mostly tap water chlorides fluorides on stainless I do not recall if brass/ammonia briefly mentioned.
PS- …on the SS tubing. Sounds to me like the chrome plating as a practical matter solves any brass/ammonia/valve issue.
Just looking at the two nil ADS-B altitudes it appears:
A – More likely ALT-OFF would still generate a valid altitude e.g. 35000 ft
B – FR24 had a file downloaded by Sladen that showed last Atlitudes all at 35000 ft but play back showed NULL altitudes 28 seconds
Apart. Also FR24 rounded all latitudes to five decimal places.
Except the last few. A computer program would not do that.
So it appears for A and B to happen the most likely explanation
Is data edits performed prior 7:24 am press release.
See also ads-b_nil_altitude_v02.pdf https://share.google/ew46ZIQL1EO1WEZiq.
@Mike Collins:
A: ALT-RPTG OFF produces no barometric altitude.
B: We have the ADSB data received by Malaysian ATC at Terengganu, which is similar to the FR24 data but the spacing between data points is shorter, and the data are the raw values that were actually broadcast by the aircraft. The barometric altitude goes to zero for the last data packet. Read more here:
https://mh370.radiantphysics.com/2019/04/03/insights-from-new-mh370-tracking-data/
There seems to be some doubt about what ADS-B altitude is output when in the ALT OFF position. I instead believe it much more likely to be the normal Baro altitude.
In addition there are issues with the recently released data (after 11 years since the aircraft went missing).
Raw Asterix Cat 21 data – should cover MH370 and other aircraft. Instead covers a totally different time frame.
Ascii file showing MH370 Asterix Cat 21 data. Firstly in aircraft accidents one would expect to see the raw data. Secondly it appears the track fields that are recalculated i.e. not directly from ADS-B data have too many duplicate tracks. So these could not have been transmitted by the aircraft. Also the ground pseudo ground speeds statistically have too many unique values compared to the 47 aircraft in the raw Asterix file.
The Genting ADS-B file has multiple location records that are in time to short. One expects the shortest gap to be around .4 of a second some are much shorter e.g. .2 of a second and much less. So to sum up it appears the data as it is could not be transmitted by the aircraft.
@Michael Collins: There are no reasons to believe that the ADS-B file that I shared back in April 2019 (from the ATC receiver at Terengganu) is not authentic. The ADS-B specification has a minimum update rate 2 Hz, i.e., every 0.5 seconds, with some variability introduced to prevent synchronous interference. To claim that “the data as it is could not be transmitted by the aircraft” is false.
I won’t let this website be used as a platform to make false claims. There are other platforms and user groups that seem to encourage this kind of silliness.
You mentioned that the Terengganu ADS-B data had valid time intervals
I totally agree with you on this. But I was discussing the Genting ADS-B file. Please have a look at the time gaps and you will see many that are too short. If this is not the case please let me know. Also in your last response you did not comment on the issues I mentioned about the Asterix files (the raw one and the ascii one). Can you please let me know if you agree with my comments or whether you believe there is nothing wrong withbthese files.
@Michael Collins: If you are asking me about the data provided by the CAPTION crew, I won’t comment. After they falsely accused me and others of changing our search recommendations based on their work, which is a batshit crazy accusation that is easily refuted just based on the chronology of articles and reports we’ve written, I refuse to read any more of their work. My comments to you are based on the data where I know the source and treatment of the data. I have no idea what the CAPTION crew is now claiming, and frankly, I don’t care.
Others like @Don Thompson may choose to weigh in.
You make a good point that the data I mentioned originated from CAPTION. Can you please put me straight on generally where does the MH370 data thatwe are discussing originate from. It appears a lot of it was made available to us via you and others via CAPTION. Say with regards the Genting ADS-B data and the Asterix data files (raw and ascii) which were published on the CAPTION website what do you think about their authenticity. For example the Genting ADS-B file would most likely be held on the Malaysian servers. Do you think that Malaysians gave the datacto CAPTION or do you think it was just made up by CAPTION?
@Victor
The point I would make, your post on April_2019 presents basic evidence that strongly suggests manual action by a pilot, secret until over 5-years after the accident. Also the home sim data is yet to be fully disclosed. By this time, public opinion was set in stone 11-12 years ago. Really we crowd sourcers and searches got off on the wrong foot from the start, likely due to the sensitivity of blaming pilot by many.
@Michael Collins: The ADS-B data published on this blog was confidentially provided to several of us and consists of the ADS-B data received by the ATC receiver at Terengganu. As I said before, I don’t know what CAPTION has published, including the source, the type of data, and whatever processing there was on the raw data. I don’t read the material published by the CAPTION crew. I doubt that the data was fabricated by CAPTION, but other than that, I have no idea what they published and what they are claiming.
CAPTION received the same Asterix we did, from the same source, but several years later. They parsed out some extra files (mainly on other aircraft) we did not deem necessary. On the files that matter, their raw results were essentially the same as ours. They claimed superior processing, but that is 100%BS. Their interpretation of the KB Civil PSR data is total garbage.
@airlandseaman: Perhaps you or Don can explain why the ADS-B data CAPTION has presented seems to differ from the Terengganu ADS-B data we have.
Flatpack: An excellent theory.
@Victor: My understanding is that CAPTION decoded the Terengganu ADS-B data from the Asterix Files. The file we received was the raw Terengganu ADS-B data from the receiver at Terengganu. According to their Report “Analysis of the 7 march 2014 raw ATC radar data”, the Asterix data files have evenly spaced time stamps (0.5 sec?) while the raw data we have has time stamps that vary slightly (by intentional design) between ~0.4 and 0.6 seconds. The transponder varies the true times of transmission to avoid systematic “collisions” in the random access channel. Apparently, the data in the Asterix file is interpolated to produce evenly spaced samples. AQs far as the actual path, there is no difference.
@Michael Collins
Marchand et al stated that the last ADS-B message received by the Genting station was at 17:16:20 UTC. They also stated that the last two ADS-B messages (with the missing barometric altitude data) were received by the Terengganu station at 17:20:34. The Genting ADS-B data is not relevant to a discussion about the missing barometric altitude data in the last two messages.
Marchand et al analysed the cause of the missing barometric data and unequivocally stated:
Therefore, the only possible remaining cause of the last two messages is that the transponder was in “ALT OFF” mode before it stopped emitting. Consequently, the end of transponder data is necessarily due to a deactivation from the cockpit.
@airlandseaman: Well, that would be odd if the CAPTION crew did not use the ADS-B data that was actually received at Terengganu since that data has been public since April 2019.
@Victor
When comparing the ADS-B data published here with their own analysis, Marchand et al stated:
In 2019 the Independent group published on mh370.radiantphysics.com website the ADS-B data from KT…the published data itself is very close to what is contained in the ATC-Caption-2025 recording raw data.
It seems to me their data is much the same as that published here and that previous comments by @Michael Collins have only caused confusion.
@Andrew So typical of the Caption folks. They have it exactly backwards. They say: “…the published data itself is very close to what is contained in the ATC-Caption-2025 recording raw data….”. But the ATC-Caption-2025 recording is not the raw data. What we published from Terengganu is the raw data, and what they plucked from the Asterix Files is “processed” data to fit the 0.5 second rates. It’s all a waste of time. We know thew path from multiple sources to 18:22.
A waypoint-based LNAV synthesis of post-Penang MH370 paths (work in progress)
Over the past several years, many members of the Independent Group and related contributors have explored candidate post-Penang paths for MH370 using a variety of methods: geometric constructions, kinematic spreadsheet models constrained by BTO/BFO, offset-path scenarios, and analyses of FMS behavior.
I have been revisiting some of that earlier work with the aim of stating, in operational terms, whether a single, continuous LNAV waypoint-defined route can account for the available evidence.
The intent is not to introduce a new path, but to consolidate and restate what many of us had almost converged on earlier, in a form that is explicit about FMS behavior and minimizes free parameters.
The draft treats the Lido Hotel image as a purely geometric waypoint problem and the Inmarsat as a constrained kinematic problem based on Inmarsat BTO/BFO data, and then shows that the same lateral path emerges from both approaches. It does not attempt to uniquely resolve post-ISBIX behavior.
A work-in-progress draft and supporting material are available here:
https://www.dropbox.com/scl/fi/okvorukznc0aat2b95o0z/A-Voyage-to….022226.pdf?rlkey=exm6lvua779hvkaim9p4onlt5&st=yibv2egu&dl=0
Corrections, clarifications, or references to prior work that I may have overlooked or inadequately cited would be very welcome. I am preparing a repository of publications that I believe are relevant.
@Sid Bennett
There is a conflict between the Honeywell and Boeing manuals regarding LNAV behaviour at a route discontinuity. Honeywell states “the aircraft maintains the existing track”, but the Boeing FCOM states “LNAV maintains the current heading”.
I queried this conflict some years ago with the technical training specialists at the airline where I worked. They queried Boeing, who confirmed that LNAV maintains heading not track. They said: “LNAV remains engaged after the airplane passes the last flight plan waypoint or the first waypoint preceding a flight plan discontinuity. Upon passing this waypoint, a CDU message is displayed, LNAV guidance outputs a heading command and maintains heading until LNAV is disengaged through the MCP.”
This behaviour was tested in a simulator, where it was found that LNAV does indeed maintain heading at a route discontinuity. It was also found that with the heading reference switch in NORM and outside the polar regions, LNAV maintained magnetic heading. The behaviour with the heading reference switch in TRUE was not tested, but in that case I suspect LNAV would maintain true heading.
@Andrew
I remember this discussion. I also have reference to the Honewell/Boeing manual which I believe supports my view. This needs to be sorted out based on documentation not opinion.
Magnetic headings have not led to much, perhaps because they are hard to compute :-). But even if it defaulted to a heading hold, and it was true heading, the rhumb line would proceed on the current gcp heading and the cross track drift would exist only due to the change in crosswind from the wind at the time of the default.
So, when people say it goes straight ahead, they are correct, but after a while there is difference between the modalities.
I believe I can find quotations from the Honeywell manual to support my position.
What is probably needed is a careful read through the ARINC spec which is controlling.
@Andrew
On the other hand at an end of route the navigation does drop out of LNAV, so there is a distinction between the end of route, which is a normal event and the route discontinuity which results from a programming error or an intentional deletion of a waypoint or the insertion of a direct-to command in an active route where there is no further waypoint.
@Sid Bennett
If I may make a comment on your Figure 2 Lido Hotel Photo, the blurry “head” in the lower left foreground looks to me like two hands holding a smart phone in landscape picture mode, with the holder’s right hand thumb and forefinger particularly visible in typical holding pose on the right. The implication is that photograph(s) in addition to the Fig 2 image were made by at least one other attendee at the briefing.
I think your waypoint-based LNAV route is unlikely to be controverted, but on the other hand there may be corroborating evidence. Do you know if any effort been made to canvass likely attendees at the briefing to see if more information can be gleaned from that session?
@Sid Bennett
With the exception of the last sentence, my comment was not an opinion.
As I said, there is a conflict between the Honeywell and Boeing manuals. Page 4.10.7 of the MAS Boeing 777 FCOM states:
LNAV maintains current heading when:
– passing the last active route waypoint
– passing the last waypoint prior to a route discontinuity
– passing the last route offset waypoint
– activating the inactive route or activating an airway intercept and not within LNAV engagement criteria.
Page 11.42.7 of the same manual states:
If the airplane passes the last active route waypoint (or offset) or the last waypoint prior to a route discontinuity, LNAV maintains the current heading and the scratchpad message END OF ROUTE displays.
As I also said, the conflict between the manuals was queried with Boeing. Boeing confirmed that LNAV maintains heading at the end of route AND at a route discontinuity. That behaviour was tested in a simulator, which confirmed the Boeing FCOM and the advice provided by Boeing specialists.
@Sid, I am unable to read your latest as dropbox says I need to request permission.
If you review our recent correspondence on this blog, you will recall that I have laid out the case for a waypoint-constrained route derived from BTO-optimal fit from Arc 1. There is a unique solution, namely: MEKAR-SANOB-IGEBO-RUNUT-XX. There is no named waypoint in the relevant part of the SIO to serve as the final waypoint. I have proposed 40S 85E and this produces excellent fit to the BTO data in LNAV until end of flight. The flight (from Penang onwards) is M0.84, FL360.
BTO analysis shows that from RUNUT (or even from IGEBO) the BTO-optimal solutions take you to the same place on the 7th arc: 39.60S and the terminus latitude confidence interval of BTO-compliant paths varies by a little over +/- 0.1 degree, even if we don’t know the specification of the final waypoint.
The fact is that the waypoint-constrained BTO optimum takes you to the only part of the 7th arc (within “search area”) that has not been covered. The route, speed and altitude predicted are all highly plausible.
@Paul Smithson
I had the same problem. The hyperlink Sid posted seems to be truncated. Try pasting the entire link into your browser:
https://www.dropbox.com/scl/fi/okvorukznc0aat2b95o0z/A-Voyage-to….022226.pdf?rlkey=exm6lvua779hvkaim9p4onlt5&st=yibv2egu&dl=0
That worked for me!
@Sid, correction – I was able to open when I highlighted the entire url:
https://www.dropbox.com/scl/fi/okvorukznc0aat2b95o0z/A-Voyage-to….022226.pdf?rlkey=exm6lvua779hvkaim9p4onlt5&st=yibv2egu&dl=0
Paths that optimised for BTO+BFO combined and no loiter will inevitably end up at about 38S, per IG’s original “hotspot” and the mode of DSTG’s pdf. That is because BFO and BTO optima are somewhat divergent (approx 39S and 37S respectively). If you optimise for BTO only you end up at 39.3 +/- 0.3.
There are several grounds to be sceptical about the fidelity of BFO, for reasons stated previously.
@Andrew
Thank you for laying out the Boeing FCOM citations clearly. I agree with you on one important point: LNAV does not invent a new lateral path at either an end-of-route or a route discontinuity, and Boeing documentation is explicit that LNAV outputs a heading command in those cases. Even if one assumes a transition to heading hold at end of route, the initial held heading is necessarily the inbound leg heading, so the aircraft initially continues along the extension of the great-circle leg. Any later divergence requires a change in wind or an explicit mode change.
Where I think the public discussion has drifted is in conflating three distinct things that the manuals themselves keep separate:
1. END OF ROUTE
2. ROUTE DISCONTINUITY
3. The geometric consequence of “maintains heading”
My paper hinges on keeping those distinctions explicit.
First, an END OF ROUTE is a normal, annunciated condition indicating that no active leg remains. Boeing explicitly treats this as a termination of the route structure.
A ROUTE DISCONTINUITY, by contrast, is defined by Honeywell as occurring when no defined path exists between successive waypoints, and it can be created intentionally by waypoint deletion or route modification. Importantly, the discontinuity exists beyond the aircraft, not beneath it, as long as the aircraft is still flying an active leg up to that point. See Appendix to my paper. In LNAV, lateral guidance is defined by the active leg geometry; in this context that leg is treated as a great-circle leg.
Second, with respect to LNAV behavior at a DISCONTINUITY:
Honeywell states that when LNAV is engaged and a route discontinuity is entered, the aircraft maintains the existing track while DISCONTINUITY is displayed. Boeing describes the exact same event as LNAV maintaining current heading. These are descriptions of the control law output, not contradictory navigation outcomes. The LNAV control law for waypoint navigation in our hypothesis is a great circle track.
If one assumes a transition to heading hold for the END OF ROUTE case, the held heading is necessarily the inbound leg heading, so the aircraft would also initially apparently continue along the extension of the great-circle leg. Any later divergence requires a change in wind or an explicit mode change.
So when people say “the aircraft goes straight ahead,” that statement is correct in both manuals’ terms. Where I part company with some interpretations is the implication that LNAV somehow terminates at a route discontinuity. Rather, the documentation supports: LNAV continues to command straight-ahead flight along the extension of the active leg until another mode or command intervenes.
This is sufficiently arcane that I would really like to talk it over with designer of the FMS. ARINC 474 might be a help also.
@Paul Smithson
I think we are addressing different questions, and it’s useful to keep them distinct.
Your work emphasizes trends in inferred end latitude driven primarily by BTO optimization, and you’ve been clear about having limited confidence in BFO when it begins to dominate a solution. BTO-only fits show relatively broad latitude sensitivity.
What I’m presenting here is something narrower and more constrained: a single, continuous, LNAV waypoint-defined route that satisfies both BTO and BFO simultaneously, and does so with a very sharp minimum in the combined goodness-of-fit.
The metric used was the root-sum-square of the normalized standard deviations of BTO and BFO, shown in the dashboard (Fig.8) of the paper.The theoretical lower bound is √2 ≈ 1.414, and the solution has a cost of ~1.62 with 21 contributing data points.
I don’t want to overplay the absolute value, but two things are worth noting. First, I have not previously seen a cost this low in my own work. Second—and more importantly—the minimum is sharp: small perturbations in start time, start location, or post-FMT south azimuth produce a rapid and noticeable increase in cost. This behavior did not arise from adding degrees of freedom, loitering, or path complexity; it emerged from small adjustments within an otherwise conventional LNAV framework as described in the paper.
So my point is neither that BTO-only trend analysis is wrong, nor that BFO should be treated as decisive in isolation. It is simply that there exists a waypoint-defined LNAV route for which both observables are simultaneously and tightly satisfied, and that coincidence seems difficult to dismiss as accidental.
@Sid Bennett said: If one assumes a transition to heading hold for the END OF ROUTE case, the held heading is necessarily the inbound leg heading, so the aircraft would also initially apparently continue along the extension of the great-circle leg. Any later divergence requires a change in wind or an explicit mode change.
I believe the last sentence is not correct for several reasons:
1. Following a great circle path requires a change in track as the longitude changes (except when following the equator) so the heading changes even if there is no change in wind when following a great circle.
2. The FMS follows a great circle leg by correcting for the cross-track deviation from the leg defined by two waypoints. After a route discontinuity or end of route, the cross-track deviation is not defined, so a great circle path is not followed.
3. The magnetic variation will vary along the route, which will change the true track when following a magnetic heading.
@Sid, thanks for reply.
Your dashboard at figure 8 says it shows the model residuals for your optimal (lowest cost) solution. It shows BTO errors of 3.2 sigma, 2.4 sigma, 2.5 sigma at 18:27:00, 18:28:15 and 19:41:00 respectively. A path with errors as large as that is highly improbable. I can’t agree that this constitutes “tightly satisfying” the BTO data.
You also say “BTO-only fits show relatively broad latitude sensitivity.”
If the top end of the path is anchored by last known position (or, in my case, assumed Arc 1 crossing latitude), BTO-compliant paths actually produce a fairly tight terminus latitude range, as I have previously shown. That optimal latitude lies 1 – 1.5 degrees south of the BTO+BFO optimum, adjacent to the segment of arc that hasn’t been searched.
It is true that the distribution of BTO-compliant paths is broader if you start at an arbitrary latitude on Arc 2, implicitly requiring an manoeuvre between Arc 1 and Arc 2. But even here, it produces the same model optimum (39.5S) at which the BTO fit is 8-10 times more probable than the paths out at the northern tail (33-34S).
I think many folks here have mistakenly adopted the view that BTO fit is more or less equivalent across the whole range of interest.
If you want a latitude-discrimination filter, how about excluding the length of Arc that has been searched out to 25NM either side?
@Andrew @Sid
Andrew- do we have a time/date stamp on your internal discussions/sim cases re: discontinuity? But I know that was about 2017 when Victor started this blog.
The point is: historically up to about the 2018 OI search, there were quite a few “straight path” “ghost” flight estimates based on deceased pilot with autopilot continuing on True Track after the last waypoint (now much less popular as we think it is not Straight after last waypoint).
Now we have more consensus that a Straight flight path must likely be a deliberate ghost flight setting, what I call a “hybrid” ghost flight, to South Pole or other distant waypoint set in LNAV.
Interestingly, Peter Turner has the last remaining “discontinuity ghost flight” and he is suggesting big “boomerang” huge curved path 192 magnetic heading via deceased pilot over-flight of WITT Banda Aceh, which he suggests starts southwesterly and bends back southeasterly curving over to 34-35s near IG POI. I have verified via quick flight sim test that Turner’s basic idea works due to magnetic heading curve and winds aloft.
Although I personally favor a “curved” (or heading-maneuver-to-southeast) explanation, a curved path requires a fairly drastic slow down, or it will hit Arc6 too soon. If, like Peter Turner, you want to postulate a curved “ghost” flight, a ghost flight requires autonomous slow down of the aircraft before Arc6 for some reason you need to suggest. I personally have no problem with the curve to Arc6 timing since I believe an active pilot did that maneuver with a big descent and slow down pre-Arc6.
@Sid Bennett: Perhaps when you said continue along the extension of the great-circle leg you meant along a path of constant track and not along a great circle. But even that is incorrect, as the aircraft would follow a constant magnetic heading, not a constant track after reaching end of route or a discontinuity. The inevitable change in winds and magnetic variation would drastically change the track.
@All
Not sure of anyones Noticed there is a photo of Armada 86 04 on Marine Trafiic taken today in Port Louis
https://www.marinetraffic.com/en/photos/of/ships/shipid:9670290/shipname:ARMADA%2086%2004?order=date_uploaded
@Paul Smithson
The quasi time series of BTO errors is −37,−93,−69,−73,7,27,−43,−4,49(μs). It has a mean value of -26.6 μs and a normalized (w.r.t. 30 μs) standard deviation of 1.6 with respect to the mean
The mean error is likely to be the result of a 15-30 second miss-registration of the track.
Sounds good to me.
All,
Sorry about the broken link. You need to copy the entire link and paste it in your browser.
@TBill
As far as the earlier discussions go, I would prefer to start the conversation with the information in the appendix to my recent paper.
@VictorI
I have re-read the primary Honeywell FMS documentation, and I think it resolves the issue cleanly once the stated assumptions are kept explicit.
Honeywell states that when LNAV is engaged and a ROUTE DISCONTINUITY is encountered, DISCONTINUITY is displayed and the aircraft maintains the existing track (Honeywell, Boeing 777 Flight Management System (FMS) Pilot’s Guide, Doc D201201000013 Rev 1, Jul 2014, Preflight, p. 4-36, Fig. 4-27).
https://studylib.net/doc/25749428/777-fms-pilot-guide?utm_source=chatgpt.com
There is nothing in this document that suggests that a route discontinuity causes LNAV to exit, nor that lateral guidance reverts to panel-selected heading at that point. The only change is that there is no next defined leg; the current leg guidance is not abandoned.
In the hypothesis presented in my paper, the active waypoint leg prior to the route discontinuity is a great-circle leg. If LNAV continues to “maintain the EXISTING track [emphasis added]” after the discontinuity, then the aircraft simply continues along the extension of that same leg geometry. If it was a great-circle path before the discontinuity, it remains a great-circle path after it. No discrete turn or change in lateral behavior is implied by the discontinuity itself.
The introduction of magnetic heading, wind-driven track drift, or magnetic-variation effects only makes sense under a different assumption: namely that LNAV leg guidance has ceased and the aircraft has reverted to AFDS/MCP heading-hold control. This assumption is not implied by a ROUTE DISCONTINUITY in the Honeywell documentation.
On the other hand:
If ISBIX were an END OF ROUTE, then LNAV leg guidance would indeed no longer apply, and subsequent lateral behavior would be governed by AFDS control laws and MCP panel settings, at which point magnetic heading and wind effects would become relevant. That is not the model being proposed.
So the disagreement is not about Earth curvature or wind physics or geomagnetic coordinates; it is about whether LNAV leg guidance continues at a route discontinuity. The primary documentation states that it does.
@Sid, if you calculate the probability of observing a set of BTO residuals with that variance using chi-sq goodness of fit test (2 tail) you get a p value 0.0014.
By way of contrast, paths with good BTO fit have p values 0.5 to 0.8 using the same metric.
@Sid Bennett:
1) Maintaining existing track after the discontinuity does not mean maintaining a great circle path, which would require a CHANGE in track.
2) As I said, the FMS holds a great circle path by minimizing the cross-track error of the leg defined by the start and ending waypoints of that leg. There is no ending waypoint at a route end or discontinuity, rendering the great circle path impossible.
3) We long ago acknowledged that the Honeywell documentation (stating a continuation of track) differs from the FCOM documentation (stating a continuation of heading). That’s why Andrew did the experiments in the Level D simulator and also why he queried Boeing. The experiments and inquiry both yielded the same answer: the heading is maintained after a discontinuity.
@Sid Bennett,
1. Your proposed cost function is inadequate in the sense that predicted dispersions smaller than expected are rewarded, not penalized as they should be.
This can occur when fitting a small number of measurements because systematic route errors can cancel part of the random error, especially at the ends of a sequence, artificially creating smaller residuals for a few of the data points.
2. The weighting of the BTO and BFO terms in your cost function should factor in their relative uncertainties. This can be done by dividing the sigma squared terms by the number of residuals of each type.
3. If the residual sigmas in your cost equation are calculated standard deviations, then you are not penalizing for having non-zero means for the BTORs and the BFORs. Of course, if you allow the BFO Bias to be fitted, you can always force the mean of the BFORs to be zero. You can’t do that for the BTORs, since the BTO Bias is fixed by the pre-flight calibration at the gate at KLIA.
@Sid Bennett: I missed something in your statistics calculation for the BTO values. The standard error of the mean is expected to be sigma/(n)^0.5. In this case, sigma = 29 us and n = 5 (for arcs 2-6), so you would expect the standard error to be 13 us. Your mean is 26.6 us, or 2 sigma from the expected mean of 0. That makes it improbable.
@Sid Bennett
“The introduction of magnetic heading, wind-driven track drift, or magnetic-variation effects only makes sense under a different assumption: namely that LNAV leg guidance has ceased and the aircraft has reverted to AFDS/MCP heading-hold control.”
That claim is not true. LNAV remains engaged at a route discontinuity, but maintains the current heading instead of a great circle track to the next waypoint. There is no change to the engaged AFDS mode and the LNAV heading guidance is not affected by the MCP setting. LNAV maintains whatever heading the aircraft was flying immediately before the discontinuity, irrespective of the MCP. That continues until such time as the pilot enters a route modification such as a DIRECT-TO, or engages another AFDS mode such as HDG SEL.
If the aircraft is maintaining a heading, the resultant track will be affected by drift and magnetic variation.
“If ISBIX were an END OF ROUTE, then LNAV leg guidance would indeed no longer apply, and subsequent lateral behavior would be governed by AFDS control laws and MCP panel settings, at which point magnetic heading and wind effects would become relevant.”
That claim is also incorrect. In the END OF ROUTE case, the LNAV guidance also reverts to the current heading. Again, there is no change to the AFDS mode and the LNAV heading guidance is not affected by the MCP.
@Paul Smithson
For the past decade we have been using the relative value of “cost” in our papers as a criterion when evaluating models, particularly for the FMT time and initial southern azimuth. When we were evaluating small changes in parameters, it was used as an objective discriminant.
Here we are venturing into the quantitative interpretation of small number sets. I should say up front that I had some help sanity-checking this — I’m not a practitioner of statistics (or sadistics…as I prefer).
The very small p-value you quote assumes the BTO residuals are IID (independent and identically distributed) with a fixed σ = 30 μs and ZERO mean. That’s a stronger assumption than I think is warranted here. A small constant bias in BTO (start time, start position, calibration) or for that matter in the BFO is entirely plausible, so treating the mean as an unknown and removing it may be a better formulation. (If I had my druthers, I would throw out the value at 00:19:30, but we have such a small data set…that’s also why we combine the BTO and BFO error statistics.)
Done this way, the variance test gives p ≈ 0.014 for σ₀ = 30 μs — suggesting some excess scatter relative to that nominal scale, but not something implausible for only nine points, especially since 30 μs is a weighting convention, not a guaranteed physical σ.
The result is also quite sensitive to the assumed σ₀. Using 40 μs, the same residuals gives p ≈ 0.32, which is entirely unremarkable.
So, I don’t think the χ² result undermines the solution; it mainly highlights how dependent the inference is on treating IID behavior and 30 μs as strict truths.
@VictorI, @Andrew
One point that may help clarify the discussion is that heading and track are not interchangeable quantities in either aviation practice or control theory. “Heading” is the orientation of the aircraft’s longitudinal axis, whereas “track” is the projection of the aircraft’s path over the ground.
This distinction is explicit in ICAO and FAA material: ICAO states that procedures are defined in terms of tracks or bearings, and that pilots maintain the track/bearing by applying corrections to heading to compensate for wind (ICAO Doc 8168, PANS-OPS, Vol. I, §1.1.2: https://ffac.ch/wp-content/uploads/2020/11/ICAO-Doc-8168-Volume-I-Flight-Procedures.pdf).
The FAA similarly defines heading as where the aircraft points and track as the actual ground path (FAA Pilot’s Handbook of Aeronautical Knowledge, Ch. 16 – Navigation: https://www.faa.gov/sites/faa.gov/files/18_phak_ch16.pdf). A concise ICAO-sourced summary of these definitions is also given by SKYbrary (https://skybrary.aero/articles/heading-track-and-radial).
I raise this not as a semantic issue, but because it matters operationally. Saying that a system “maintains heading” is fundamentally different from saying it “maintains track.” The former allows ground track to drift with wind and reference effects; the latter implies active correction of heading to preserve the ground path. In the case we are considering, the ground path (track) is a great-circle path between waypoints.
And, with that, I say good night to all….
@Sid Bennett
I can assure you I am very familiar with the meanings of heading and track and the implications of each.
With all due respect, you are over complicating a simple concept. At the end-of-route or a route discontinuity, LNAV will maintain the current heading. It is that simple.
@Sid Bennett: And I can assure you that EVERYBODY here understands the difference between heading and track.
As @Andrew has said: at the end-of-route or a route discontinuity, LNAV will maintain the current heading. It is that simple.
We’ve known this for years based on the confirming work of Andrew and his Level D simulator experiment and as further confirmed by Boeing.
@Sid. The path fit isn’t even at the margins of possibility. Changing the empirically-proven reference distribution of BTO errors (mean zero, sigma 29 microsecs) doesn’t wash.
But leave aside the probability and the absence of waypoint to sustain LNAV, I don’t understand why you are pursuing an analysis that will end up recommending a terminus near 38 South. It has already been searched out to 45NM (outside) and 30NM (inside the arc) and the plane ain’t there.
@Andrew, @VictorI, @Paul Smithson
Are you all in agreement that if ISBIX is the last waypoint, then LNAV is still active? If so, and the nav system maintains the current True heading, the plane will follow a rhumb line thereafter with cross-track drift due to winds where the heading is maintained at the value to compensate for the cross-track wind at ISBIX?
@Sid Bennett: After an end of route or discontinuity, the plane will likely follow a magnetic heading unless the NORM/TRUE switch has TRUE selected.
@Sid Bennett
Yes, LNAV will remain ‘active’ and will maintain the current heading at ISBIX. That heading would include any drift compensation required at ISBIX.
The heading reference for the PFDs, NDs, AFDS and FMCs is determined by the HDG REF switch, which is usually selected to NORM. With NORM selected, the heading reference would be magnetic, as Victor mentioned.
Further to my comment above, under normal circumstances the HDG REF switch is only selected to TRUE if the HDG SEL or TRK SEL AFDS modes are needed while operating in the polar region. The switch remains in NORM at all other times.
The deafening silence of Ocean Infinity on its failure to find the MH370 wreck?
Yesterday, on February 23th at 9 am local time , ARMADA 86 05 docked in Pago Pago US Samoa.
https://www.marinetraffic.com/en/ais/home/centerx:-170.684/centery:-14.278/zoom:16
4 red dots are fishing boats ; 2 green are cargo vessels; mauve is pleasure boat
the blue dot tug & special craft is ARMADA 86 05;
As all the way from Australia , without its identification.
@Jean-Marc Garot: We’ve known that OI is prioritizing commercial work over the MH370 search. There’s no reason to believe they won’t restart the search when vessels and equipment are available and weather is favorable. I’m not aware of statements to contrary.
@Victor:
Given the transit times involved, if neither A8604 deploys from Port Louis, nor A8606 deploys from Singapore, by the end of February 2026, it’s over, period, because, realistically, there are only 9 (possibly 10) weeks left in this ‘search season’ which is barely enough time to transit to the search area and complete the remainder of the stated 55 active search days.
Further, don’t forget that the 18-month contract ends in a few months’ time (September 2026 ?), which is well before the next ‘practical search season’ could open in late November 2026.
I agree with JMG that the ‘deafening silence’ from both OI and the Malaysian Government is telling.
The ‘elephant in the room’, the real question that should be asked is:
‘Why have both OI and MGov steadfastly refused to say anything ?’
Even if both OI and MGov have nothing but distain for ‘the wider MH370 community’, it is, if nothing else, clearly ‘cruel and inhumane treatment of the NOK’, especially in light of the MGov’s glowingly ‘self-touting’ of it’s bona-fides towards the NOK in the MGov media release of late December 2025. The word ‘hypocrites’ now seems rather apt.
Therefore, sadly, I think that your apparent ‘optimism’, is little more than unhelpful ‘wishful thinking’ at this stage of proceedings.
@Paul Smithson
I believe thatthe criticism that the BTO residuals have an average offset of ~30 µs is being overstated. As a back-of-the-envelope check:
In MH370 analyses the BTO residual is treated as an effective round-trip timing term, so a timing residual Δt maps to an aircraft–satellite slant-range residual of approximately
Δ𝜌≈𝑐Δt/2
Numerically, 1 µs corresponds to ~150 m of range, i.e. 0.081 NM per µs. Thus:
10 µs ≈ 0.81 NM slant range
30 µs ≈ 2.43 NM slant range
Even allowing for projection onto the Earth’s surface (a factor of order unity near equatorial geometries), a 30 µs mean error corresponds to only a few nautical miles, not a large displacement.
The samples are not independent in time.
The first three BTO measurements occur within a few minutes, while the most of the remainder are roughly hourly. With this clustering, it is not appropriate to treat the residuals as i.i.d. noise. Any modest track miss-registration or constant bias (e.g. 15–30 s timing offset) could push closely spaced samples in the same direction and naturally produce a non-zero mean.
Mean offset itself should be treated as a nuisance parameter.
Unless there is a compelling reason to assume the true mean must be zero, a non-zero average residual may suggest an un-modelled bias term. In that case, goodness-of-fit should be assessed from the scatter about the mean, or by explicitly fitting a bias.
For the present data set, a bias of ~−25 to −30 µs together with a standard deviation of ~45–50 µs appears to be a conservative and internally consistent interpretation of the residuals. Interpreted this way, the data are not anomalous.
@ventus45
My take:
MH370 remains a very sensitive subject in Malaysia. Hopes have been raised several times, only to have them dashed when search efforts came up empty-handed. That becomes very difficult for the Malaysian Government to manage at a political level. Its response has been to keep on-going search activities very low-key and to tightly control the messaging. I suspect that includes restrictions on media statements by OI. OI, in turn, are simply abiding by the terms of their contract.
Perhaps the lack of messaging is a good thing, in the sense that OI’s silence may be an indication they are trying to stay in the Malaysian Government’s favour in the hope of a renewed contract at some point. Who knows when the search might continue, but at the end of the day OI is a business and needs to prioritise commercial opportunities to generate revenue.
@Andrew @VictorI
I am taking your comments seriously and rethinking the “after ISBIX” portion of the paper. The crosswinds during the later portion of the flight seem to make a maintain heading approach impractical, not the least due to the uncertainty in the winds for modeling.
But, I still am trying to clarify the role of the cockpit switches. It seems to me that they have no effect on the LNAV mode and would only come into play if the FMS exited the LNAV mode.
Do I qualify as a contender for the Limpet award?
@Sid Bennett
“But, I still am trying to clarify the role of the cockpit switches. It seems to me that they have no effect on the LNAV mode and would only come into play if the FMS exited the LNAV mode.”
Are you referring to the HDG REF switch? If so, that switch does control the heading reference used by the FMC for LNAV commands. It also controls the heading reference used by other AFDS lateral modes, such as HDG SEL.
In NORM, the heading reference is magnetic north, unless the aircraft enters one of the polar regions (82°N/S or 70°N/60°S near the magnetic poles), in which case the heading reference automatically switches to true. In TRUE, the heading reference is true north regardless of latitude.
@Sid Bennett
The MCP heading/track selector and heading/track select switch, on the other hand, have no effect on the LNAV heading commanded by the FMC.
@Sid Bennett said: Mean offset itself should be treated as a nuisance parameter. Unless there is a compelling reason to assume the true mean must be zero, a non-zero average residual may suggest an un-modelled bias term. In that case, goodness-of-fit should be assessed from the scatter about the mean, or by explicitly fitting a bias.
We don’t see this offset in the DSTG test flights in which the bias does not change on the ground before departing and through the entire flight.
The fact that your path has a significant mean BTO error suggests a systematic error was introduced in your path selection. Perhaps you can change your cost function to be RMS errors for the BTO and BFO so the mean error is also minimized.
Armada 86 06 is leaving Singapore. VesselFinder lists “Vung Tau offshore” as next destination.
@Sid. I profoundly disagree with your comments above claiming that the BTO fit of your proposed path is fundamentally satisfactory. It isn’t. Further response from me will be futile, so I will stop here,
Reading what @ventus45 said on February 24th at 7.02.pm, I have to say that I share his pessimism about the timing of the future search. Regarding the silence from OI, the simplest explanation that comes to my mind would be that since the report of January 19th they have not actually scanned any part of the seabed, due to bad weather and poor interpretation of the weather forecast and that they simply consider silence to be better than admitting it.
They simply have nothing to say.
Besides, I have the impression that, apart from a couple of dozen participants here and on a few other spots, I do not see any great public interest in this matter.
For me, the good news would be the ASAP departure of an Armada vessel to the search site, equipped with ROV, which would show me that the OI really believes that they will find something (or, who knows, that they found something in the previous phase).
Of course, this is speculation and only one of the possible interpretations of the events surrounding the search.
@Vlad.BG: I understand your skepticism. For the time being, I am leaning towards OI’s continuing the search when able, but nobody knows for sure.
Concerning the last two KT ADS-B messages, as Don recalled, they do not contain an altitude of 0ft, their barometric altitude is just missing (i.e. value = 0 by convention, please note that 0ft would be coded as 12 by convention). As Captain Turner’s test was to see if the Alt off position would produce a 0ft altitude (i.e. value=12), the test is invalid.
The field in question is the altitude field in BDS 0,5 messages, which can contain either
-the barometric altitude (normal mode)
-the geometric altitude (indicated by the TYPE code)
or
-a “not available”/”invalid” altitude.
The absence of barometric altitude in the last two CAT021 messages indicates that the barometric altitude was not sent. The NUCp value of 7 shows that the geometric altitude was not sent either (in which case the NUCp should have been 0, 8 or 9).
ICAO documentation explains that if the geometric altitude is available but the barometric altitude is not, the transponder must send the geometric altitude (and change the TYPE code to indicate this), which did not happen. This should therefore mean that neither altitude is available.
However, a NUCp value of 7 is impossible without GNSS; therefore, we know that geometric altitude was available.
Thus the only other possibility is that the altitude field is marked as “unavailable”/”invalid” (i.e value=0) due to the “ALT RPTG OFF” position of the knob.
One has to keep in mind that on the ground, the aircraft send “BDS code 0,6 – Extended squitter surface position” while in flight they send “BDS code 0,5 – Extended squitter airborne position”. Following the specifications, “BDS code 0,6” cannot send the FL, so no surprise for its absence. A valid test would be recording the ADS-B data in flight only in ALT RPTG OFF.
The ATC-caption-2025 data has not been manipulated. It is genuine, and efforts have been made to make the CSV (and JSON) transcriptions as close as possible to the raw data. In the case of KT ADS-B, the dataset includes the raw binary data files in addition to the csv files. They certainly have NOT been re-arranged or re-scaled in time. The time tags are the original values. In the case of KT ADS-B in particular, the times are basically the same as published on this website in 2019.
Nevertheless, Michael’s point is noted and we are currently checking if an insidious tiny time shift did not occur here and there in the genuine Genting ADS-B data at the time of the data formatting and recording.
@All
Armada 8604 has sailed from Port Louis for Las Palmas, SPAIN eta 30 Mar 2026.
Armada 8606 has sailed from Singapore for offshore Vung Tau VIETNAM eta 28 Feb 2026.
@Andrew
Andrew, I have not ignored your “take” of (February 24, 2026 at 8:19 pm) but it has taken a little time to compile a response.
Firstly, I don’t think anyone would disagree that there is no doubt that MGov have frustrated all attempts to find MH370 from day one. The fact that ‘MH370 remains a very sensitive subject in Malaysia’ is mainly because of the secrecy and duplicity (not just the apparent incompetence or the ‘changing narratives’) that has also prevailed, from day one.
The best example of this came very early in this saga, when Hishammuddin Hussein’s interview in an Australian Broadcasting Commission 4 Corners program about a month after the event said it all.
Watch (https://www.youtube.com/watch?v=nukuLJ39i0U) from 32 min 40 sec to 35 min 23 sec. Then watch from 40 min 05 sec to 40 min 41 sec, where Hishammuddin Hussein puts a very ‘clever’ spin on the Inmarsat data, (to deny that the aircraft had flown for many hours) by dismissing it, by ascribing it ONLY to ‘engine data’ (which we now know obviously had not been received because ACARS was down).
By that stage, ‘blind Freddy’ could have deduced that the MGov obviously knew ‘who’ took MH370, ‘why’ MH370 was taken, and quite ‘possibly’ (roughly) where it might be.
I can’t see any logical reason why you should be (apparently) apologetic for MGov’s actions, either past or present.
Then we move to the SIO air search, which was not MGov’s idea at all, it was literally forced on Malaysia by Obama, and he told AusGov to have AMSA run the search. The initial air search was concentrated in the south (37S – 40S), supported by the NTSB ‘two tracks’ (which have mysteriously ‘vanished’ from the public record – funny about that) and other sat data.
Indeed, Chinese sat data indicated even further South West, but MGov did everything that they could to ‘force the search north’, and eventually won.
Why ?
The Chinese IL-76’s weren’t allowed to go out again, they were sent home, the air search was cancelled, the ATSB dumped AMSA, and thus began the saga of the 7th arc towed sonar’s.
The ATSB controlled FUGRO search covered a long narrow strip close to the arc. When PM Najib Razak and Hishammuddin Hussein came out here at separate times during Fugro’s search, it was clear from their body language, that they actually feared that MH370 might be found.
But, alas, Fugro did not find it, and the ATSB search was eventually terminated, and MH370 became a ‘cold case’, much to MGov’s relief, and China’s all too obvious studied inscrutable indifference.
Then, in 2018, OI sprung up out of nowhere, apparently funded by cashed up investment bankers, whom it was rumored, had a penchant for prior ‘clandestine missions’ exploiting (or attempting to exploit) historical shipwrecks with the potential for ‘lucrative payoffs’. Rumors suggested that some of these ‘missions’ had run fowl of various ‘authorities’.
It could be argued that OI was created as a means for them to ‘emerge from the shadows’ by bursting onto the world stage in such ‘a noble quest’ that no one would look deeper. Hiding in plain sight is perfect cover, and although the company was financially HQ’d in London’s ‘City’, it’s operations from 2018 until at least 2020, were apparently run out of Texas.
Anyway, OI made a proposal to MGov, but MGov stonewalled for months. In the meantime, OI prepared Seabed Constructor. When it became clear that MGov would not ‘come to the party’, OI sailed Seabed Constructor to Durban, effectively giving MGov ‘one last chance’ to accept their ‘no find no fee’ offer.
MGov was unmoved, so Seabed Constructor deployed from Durban for ‘the arc’.
By the time Seabed Constructor was about half way across the SIO, MGov knew that they could not ‘stone wall’ any longer, they caved, they signed, and two Royal Malaysian Navy Hydrographers were frantically flown to Perth, put on an ocean going tug, and eventually rendezvoused with Seabed Constructor at the arc, and boarded Seabed Constructor as Malaysia’s ‘official contract observers’.
So, the search for MH370 became OI’s backers / funders first ‘open mission’.
But while Seabed Constructor was searching, back in KL, MGov militarized it’s AIB, and when Seabed Constructor made a port call, they attempted to replace the two Navy Hydrographers with half a dozen Air Force Officers. That didn’t go well.
In the years since, OI has grown by leaps and bounds, with ‘heavy’ capital investment. If you look at their incredible growth rate, I doubt that the revenue generated by their operations to date has come anywhere near paying even a respectable interest return on those investments, let alone made any progress on paying back any of the seed capital. In other words, whoever is funding OI has very deep (bottomless ?) pockets, and is obviously playing a very long game.
Therefore, I very much doubt that your take’s central tenant that OI ‘needs to prioritize opportunities to generate revenue’ holds much water, and is highly unlikely to be ‘a valid escape clause’ for terminating the latest search (nor 7805’s aborted search last year for that matter).
Furthermore, quite clearly, a potential $70m ‘payoff’ for finding MH370 will not come anywhere near covering OI’s MH370 ‘sunk’ expenses from 2018 until now. So why are they persisting with this quest ?
In purely business terms, the only strategy that makes any kind of sense to me is that the company must consider ‘their MH370 project’ as a straight out ‘loss leader’, but the question becomes – for what ?
Perhaps (as you suggest) they are ‘currying favor’ with MGov as a legal contractual lever to guarantee them the ‘exclusive salvage contract’ when and if 9M-MRO is found (by anyone ? – it’s possible an oceanographic research ship might just ‘stumble’ upon it), but even that seems a bit farfetched.
In any case, there is no doubt that MGov do not want MH370 found, not ever, which effectively means, that if OI is only allowed to search where MGov is willing to let them search, then the MGov must be very confident that MH370 is definitely not there, which means that OI are on a hiding to nothing, they will never find it themselves under an MGov contract.
So that only leaves OI having ‘an in place first right to salvage agreement’ of some kind (if someone else finds it).
@Victor
“We’ve known that OI is prioritizing commercial work over the MH370 search.”
That does seem to be the case now (commercial work priority) but I do not recall knowing that explicitly, since of course the contract terms have not been disclosed.
@TBill: I have no knowledge of the contract details. But based on the tracks of the Armada vessels, the searches are fit in between commercial jobs. Considering the way Malaysia jerks them around each time there is a contract to be signed, wouldn’t you do the same?
@ventus
I suspect you are in line for a creative writing award – there are logical fallacies involved when promulgating a line of argument based on former asseverations.
However it seems to me that your observations of the growth of OI mirrors the explosive expansion of the Victorian branch of the National Safety Council of Australia in the 1980s Within just a few years from 1982 the NSCA went from a NGA advising on the needs of personal safety for front line emergency workers, to being a “Thunderbirds” organisation with helicopters, aeroplanes, even a submarine and technology not in use anywhere else in the world at the time, even contracting rescue services to sovereign military forces in Australia and internationally. The whole shebang disintegrated in 1989 when creditors realised they had been taken for a ride, with close to $½billion disappeared.
a naîve question :
of the 13 ARMADA ships, 8 first generation ARMADA 78 (from 01 to 08) and 5 new generation ARMADA 86 (from 01 to 05) , only one : ARMADA 05 has not been “displaying” its “identification” , for 16 days , from South West Australia to Pago Pago US Samoa (US semi independant Stat)
I an not a fan of conspiracy theory , but it puzzles me .
The only “event” which “may” have some bearing : Trump’s decison in January to relax the constraints for the exploitation of the ocean …especially in the Pacific Ocean
https://www.nrdc.org/bio/rebecca-loomis/trump-administration-racing-launch-deep-sea-mining-industry-despite-huge-risks
I could have changed the OI priorities and in the meantime the reason for ARMADA 05 to hide until …the ash falls down , something the Malaysian government may be pleased as well
@Jean-Marc Garot
I agree with some of your speculations. It may potentially be an important development.
@John
Nice description.
@ventus45
Have you considered alternative interpretations of your basic observations?
until further announcement? ARMADA 05 didn’t find the wreck.
Several options:
the wreck is there , but ARMADA didn’t find it . A blow to OI new Armada capabilities?
the wreck is not “there” , but what means “there”, since OI didn’t make public the zone , more or less in the same area and why ? …as far as I know
the wreck is in a area totally different
back to square one , of the thousand , Inmarsat compatible, plausible trajectories
better not to reopen the pandara box ? and conspiracy theories ?
@John
Wow, a fascinating read, an amazing story really !
https://en.wikipedia.org/wiki/National_Safety_Council_of_Australia
The Deep-Sea Salvage of SAA Flight 295
https://www.youtube.com/watch?v=EsZfdLiNR-Q
The Deep-Sea Search for Air India Flight 182
https://www.youtube.com/watch?v=cwBD8oIzjLQ
@All
https://www.dropbox.com/scl/fo/7oz4sknn37iw7nxexwk1o/APxj_zffRAbGix2qm_Zb31w?rlkey=o24jr5qv23utxrksga30et7ht&st=wmhxv35a&dl=0
I have taken the comments on my recent paper “A Voyage…” into account and posted a revised version of the paper to my dropbox folder.
When using the above link, I suggest copying the entire link and pasting it into you browser window.
@Sid
What date is the latest version?
https://www.dropbox.com/scl/fi/q08uppjhtd2ssoa1bcm6e/Voyage-to-revised-022628-r1.docx?rlkey=ru4zraeuyj85p9308cz0jj2du&st=194c7ozd&dl=0
Here is a direct link. I removed the previous version.
I’ve been following MH370 since 2014 and came across this sub. Great work to all the writers and publishers of this website; many individuals learn and draw their own conclusions. Regarding MH370 theories, I’ll share some insight consistent with my undergrad studies in both criminology & psychology.
(Informative information for the forum):
There is a consistent psychological trait associated with suicidal intentions. High intellects engage in methodical planning in their controlled demise (from AGARI onward). This phenomenon, termed the “suicide bounce,” describes an emotional uplift triggered by the relief of having found a definitive escape from their suffering. For Zaharie, a turbulent personal life involved his divorce and increasing isolation.
(My own assertion & theory):
Based on psychological fundamentals, a disturbed individual will seek alternatives in their disappearance. Psychologically, the removal of oxygen or the depressurization of the cabin is the “least mentally inducing” way out because it bypasses the “survival instinct” of the lizard brain. Therefore, why would Zaharie ditch MH370, given he’d drown if he survived the initial impact? BFO data already confirms a high-speed dive, consistent with no pilot at the controls (assumption).
Likewise, it is logical to “assume” that the following measures Zaharie took
1) Final military primary radar indicates a manual sharp turn.
2) Zaharie manually inputs a waypoint to the SIO.
3) Removal of the oxygen mask (painless way to end his mission)
(Evidence supporting this series of assumptions):
-Inmarsat data, i.e., MH370’s endpoint in SIO, is irrefutable.
-MH370 flies until fuel starvation. What occurs after the final ping is unknown, other than BFO data, suggesting a high-speed dive.
-The “Easy Way Out” replaces a violent impact with a painless physiological “fade.” Lack of radio contact: the “Ghost Plane” flight profile. This theory rejects a controlled ditching scenario.
-Lack of sizable debris (rules out a controlled ditching)
-Manual sharp turns, skirting Thai and Malaysian borders (as shown on radar). Again, these actions are not consistent with an incapacitated pilot.
-Zero distress calls (Squawk 7700)
I encourage anyone to disprove anything I’ve listed with evidence and an extensive psychological background.
-Ethan
@All
I looked at the posted paper and the first appendix was truncated. I reloaded in and this is the link..Sorry for the confusion.
https://www.dropbox.com/scl/fi/q08uppjhtd2ssoa1bcm6e/Voyage-to-revised-022628-r1.docx?rlkey=ru4zraeuyj85p9308cz0jj2du&st=6tyq1frx&dl=0
@Sid Bennett: If you minimized the RMS error rather than the standard deviation, especially for the BTO residuals, you would eliminate a lot of the criticism of this work. The path might not change very much.
@Sid Bennett
I note that your path ISBIX –> 43S088E passes within 4.02nm (per G-Earth) of my second last waypoint of 37S089E (which is ~31nm south of the 6th arc, prior to FE). (For information, 37S089E is the “head” of the GFZ and is my “final turn point” to then track directly towards my last (unreachable [due to FE} waypoint) which is IGPOL (41.89666S 085.000000E). The track 37S089E –> IGPOL is essentially the axis of the GFZ trench.
Question:
Does your “cost” increase or decrease if you use 37S89E as the “end of cruise south point”, and if so, by how much, and does the magnitude of the change rule 37S089E “in” as plausible, or rule 37S089E “out” as implausible in your view (perhaps given other constraints as well) ?
@Ethan Larsen
G’day Ethan,
What exactly are you trying to demonstrate? You are speculating about a matter for which there can be no definitive answer. That sort of endeavour tends to be right up there in both the total waste of time and much heat, very very little light categories.
You’ve invited contributors to “disprove anything I’ve listed with evidence and an extensive psychological background”. That’s just a bit rich, on the basis that you’ve provided no evidence at all to support your contentions, and there is nothing in what you have written that suggests an extensive psychological background.
To address some of your speculations and suppositions:
“There is a consistent psychological trait associated with suicidal intentions.”
Research has repeatedly demonstrated that far from there being a consistent psychological trait associated with suicidal intentions, there are in fact myriad factors that come into play. To the extent that there is one generally common trait, it is almost certainly a sense of hopelessness.
“High intellects engage in methodical planning in their controlled demise …”
I’ve yet to see any good evidence that supports that contention unconditionally.
“For Zaharie, a turbulent personal life involved his divorce …”
He wasn’t divorced.
“Based on psychological fundamentals, a disturbed individual will seek alternatives in their disappearance. Psychologically, the removal of oxygen or the depressurization of the cabin is the “least mentally inducing” way out …”
I’ve yet to see any good evidence that supports either of those contentions. Some suicides are meticulously planned, many (if not most) are largely, if not entirely impulsive. In terms of choosing a method for suiciding, access and availability tend to be the overriding factors.
With regards to choosing “the “least mentally inducing” way out”, here we might look at the data for mass murder-suicides by commercial pilots. Thankfully such events are vanishingly rare. Leaving aside MH370, since 1982, there have been seven (perhaps eight) incidents where a flight crew member has carried out or attempted mass murder-suicide using the aircraft they were crewing.
Despite wide ranging ages (27 – 59 years) and diverse cultural backgrounds, the methodologies have been stunningly consistent. There has never been any attempt to utilise “the “least mentally inducing” way out”, rather, at an opportune moment the perpetrator has simply pointed the aircraft at the ground/sea and flown it to impact.
@Paul Smithson @VictorI
The written paper says RSS and is what is actually computed. My statement that the data is de-meaned was from memory and incorrect. I computed the RSS of the BTO residuals in the image of my spread sheet and the normalized RSS matches the value in the cost column.
As I mentioned, initially the cost was used in a semi-quantitative to compare paths when we studied the turn time and initial gcp angle after the FMT. It is only now where changes in paramaters that are very small are involved that one would like to have physical interpretation. But the orientation of the route wrt the ping rings varies from tangent to orthogonal. Nevertheless the error is small.
I have tried adjusting the model of the S turn for the offset and have found that what improves the BTO data degrades the BFO data. The 15 sec granularity of the parameters is the limitation.
I feel that reducing the granularity of the kinemetric model to 1 se would improve the simulation but not change the result.
Thank you both for making me think more.
@All
If you want to understand the sensitivity of the cost results to parameters, I suggest that you download a copy of the spread sheet whose link I posted, make a working copy and at the MAIN page left hand block alter the clock start time. Start in 15 sec increments and observe the behavior of the cost.
I would think that people with sophisticated simulators could create good replicas of the waypoint path and investigate various ways of cancelling the offset at IGOGU, and other similar geographical end points that would bive similar azimuths after ISBIX.
@Sid Bennett: Your cost function as defined in the paper is the root sum squared of the calculated standard deviation of the BTO and BFO errors, both normalized by the respective expected standard deviation. That means the mean errors for both the BTO and BFO are ignored. If you mean something different, you should change the equation to avoid confusion.
@Mick Gilbert
Hi Mick,
Your cited precedents actually prove my point. SilkAir 185, EgyptAir 990, and Germanwings 9525 were ‘hot’ impulsive acts where the crash occurred within minutes of the pilot taking control. MH370 is the sole outlier because it involved 7 hours of active, methodical navigation. It is the only B777 to have its transponder and ACARS disabled manually at the exact ‘dead zone’ between ATC jurisdictions.
@Ethan Larsen
So what’s your argument now, Ethan? That events that share none of the observable characteristics exhibited by every other event in a set, are, in fact, part of the set because, well … outliers! I’m mean, what is an apple if not an outlier orange?!
And how can you conclude that SilkAir flight 185, and Germanwings flight 9525 were ‘hot’ impulsive acts? Because “the crash occurred within minutes of the pilot taking control“. Every recorded incident where a flight crew member has carried out or attempted mass murder-suicide using the aircraft they were crewing has concluded with a crash occurring within minutes of the pilot taking control. Were they all ‘hot’ impulsive acts?
And is it your contention that the airspace between ATC jurisdictions, typically referred to as a Flight Information Region (FIR) boundary is, how did you put it?, a ‘dead zone’?! That is patently preposterous! Outside of the controlled airspace around airports, FIR boundaries are probably the next most stringently and actively managed.
@Ethan Larsen
I would contend that “… within minutes of taking control …” is NOT impulsive, far from it.
In all other known cases, I beleive doing it earl was preplaned to ensure minimum time for intervention by other crew.
It was not an impulsive action at all, the perp had decided to do it, and do it that way, long before the flight took off.
The MH370 perp had decided to do it, and do it the way he did, as soon as he saw Hamid woulld be his FO.
New boy, needs to impress, will do anything the boss wants, send him out, secure door, depressurize, done.
Proceed with ‘Operation Disso’.
(“Doing the disso” is Australian slang for disappearing or leaving an event abruptly).
Mick Gilbert said: And is it your contention that the airspace between ATC jurisdictions, typically referred to as a Flight Information Region (FIR) boundary is, how did you put it?, a ‘dead zone’?! That is patently preposterous! Outside of the controlled airspace around airports, FIR boundaries are probably the next most stringently and actively managed.
The disappearance of MH370 has demonstrated that the hand-off between ATC controllers at the FIR boundary between Kuala Lumpur and Ho Chi Minh airspaces could lead to confusion if the plane does not check in with the next controller and the transponder is inoperative. Also, the Malaysian military radar was at its limit of detection, so referring to it as a “dead zone” is not that inaccurate.
@ventus45
RE: “Doing the disso” is Australian slang for disappearing or leaving an event abruptly.
I have never heard that expression in my 62 years as an Australian citizen!
@Andrew,
I’m 71yo and my grandparents used it often !
@Ethan Larsen
I try to list (the many) reasons for denial of apparent-pilot deliberate action. One item I call lack-of-precedence denial. This is common false belief all such accidents must have exact prior example, if not, then it did not happen this time.
MH370 looks a little like 9/11 where Xponders were cut off and there was a plan of continued flight for a target. Of course 9/11 was pre-reinforced cockpit doors which have so far proven nearly completely effective in preventing passenger tampering by getting into the cockpit during flight.
@Victor Iannello
Victor, referring to a FIR boundary as a “dead zone” is most assuredly misleading unless, of course, you’re just interested in further propping up a specific speculative narrative. That sort of subjective distortion is sadly commonplace across the MH370 community.
@TBill
Thanks for raising the “lack-of-precedence denial”, Bill.
A wonderful example of how logic is applied asymmetrically (and I would argue, quite dishonestly) by proponents of specific speculative narratives in that “lack-of-precedence” is happily deployed to counter all the mechanical/electrical failure scenarios, and yet ignored or rejected completely when it comes to their preferred scenario.
Yes, Mick Gilbert and TBill: Agree. There can be many variations.
Captain Zaharie wasn’t divorced, but his wife and children had reportedly moved out of the house.
Zaharie had also reportedly become obsessed with young twin Instagram models, and engaged in affairs (obsessive and hypersexual components of bvFTD).
He also reportedly paced constantly around his empty house between flights (common motor stereotypy).
He was at the right age (50s).
We have to know if he started spending a lot of money or making bad investments.
But mainly if his personality changed: Did he become taciturn, lose his sense of humor, start snapping at people? Did he withdraw from family and friends?
I can understand why people may think he knocked himself out before he took MH370 down to the water, but we must remind ourselves that he was irrational and so far gone by then for us to even begin to guess. But, if I were to guess, I would say he went down with the ship, aware and awake.
I await final reports of China Eastern, Air India and EgyptAir. Eventually.
https://www.facebook.com/100064746874584/posts/1345245560976994/
Official communication to the families. It doesn’t say so explicitly, but it looks like search is over, for now.
@Paul Smithson: It’s disappointing that the current search area was not completed, but I understand OI’s decision.
Certainly this reduces the prospects of additional searching by OI.
@PaulS
…also my current interpretation, the blues of the 12th?
And , if OI didn’t find the wreck , may be it is because the wreck is not in the search zone?
And if the search zone is not the right one, may be the trajectory which led to it is not the right one?
And if the trajectory , even it fits with the BTO and BFO ,is not the right one,
amoung the thousand (as written in the Inmarsat study) compatible with the BTO/BFO maybe there is an other one?
Statement from Ocean Infinity
https://oceaninfinity.com/news/conclusion-of-the-search-for-malaysian-airlines-flight-mh370/
Thanks for sharing. The important bit is at the end: “Although this phase of the search has concluded, our commitment has not. We’re continuing to work with the Malaysian Government in the hope of being able to return when circumstances allow.”
so what : it doesn’t answer to my questions
@Jean-Marc Garot: You are asking questions that cannot be objectively answered. Either the MH370 debris field is in the proposed search area and was missed or not yet searched, or it was not in the proposed search area. If not in the proposed search area, there MAY be other solutions involving automated flight that satisfy all constraints, i.e., BTO, BFO, navigation, weather, fuel, drift, etc. Or, the flight was not automated until the end, in which case the problem becomes fairly unconstrained and many solutions are possible, and we are left to following hunches to reduce the search area.
Frankly, my advice would be to complete the current search before moving on to other areas.
@paul smithson
I agree, Paul. Glass half-full, although we can only wonder what the bar is now in terms of credible new information/evidence.
Clearly, and unsurprisingly, the WSPR Ouija Board for One does not clear the bar.
@Jean-Marc Garot
I agree with your inference: it was probably a different (active pilot) flight path. I am not optimistic as far as actually finding. Even if we had a better handle on actual flight path, it might still be near-impossible to find. Thanks to Inmarsat and debris finds, we already know the basic generic answer to where/what happened.
The only words in that OI release that are relevant, specifically, the charging herd of elephants, trumpeting, with tusks at the ready, are the very last:
“when circumstances allow”.
This begs the most obvious of all questions:
“what circumstance(s) DIS-allow”.
To my way of thinking, there are only 5 broad categories of candidates:
1 – Weather: – no – season is still open
2 – Vessel: – no – 05 and 06 still viable – (06 closer / sooner)
3 – Equipment: – no – proven technology
4 – Company: – no – keen to continue
5 – Malaysia – contract – probable – unacceptable (to OI) variation of terms ?
I could never understand why OI accepted a contract which obviously has conditions which prohibit them from making ‘unauthorized’ public statements.
Perhaps OI initially thought that it was in their interests to maintain a kind of “doctor – patient” confidentiality, but clearly that has backfired on them.
OI now face a possible PR disaster, a bit like those doctors who kept Andreas Lubitz secrets from authorities. I might add that those who sign NDA’s run similar risks in the long run.
@All:
I have written an open response to Ocean Infinity’s conclusion of the current search. It’s a bit long, so here’s a link to GoogleDocs:
https://tinyurl.com/2026Mar9-Java-Candidate
The Java anomaly would seem to be the only search candidate in calm waters that is specific enough to be searched en route to other work before the $70M contract expires in June.
If ever there were a time to recommend to OI that it eventually be searched, I believe it is now.
@370Location: I think that letter is fair. I hope your candidate area is searched before the contract expires.
@370Location
Good letter, Ed. While Paul S’s site doesn’t get the favourable weather tick that yours does, I think that his and yours should definitely feature in any renewed search effort.
You might recall that I put a few questions to the forum earlier in the year. Anyone who is confident that the 7th arc is accurately mapped, and that the final BFOs suggest a rapid, and rapidly increasing, rate of descent, should be throwing their weight behind any previously unsearched areas close to the arc, be they far north or far south.
@370location – Congratulations on such a clear document that even a reader of the blog like me can understand. Hope this document garners the attention from OI that your candidate site deserves. Wishing the best to you, Victor and all contributors here for the tireless efforts in resolving this mystery for the sake of the families.
@370Location
I also hope that your proposed site is searched at the earliest opportunity.
If the accident site is 42 miles off the coast of Java, it is outside Indonesia’s territorial waters. Indonesia would only have responsibility for the investigation under ICAO Annex 13 rules if the accident site were within its territory. In this case, Malaysia would still have responsibility, as the State of Registry of the aircraft. Indonesia would, however, be required to provide assistance as the State nearest the scene of the accident.
@Victor, @Mick, @ST, Andrew:
Thanks for the words of support!
I’m hoping that OI and Malaysia will see the writeup as credible new evidence for a specific site.
@Mick:
I do recall that discussion. OI has effectively said that they don’t have time for an extended search in rough seas. I don’t think any other viable site is as specific as mine. If it’s the only site searchable before June, then I think that makes it the only priority site at the moment.
My recommendation on the earlier discussion was to start at Java end of the 7th Arc and do a single seabed swath en route to the previous search area. They might do a variant of that before June if they have time, but I think the swath is max 1 km wide.
@Andrew:
Thanks for the correction on the Annex 13 being 12 nm territory. I wrongly assumed that it extended to the full 200 mile Exclusive Economic Zone which covers seabed surveys. (And obviously not 200 miles near Christmas Island.)
I can only admire such great minds, and truly enjoy this site for more than 10 years now. I miss Dennis Workman , with his whit, intelligence and sharp tongue. So thank you for also educating the lay person on scientific and academic subjects related to MH 370. It never dulls. And thanks for always reminding us that data and facts advance theories by confirming or denying them. Sadly we will likely not know with certainty what really transpired. I grew up in his contrails so can say with some certainty that his deed had to be hidden to protect his direct family and children from shame of his actions, to be ostracized .In that he succeeded. We are simply not meant to find out . It is and will remain a touchy subject for Malaysia. Profiling Z as a commenter wrote the other day is futile imo. Speculating on his ‘end’ game gives so many variables that it is best to stick with what we know. If human behavior is the benchmark , he would have ended it on his terms, at his chosen end point with a big middle finger to MA and MY government. If anyone can figure the location out, It’s you!
[Comments here are closed. Please continue the discussion under the new article.]