Possible MH370 Debris Found in Tanzania — UPDATED

Pemba 1

Via alert reader @Susie, photos have emerged on Tanzanian social media of an object that looks very much like a control surface from an airliner. Here’s what Bing Translator makes of the original Kiswahili:

Wing of the plane have been conflict and civilians today in the Indian Ocean on the island of Kojani. Made known is what airlines.

A place where it is believed to the remains of the plane have been caught in the Indian Ocean Island beach in Kojani.

Wing of the plane was found in the island of Kojani, is eliciting a great debate among the inhabitants of the island with many believing it is the wing flight of malaysia which was lost without a known future. Airlines of Malaysia Airlines with type MH370, it had disappeared March 8 in 2014 has never been visible until today.

Though still no certainty is what bird fossils, experts of air travel have started initial stages of the investigation of the wreckage of the plane.

Reports say security officials already have started to investigate the wing and probably not long we get enough information from entities involved

Kojani is a small inhabited island near Pemba, about 50 nautical miles north of Zanzibar and 500 miles north of the beaches in Mozambique where MH370 debris has previously been found. It has been described as “one of the least accessible villages [of Pemba], located on an islet off the eastern coast of the main island. At the last count Kojani was home to more than nine thousand people.” While still south of the equator, it is by far the northernmost debris from MH370 identified so far, if that is indeed what it is.

Pemba 2

Its appearance is strongly reminiscent of the flaperon found on Réunion island, although there seems to be none of the broken-off hinge attachments and so forth that were visible on the ends and underside of the flaperon. Also, there is a very visible waterline, which the flaperon lacked. It would be interesting to know if this waterline corresponds with that observed by the French investigators when they put the flaperon in their test tank in Toulouse.

So what is it, exactly? Commenter @Rob suggests it “Might be a piece of inboard flap.” @Ken Goodwin writes “Though the part has the shape of a wing part. It does not jog the memory. Closed large end with no fittings. Surface with no fittings. ???” Of course it might not be from MH370 at all. But if it is, it breaks from the recent trend of debris items being small enough to hold with one hand.

I hope that somehow this object finds its way into the hands of independent investigators who can examine it before it disappears into the black hole that is the Malaysian investigation.

UPDATE 6/24/16: New photos from Jamilforums below.

Flap 2


Flap 3


Flap 4

For references, here’s a shot of the outboard end of the right flaperon found on Réunion:

Outboard leading edge crop

In both pieces it seems that the main structure is aluminum, with the curved leading edge made of composite material.

229 thoughts on “Possible MH370 Debris Found in Tanzania — UPDATED”

  1. @VictorI,

    I agree that the assumptions involved in solving for a specific route need to be re-examined. I am also doing this by exploring all possible combinations of speed mode and LNAV mode. As you said, other route solutions are also possible.

    I have several questions related to your magnetic heading route:

    1.Can you think of any reason why a pilot would enter a desecent rate of -100 fpm?

    2.How would the speed change from M0.84 to 310 kias be implemented? Would this require pilot input between 20:41 and 21:41 to the MCP? Could it be preset beforehand in any way using the FMC to occur later?

    3.You indicate an ascent occurred between 18:40 and 19:41 after the descent at 18:40. Can you think of any reason a pilot would descend (at 18:40), climb (by 19:41) and then descend again very slowly (after 19:41)?

    4.The change to magnetic heading at ~19:41 could have occurred by a route discontinuity in the FMC after passing the last entered waypoint. How close is the nearest waypoint to the track near 19:41? If there is none very close, then it would seem pilot action would be required at about 19:41 to change the LNAV mode, presumably using the MCP, right?

    I have generated similar routes with northwesterly bearings at 18:40. One of my concerns all along has been that they require a brief but high-rate descent exactly coincident with the 18:40 phone call. While this is certainly possible, it seems an unlikely coincidence to me.

    In my opinion BFO errors as large as 20 Hz are quite unlikely. It would be helpful if you would post the BFO errors for your route. Generally speaking, I would expect BFO errors less than 7 Hz and without an obvious linear variation over the post-FMT route.

  2. @gysbrect
    ‘An absolute impossibility’
    What in that statement makes you think i believe the suggestion to be a viable theory????


    The linked image of the concentric circles showing the ping rings is the perfect evidence to show why it is not possible.

    Try reafing posts before atacking people.

  3. @Bobby Ulich: Responses to your questions below.

    1. I see no sensible reason for a pilot to intentionally send the plane into the SIO, no matter what autopilot setting. That said, if a simulated flight path that ended in the SIO was found on the pilot’s home computer (which I believe did indeed occur), his behavior might not have been what I would consider to be sensible. A slow descent might have been chosen to minimize debris from an impact, even though the aircraft was near 11,000 ft at fuel exhaustion.
    2. There is no required pilot input to transition from M0.84 to 310 KIAS during a descent when the autothrust mode is in SPD. I have intentionally looked for a flight mode requiring no pilot input after the turn at BEDAX.
    3. I believe that changes in altitude occurred at various times after the turn back at IGARI, as evidenced by variation in speed (such as a possible chandelle turn at IGARI), intermittent radar captures along the path, and a cell phone connect near Penang. Knowing this, I think it is reasonable that there were changes in altitude at 18:40, possibly to evade radar.
    4. If the final waypoints were BEDAX-ISBIX, and the waypoint was overflown, then the heading would be about 182M, and the fit is not as good. I am proposing that the pilot intentionally set a heading of 180M and a descent rate of -100 fpm after BEDAX.

    Other comments:
    I am not sure why you are willing to accept a climb during the log-on sequence after 18:25, but not a descent at 18:40. There were many things happening in this time frame.

    The BFO errors are 9,6,4,4,3,3 Hz at times 19:41, 20:41, 21:41, 23:14, and 00:11. I referenced 20 Hz only because that is the number used by the DSTG. From a practical vantage point, it is difficult to use the BFO to discriminate between acceptable paths because of the relatively errors in BFO that are possible. In this sense, I remain consistent with the DSTG study.

  4. @Jeff
    ya, accepted; at least, I dont have any joy that @alsm doc mentions the eliptic access port and Dubai take-off video as it all goes against my crazy no-crash theory; so sometimes posting something about hope – but it may be all wrong.

  5. @Victor: thank you for your paper, which I read with interest. Here are my quick reactions:

    1) I see no mention of an Arc 7 / fuel limit intersection – originally, a flagship criterion for adjudicating scenarios. If the plane follows your path, do fuel models you’ve studied predict exhaustion at [00:19 minus reboot time]? If so, that would be surprising, since the ATSB’s fuel limits bulge out considerably beyond Arc 7 by 31S – meaning MH370 would have had excess fuel if it took such paths. I’m sure you’ve thought of this – can you reconcile their limits to your theory?

    2) Personally, I am skeptical of the “pilot sim” story. I mean, it’s not exactly swimming in accountability, here – no actual person stepping up to the mic; no actual numeric data – just a very vague statement seemingly designed to create a very strong impression of pilot guilt. Made at roughly the same time every single nation involved was chanting in unison that the hard drive contents CLEARED the pilot of any suspicion (how soon we forget…). Until someone in authority – who has the guts to use his/her name – specifies an entire and precise flight path, I am going to assume that this supposed FBI claim is not credible. And wonder why it was made in the first place.

    3) Richard Godfrey is to be commended on his work, which attempted to find the most likely spot on Arc 7 to search, given 9 suspected or confirmed MH370 debris finds. (Also Henrik Rydberg, months earlier, using the flaperon only.) My concerns are a) he only looks between 28s and 40s on Arc 7: areas further north on the arc – and MUCH further north, OFF the arc – are each deemed by his own method to be far more likely than any points on his tiny subset. I can’t argue with the statement, “if we’re determined to look between 28s and 40s on Arc7, then s30 is the least bad place” – but I can – and will – show how myopic this logic can look, when put in broader context.

    4) Even points as far north as 31s still run into serious issues trying to explain zero debris on Australian shores. The UWA graphs I posted paint a very bleak picture.

    5) The 100fpm assumption is a new degree of freedom introduced to makes the data fit. It is hypothesized that a decision to descend at this painfully slow rate may have been a conscious decision to minimize debris. While I am all in favour of relaxing prior assumptions, I do not find the “maybe he wanted to minimize debris” argument persuasive, and thus fail to see any reason for adopting this assumption other than that it creates an autopilot-generatable flight path that misses the searched out segments of Arc 7. Victor, I know you have conceded as much in past statements; so I think we’re on the same page, here.

    I admit that I have an inherent bias against scenarios which pin this catastrophe on anyone unable to defend themselves, and thus demand a much higher standard of hard evidence. But even the cold hard science seems to be suggesting to me that far less bizarre explanations of the plane’s fate are readily available – all of which are variations on the simple, plausible idea that, in Florence de Changy’s words: “something happened that cannot be admitted”. Under such an idea, the signal data (and perhaps even the debris to date) are simply car keys being jangled at us – something to ogle at and reach for…until we tire, and fall asleep.

  6. @Victor: @Brock:

    The “fuel” question might have another possible explanation.

    What “if”, the ghost flight objective was to minimise debris with slow speed impact with flaps down (as the latest debris would suggest).

    In that case, after the turn, he would have to descend to a lower altitude than normal cruise, and “slow down”, very early, to flap extension speed. Then the slower ground speed and higher fuel consumption combined with a ROD of -100fpm, would, at one stroke, account for your excess 11,000 feet at the 7th arc, account for using the remaining fuel faster, to produce fuel exhaustion at the 7th arc, and, by implication, would curve the track made good even further east, thus putting the aircraft even further north on the arc, would it not ?

  7. @VictorI
    Thanks for your latest paper. I wondered whether your new trajectory now fits with Kate Tee’s observation but it doesn’t seem to be close enough (as far as I can see from various times, locations sourced from thehuntformh370). Any thoughts?

    Thanks too to ALSM, Ge Rijn and others for identifying the Tanzanian debris. If this is indeed from 9M-MRO then I can’t help feeling some of these factors must be related:
    Several pieces from R wing; comparatively few pieces of debris from elsewhere; maintenance records lost; R wing damage in China in 2012. The ditching theory may have some merit but for it to explain the RH side debris is a coincidence too far IMO.

    @Brock. Many good points, I particularly agree with your point 3) above.

  8. Jeff, 2 years ago, many people speculated that MH370 could have been hijacked to Somalia. But this was quickly dissmissed, saying that it did not have enough fuel. But this debris is getting very close to Somalia, in fact…All of the debris could fit a drift pattern from a plane that crashe short of reaching Somalia. It would not be the first time Hijackers did not believe the pilots saying there is not enough fuel, and make a desperate attempt to get to their destination.

  9. @ Ge Rijn. Outer flap part: separation. A couple more observations from this photo series, posted previously.


    The pin which has the locked nut, third photo down is, at one end of the ‘pivot link’. This connects the rear of the flap to its operating mechanism. Supposing part of the pivot link might be retained on this flap section, the evidence of its fractured surface could help with failure mode. The same applies to the fracture at the front surface of the hinge/attachment fitting (ie ‘carriage’) a second flap securing point.

    We might need to await findings.

    Also, from the top and third photos, the missing section of the trailing edge is from the outer end of this inboard section. Were flaps deployed and wings level the inner end would be lower.

  10. @VictorI,

    Thank you for responding to my questions and providing the BFO residuals.

    I have some additional comments:

    1. I would agree with other commenters here that the fuel question must be answered before any route can be seriously considered. By answered, I mean that a fuel model calculation must be done (as you have done for other routes in the past), to determine what average engine PDA will produce second-engine flame-out at 00:17:29. In addition, this fuel model must include the elevated temperature effect, which increases the fuel consumption by at least 3% for most routes. My recent calculations have demonstrated that great circle or true track routes ending near 39-40 degrees South actually cannot achieve a flame-out time that late. They appear to just make it to 00:17:29 without including this effect. Unfortunately, so far the ATSB has been completely silent or obstructive on this important point.

    2. Your BFO errors have an acceptable RMS value (one should not use the standard deviation about the mean for this evaluation). That is good, but they are all of the same sign, and there seems to be a significant systematic slope from one end of the post-FMT route to the other end. Those two characteristics indicate residual systematic trends that are somewhat uncharacteristic of the expected behavior of the correct route solution. In my opinion this argues against your route being correct.

    3. I don’t think any weight can be placed on completely unverified reports of (1) a Penang cell tower connection or (2) far south routes into the SIO on the pilot’s home simulator. You are free to believe whatever you like, but these are in the same category as garbled radio transmissions and disintegrating aircraft. To date, there is simply no verifiable basis for any of them as far as I know.

    4. I am quite willing to believe a descent near FMT, but the ROD needed to match the BFO at 18:40 is quite high – near -2,500 fpm. Since the phone call is about a minute long, the descent must have been at least 2,500 feet. In order to end up near your assumed altitude of 37,400 feet, the aircraft would have to have been at about 40,000 feet at the start of the call. For this to have occurred, the call time and the descent time would need to be the same within a small fraction of a minute. For a two-minute descent, the descent would have begun above ~42,000 feet, and that only produces a one-minute difference between the descent and the phone call. It’s not impossible, just very unlikely in my opinion. In addition, if the desire were to avoid radar, why climb to 40,000-42,000 feet to begin with, making the aircraft more detectable?

    5. While I am not familiar with the speed mode you assume, it does seem to be unusual. Maybe it is used in long descents, if those are ever done. I agree with Brock on this point that it appears that an unorthodox combination of declining speed and low descent rate is required to fit the satellite data at 180 degrees magnetic heading. There are orthodox solutions available for other LNAV modes, including some outside the current search area. I will have more to say about this later on.

    6. Possibly the final heading was set simply to make sure the aircraft did not eventually crash on land in a populated area. You could say the same thing about it heading out the Strait of Malacca after passing Penang.

    7. The intermittent radar captures just prior to 18:22 could be due to a descent and a climb, but in my opinion the altitude over the entire path from ~17:21 to ~18:15 must have been rather high (near 35,000 ft) in order to generate the very high observed average speed.

  11. I assume the more northerly suggested crash site is not close to a previous/original search zone?

  12. @ventus45: the altitude reduction in Victor’s scenario can’t be materially altered, as its pattern is needed to create the steady ground speed reduction required by any magnetic south heading/track that actually hits all BTO arcs on cue. If the plane flew one of these modes at cruising altitude the whole way, the last few BTO arcs would have been overflown.

    And if it flew the slow path you suggest, a ghost flight on a magnetic south track would fall SHORT of the BTO arcs.

    My prior post criticized it, but I do admire the intelligence with which Victor has crafted this scenario – one of the “least implausible” among those which still trust the Inmarsat data.

    Even those of us who have stopped looking up and down the 7th Arc for the truth (wreckage? maybe; the truth? not bloody likely) are deeply indebted to folks like Victor, who set for us a high water mark for plausible ISAT-compliant scenarios. If we can’t swallow scenarios like this, we’re pretty much done with “Camp 1”.

  13. @Victorl

    Thank you for that informative update..I clearly missed those latest explanations..So I stand corrected. Even though the Ping data might be correct.. I still raise the questions of
    1- Malaysian military radar track which suggests MH370 flew as low as 10,000ft

    2- Low altitude flight or even if MH370 did climb back up to approx 35000ft the fuel exhaustion would be a lot greater, therefore distance traveled is less..But the time remains the same to fuel starvation..

    I am correct in assuming that? Please correct me if I’m wrong..

  14. @Dr. Bobby: speaking of “high water marks”: since August, 2014, I’ve considered your contributions re: statistical considerations when fitting models to have been particularly invaluable.

    To wit: good comments on the residuals analysis. I agree: even if only 5 of 5 residuals have the same sign, that’s a 15-in-16 chance that it’s your model, and not random chance, that’s causing it. And noticeable slopes are just that much worse – red flag territory, I’m afraid.

    Aside: I’m surprised you shifted your 84E impact point so far east last year. Since you first suggested 84E 2 years ago, only two things have really changed:

    1) the towfish scans have falsified most COMPETING theories (but not yours; they still stop at E85, and turn around) &

    2) the ATSB has clearly liberalized their fuel limits, thus eroding one of the louder original objections to your path

    I do recall that you had rational reasons for, quite literally, coming back to the pack. But does any part of you now regret not sticking to your guns?

  15. @Brock

    Thanks, points taken. I wasn’t welded to the idea, just threw it in for “plausibility testing”, which you have debunked.

    I note your comments to Bobby re further SW beyond the search area.

  16. As per Victor Iannello, ScD,
    June 25, 2016
    have read the report and my comments below
    1. Assume that 23:41 was the place to consider a search. As compared to debris found, the south equitorial currents would only be effected if the incident occurred somewhere 20 degrees latitude.
    2. Beside that, the regular interval of signal was until 23:41 hrs.
    My estimate latitude 20 to 22 South

  17. @Dr. Bobby Ulich & VictorI

    First, it is not impossible to imagine the initiation of descent near FMT for purposes of landing, e.g. @ WITT Banda Aceh. Second, is not a RoD = -100fpm a very unique numerical value, namely the _minimum_ RoD a pilot may enter into the FMC ? So, is it not possible, that a pilot, on the verge of incapacitation around 18:40, tried to turn the a/c southwards and initiate descent, but succumbed to hypoxia and/or hypothermia with his thumb on the console ??

    Regarding BFO-implied RoC @ 18:25, an engine-unsustainable “zoom climb” of 5000fpm (declining over the next couple of minutes) could account for all BFO values on N571 towards IGOGU, and (naively) appears to be an effective method of decelerating the a/c, from essentially maximum airframe speed along the military radar track, to more plausible 800-ish kph speeds most compatible with the FMT and southwards ghost-flight afterwards, e.g. Inmarsat JoN article. I understand Dr. Ulich has estimated a climb @ 18:25 of +5000′, I integrated the implied BFO-to-RoC figures and computed +7000′ from 18:25-18:28. That would put the a/c up around 40,000′, an altitude value you both appear to be discussing.

  18. I wish to add, that the waypoints (IGOGU-)ANOKO-BEDAX are standard landing waypoints for WITT airport. If the a/c were at very high altitude, then the subsequent BTO to ISBIX @ 19:41 +/-, on a heading of ~181 degrees, would be very near the measured value. Alternatively, perhaps the pilot could have somehow entered a non-standard waypoint (e.g. on the equator) more-than-less due south of BEDAX, heading 180 ?

  19. @VictorI
    nice find!

    Just one remark : if you assume that there was a descent of -2600fpm at 18:40, then there is no way of discriminating between a southern flight path vs northern, as there could also have been a descent rate of -2600fpm (or even less required) at 19:41 and so on.

    If there was indeed a descent rate at 18:40 and if we assume that the plane disappeared from radar then re-appeared due to descent and climb, don’t we have a beginning of a pattern?

    The plane would seem to have flown on a phugoid pattern. Could be explained by a non trained pilot in command? Somehow the autopilot gets disconnected, the hijacker would be pilot now has to fly manually and can’t flight her straight and level? There are many scenarios that could fit this observation.

    Any thoughts?

  20. According to @)Oleksandrs diagram posted here before by @Gysbreght, a steep descent occured during a turn to the north after 18:28:14 with the descent continuing at least till 18:40:56 heading south and probably descending further when the SATCOM call ended.
    The average descent rate was taken at ~600ft/min in accordance with an ’emergency descent’ procedure for B737.

    Now for me it’s not possible to judge @Oleksandrs figures on absolute correctness but assuming they are fairly correct I see some possibilities to assume explanations on some points @Victorl questioned.

    First I see two possible scenarios for a descent starting gradualy after 18:27:08 turning into a steep (emergency) descent some where after 18:28:14.

    scenario 1-the plane recieved serious damage after 18:22 which decompressed the cabin forcing it to save altitude.

    scenario 2-after leaving Malaysian radar range the pilot turned and descended steep to dive under Indonesian radar horizon to stay undetected when crossing Indonesian air space on it’s way to the IO.

    In scenario 1 the plane would be restricted in altitude and speed due to the decompressed cabin thus burning more fuel. Maybe damage on the plane could have limited possible speeds and increased fuel consumption even more.
    Resulting in a more northern crash zone.

    In scenario 2 the plane would have climbed again after leaving Indonesian air space.
    A -100ft/min descent rate could have been set after BEDAX as @Victorl proposed with the goal to end up at a (calculated) certain location on a certain altitude.
    If the goal was to hide the plane, latitude 31.5S happens to be just above the 5000m deep Diamantina trench and near the more the 7000m deep Dordrecht Hole.

    -When assuming human input after FMT and BEDAX (as @Victorl does IMO) there also would be control over the amount of fuel.
    A surplus of fuel could easily have been dumped before reaching destination.

    -Considering the found debris and especialy the latest one (outboard flap) it has become more evident to take a controlled glide and an attempt to ditch the plane into serious account in every scenario IMO.

  21. @Gysbreght

    No problem. Its sometimes hard to follow conversations on unthreaded forums I know. And I share your frustration with debunked theories being raised time again.

  22. @Dr Bobby Ulich: Comments below
    1. Yes, I agree that fuel calculations are important. However, they are not as easy to do as for flights at LRC speeds and high altitudes because of the limited tabular data available. (I have had limited success in developing a more general fuel flow model that would allow a computation for any altitude, speed, weight, temperature combination.) As for there being insufficient fuel to reach 40S when temperature and PDA effects are included, I have been saying that for some time.
    2. The BFO errors are not all the same sign. I simply reported the magnitudes. On the other hand, if the oscillator drifted in the manner presented in the DSTG report (see Fig 5.4), it is possible to have large errors all of the same sign. You are trying to extract precision from the BFO data that is not possible. @DennisW has been saying this for some time based on his extensive experience with precision oscillators.
    3. I suggest we put the issue of flight data from the pilot’s computer aside for now. However, I advise you to consider the implications of those reports being true.
    4. If the altitude profile was as erratic as some of the data indicates, we simply don’t know the start and ending altitudes for the descent at 18:40. A descent of -2600 fpm is not extreme, and could have occurred for a significant length of time (minutes). We also don’t know the combination of speed, track, and vertical speed that produced the BFO value. The important takeaway is that a trajectory to the northwest at 18:40 followed by a later turn to the south is possible.
    5. There is absolutely nothing “unusual” or “unorthodox” about a M0.84/310 speed profile. In fact, it is one of the most common. With the autothrust in speed (SPD) mode, simply set the Mach number and during the descent the aircraft will automatically transition to 310 KIAS at the cross-over flight level. (The value of 310 KIAS is built-in.)
    6. In my opinion, it is unlikely that the path captured by radar is explained by a pilot trying to avoid land.
    7. When I first analyzed the radar data in detail, I found that the timing of the start and end of radar capture segments indicated an erratic speed, which I attributed to timing synchronization issues and/or position inaccuracy, and I assigned the path to M0.84 and FL340, which matched the overall path and timing. However, the DSTG report was very specific in showing that there was a significant variation in speed along the path caught by radar, including a large dip in speed near the turn at IGARI that has the features of a chandelle manoeuver. I now believe the path was not at constant level and speed, and my guess is the authorities agree. I have tried in vain to get the raw radar data, or at least the radar data supplied to the DSTG by Malaysia, to better understand the nature of the erratic flight.

  23. @Brock McEwen:
    1. I have not yet completed the fuel calculations, and until I have, I don’t know whether there would have been excess or insufficient fuel. It is not just a matter of comparing the path to the maximum range. And without knowing the speed/altitude combination chosen for the ATSB fuel calculations, I can’t comment on their results other than to say with Boeing’s help, the calculations were probably performed correctly for the assumed conditions.
    2. Yes, what has been presented in public regarding the flight to the SIO from the pilot’s computer has been fuzzy. I hope this will be made clear in the near future. My guess is that will be a bifurcation between those that believe the evidence implicates the pilot, and those that believe evidence was created to implicate the pilot. But in any case, whatever evidence was collected (or not collected) needs to be made public for a fair vetting.
    3. I am not against considering crash points further north on the 7th arc. I am less inclined to consider crash points off the 7th arc. I won’t dismiss that possibility, but it would imply deliberate corruption of the data, which in my opinion makes it less probable.
    4. Regarding lack of debris on the shores of WA for a crash near 31S, I will leave it to others like you, Richard Godfrey, and Henrik Rydberg to debate. I have not studied this in sufficient detail to have an informed opinion.
    5. I don’t know what would motivate the pilot to program a 100 fpm descent. I also don’t know what would motivate a pilot to fly into oblivion in the SIO. At the same time, I believe a technical failure cannot explain the event sequence of the disappearance.

    You are obviously very willing to believe that the US, UK, and Australia fabricated evidence to cover-up the true sequence of events. I can’t dismiss this possibility, but I think a scenario in which the pilot hijacked the aircraft to be more likely. I am not accusing and convicting the pilot, but he is a prime suspect. I think it is misguided to apply legal standards of guilt or innocence when thinking through the logic of possible scenarios.

  24. @AM2: The trajectory before 19:41 could be modified to fit Kate Tee’s observation. I have deliberately avoided doing this because it is not clear what she saw.

  25. @Sinux: Yes, a phugoid pattern could explain climbs and descents. But I have a hard time believing this occurred for hours. For me, the sequence of events suggests deliberate intervention, not a technical failure.

Comments are closed.