Turkish MD-83 Crash

[lomapaseo]

All the standard answers apply. Training, CRM, equipment upgrades (EGPWS), double checking your sums, duty time/rest, communication...



I don't agree. Most of the above are precursors to the crew error not additional barriers to an accident following common crew errors.

[pls8xx]

While I am still struggling to pinpoint the crash site, there is one thing I am rather certain about ... the heading at time of impact was generally NE, parallel to the runway.

One of the photos made on the morning of Nov. 30 had the camera pointed parallel with the crash scene. In it one can see shadows near 90 degrees to the camera angle. The shadows are 2 to 2.5 times the height of objects.

Using the solar calculator at .... www.srrb.noaa.gov/highlights/sunrise/azel.html (http://www.srrb.noaa.gov/highlights/sunrise/azel.html)


By trial and error I adjust input to get a solar elevation that will yield shadows of the length seen in the photo. This occurs about 9:45 AM. The azimuth of the sun at that time was about 149 degrees.

http://i9.photobucket.com/albums/a93/pls8xx/solar.jpg

Turning back to the photo, one can see a high point along the horizon in the upper left that is near alignment with the crash scene. Google shows a peak about 25 mi at around az 54 degrees from the general area of the crash.

http://i9.photobucket.com/albums/a93/pls8xx/heading.jpg?t=1197476726

From info on radio communications, gear down comments, an altitude less than 7000, and the flight path derived from the above analysis, I conclude that it's likely the crew thought they were on final approach.

[PBL]

Now that, pls8xx, is class! Very nice.

PBL

[Earl]

Quote:lomapaseo
Oh, and just what critical question did you expect to get answered only from the black box ?

Earl: CVR should have added quite a bit of information as to what happened.
Approach briefing to begin with, who was the PF,PNF, EGPWS warnings and what actions were taken, CRM, standard call outs, many things that would have helped this investigation, if true that they were disoriented this would be more than apparent on the recording.
This was an important piece of the investigation.
DFDR would have backed up much of this info.
Really surprising question here!
Care to expand on this comment?
Or just more Media!

[Green Guard]

Hmmmm. 18 minute racetrack, lawsuit filed, empty CVR, incomplete FDR data. Any chance the logbook with any MEL's will apppear?
You must be kidding !!
Just think why was CVR "empty" and FDR "incomplete"" !! :{

[Machaca]

Nice work pls8xx!

Using both on-site photos looking out and many photos of scenery in the surrounding locales, I used roads, terrain, vegetation, landmarks, reports, etc. to deduce Türbetepe's location and the point of impact to be:

LAT 37.873 -- LON 30.200

http://img89.imageshack.us/img89/9164/4203turbetepege01cy6.th.jpg (http://img89.imageshack.us/my.php?image=4203turbetepege01cy6.jpg)

Here's a view looking West over IPT to Türbetepe:

http://img89.imageshack.us/img89/9375/4203iptturbetepe1sl5.th.jpg (http://img89.imageshack.us/my.php?image=4203iptturbetepe1sl5.jpg)

Here's a view looking South to Türbetepe from above the town of Aydoğmuş:

http://img443.imageshack.us/img443/948/lookingsouth2212700emk3.th.jpg (http://img443.imageshack.us/my.php?image=lookingsouth2212700emk3.jpg)

Here's a flyover view of the area -- crash site area is obscured by the inboard canoe!

http://img150.imageshack.us/img150/5051/3910147bai8.th.jpg (http://img150.imageshack.us/my.php?image=3910147bai8.jpg)

From atop Türbetepe you can easily look about in Google Earth and match up the views seen in many of the photos taken at the crash site.


-=MachacA=-

[PBL]

Or just more Media!

No earl, he is not, as you could check for yourself.

Let me take this opportunity to say that I find these sorts of prejudiced comments unnecessary. Much of the information any of us have about this accident has been gleaned through the hard work of journalists. I prefer to appreciate this necessary work, not to demean it.

PBL

[MaxBlow]

www.airporthaber.com/hb/detay.php?id=22559

here're more pictures/charts showing the track flown by the aircraft.
Maybe someone can translate the text.

[hetfield]

@Max

What does the 2nd picture say about DFDR (red)?

[MaxBlow]

I'm not 100% sure but understand the following:
'Once analysed the DFDR will explain why the aircraft was flying this track'

[learsimmer]

If anyone reading this is familiar with flying in the next valley parallel to and to the West of Isparta (that's the valley that contains Cardak airfield and the town of Basmakci) could they say at what minimum altitudes an aircraft flying overhead, say Basmakci, could be expected to receive useable radio transmissions from Isparta VOR (IPT 117.50) or Isparta NDB (IPT 349)?

Nick

[hetfield]

@post 264 by Max

What does this picture say?

http://www.airporthaber.com/hb/pic.php?id=16838




- is it a guess by journalists?
- is it a radar plot?
- if so, did radar coverage stop after the aeroplane entered mountaneous terrain or after the crash?

Thx

[PBL]

The pictures from airporthaber.com are not necessarily that helpful. The first one shows the approach with a direct entry into holding from the northwest. The second shows an approximation to the actual track, with the annotation that the DFDR will give the exact data (hetfield: that is what it says).

They are the only representation I have seen that suggest the aircraft approaching IPT from the northwest. The EKSIN 1C arrival from UA-18 from the Afyon VOR comes in to IPT from the East, from intersection EKSIN on a heading of 258° on the 078° radial of IPT; the PELIL 1C arrival from CARDAK (CRD) on a heading of 080° on the CARDAK radial 080 and then the 260 radial of IPT from PELIL. There is no arrival from the northwest. If they had flow Afyon direct Isparta, as suggested by SeattlePilot, then they would have come in from the north heading about 184°, because that is where Afyon is. The magnetic variation at IPT is 3.2°E. One can see all that from the en-route chart snippet uploaded by MachacA, and the arrival and approach plates to which the links have been given in this forum.

The best info comes from the article from Hurriyet on 12 December, included by Machaca. It doesn't say where the aircraft came from, but it does say that they came in at 8,000 ft. After passing the VOR they followed a track of 253° until 12 nm out, then turned right. They also didn't stop at the inbound heading of 043°, but went a further 50° around to 093°. Then they hit Türbutepe at about the 1,500 m mark. There was a GPWS warning. At the "pull up" warning, they did indeed pull up and put in power, but the power application pitched the aircraft up and the tail struck.

Hurriyet says that this information comes from preliminary analysis of the FDR and CVR, and that these have been sent to Germany for further analysis.

So it looks as if it was CFIT, as many have thought.

The question here is: why 253°? The following was suggested to me in PM (by a participant who is a little fed up with certain participants chewing others out for not being professional pilots). This is close enough to the 258° heading required for the EKSIN 1C arrival. Perhaps they had loaded the heading for the EKSIN 1C arrival at or before Afyon, and then after they were cleared direct switched to that instrument upon station passage at IPT, without remembering or checking that this was the arrival heading they had loaded and not the outbound heading for the VOR/DME 05 approach. This is pure speculation, though.

Thanks very much to my student Tülay Inci who translated the text for me today in conversation.

PBL

[zeynep737]

This accident showed us that the Turkish government is still inadequate about solving these kind of disasters..ı m a Turkish citizen and ı can watch what our authorized people r doing here..this is the 3. plane crash in Turkey since 1998,and in all accidents the reason of the crashes were explained on the day of event but in this case noone can explain and ı still cant find out why?ı m not an aviation authority but its not just a pilot fault,there r missing parts..first they said the CVR was empty and now they say CVR is being solved(?)There were 6 Turkish nuclear physicians aboard from Boğaziçi University and Doğuş University-going to Isparta to a symposium-they were working on a material named torium,when these element is used in Turkey in nuclear stations than our country will never need petrol oil,and while our external debt is 450 billion $,with the usage of this element our country would have 70 trillion $ source...With this invent our country would be one of the most biggest energy source but somehow it was stopped(!) the cause of this accident cant be just because of pilot fault...

[hetfield]

the cause of this accident cant be just because of pilot fault...

I'm afraid, it can...

[zeynep737]

ı hope this accident can be explained as soon as possible..it was pilots first flight to Isparta which means that they didnt know the area(thats why therer r landing charts but ı guess the crew trusted themselves a lot..)

[LuckyStrike]

- It is a clear night
- "Runway in sight" reported

These are the established facts that I have drawn out from various posts here...

Now if that's the case how would they end up in those mountains as there is no limitation for the pilots to see the runway after assuming their crash heading? They would be following the runway as they were making the turn to complete the racetrack one thinks...

As far as I understood from what's plotted on the charts on MaxBlow's post, they indeed flew on the 223 heading but then, for some unknown reason, instead of turning to 43 another heading was chosen which is <43 degrees and from the plots you could see the path flown is not a 180 degree right turn.

Based on these I would personally assume that from their entrance point to the racetrack (which should be <223 degrees) took them over to the mountains on the magnetic S. Having this fact in mind, they then extended their turn in order to cut runway centerline assuming still flying on the lake. Thus confidently they started the descent but ended up unfortunately in the mountains.

I am looking the charts and see the crash site between 4000 and 5000'. Looking at D5 IPT, it shows 4560'. Maybe a longshot for many but for me it is mind bothering...

However there is a major flaw in this theory: If it was possible to follow the runway lights as there was clear skies, one would definitely observe the deviations and take corrective/evasive actions. This I can't explain... :ugh:

[Belgique]

An initial investigation showed that pilot error caused the Nov. 30 crash of an Atlasjet MD83 that killed all 57 people on board, Turkish news media reported Tuesday. The wreckage was found seven miles from the airport, on a mountain around 5,000 feet high.
.
Reflect upon the "black hole" effect and the near side of intervening hills. It is possible to be seeing quite clearly the runway lights from 0.7 to 77+ miles away on a night with good visibility. However, very much "unseen" is the unlit hill between you and the airfield that increasingly juts blackly up into your flight-path. Airplanes don't necessarily go where they're pointed...... an “into wind” or underpowered flight path can often fall parabolically short (which enables the dark unseen intermediate terrain to claim its victims). It's nothing to do with "spatial consciousness", it's all to do with unseen hazards and the perils inherent in eye-balled descents and NPA's (non-precision approaches).
.
So, all the evidence thus far would seem to indicate that the crew declared themselves to be "visual" with the airfield - and that they simply flew into the near-side of a darkened mount - as their too-steep/insufficiently powered arcing descent took them latterly beneath their visual sight-line. The only chance one ever has of seeing that very subtle development happening is if there are lights very obvious in the foreground disappearing as they are subtly obliterated by the darkened hill. However such lights can easily become: “out of sight/out of mind” and those very final developments can be quite sudden. It would be careless in the extreme ever to rely upon a night visual approach and discard any requirement to observe Minimum Sector Safe Altitudes just because you can see a distant destination airfield’s lights clearly in the distance.

[PBL]

LuckyStrike,
As far as I understood from what's plotted on the charts on MaxBlow's post, they indeed flew on the 223 heading but then, for some unknown reason, instead of turning to 43 another heading was chosen which is <43 degrees and from the plots you could see the path flown is not a 180 degree right turn.

Please read my post concerning the Hurriyet article. I don't believe the chart with track to which you referred is an accurate rendition. They apparently flew 253° outbound, not 223°.

However there is a major flaw in this theory: If it was possible to follow the runway lights as there was clear skies, one would definitely observe the deviations and take corrective/evasive actions. This I can't explain...

Aircraft don't have rear view mirrors. You can't see runway lights away from which you are flying.

BTW, it occurs to me that this progression of errors shows the value of the appropriate procedures: a parallel or teardrop entry to the holding pattern (depending on the exact direction from which they approached IPT) and *then* flying the 12 nm outbound leg of the approach from holding. The holding pattern is well protected and you get a chance to reconfirm outbound and inbound headings for the approach itself.

PBL

[hetfield]

The plane disappeared from radar screens shortly before it was due to land at Isparta

Extract from posting #1.

Was it Antalya Radar?
Did the plane disappear after it entered mountaneous terrain or after the crash?

[PBL]

The plane disappeared from radar screens shortly before it was due to land at Isparta ........
Was it Antalya Radar?
Did the plane disappear after it entered mountaneous terrain or after the crash?

I thought I had answered this above, but it seems I didn't. Hurriyet suggests that radar contact was lost as it descended below 8,000 ft approaching Isparta. I presume this is normal.

Since people seem to be thinking about the chart from airporthaber.com, it may be worth while to be specific about why it is likely to be misleading.

First, we do not know for sure that the aircraft approached Isparta direct from Afyon. Indeed, MachacA's chart snippet showed it on UW91, departing UW91 shortly before CRD to fly direct. So the airporthaber.com chart would be consistent with this direction of approaching Isparta.

But then the chart has him turning on the outbound course *before station passage*. Now, one supposes the runway lights were on so these might act as a reference point to anticipate the turn, but why would you anticipate the turn? You're going to have to pull a 180° procedure turn at the other end, so why do something that will misalign you on final?

But let's say he does that, and proceeds outbound on the inbound track as airporthaber.com shows. He then pulls a procedure turn to the *right* to get on the inbound track, and ends up 7nm or so away. We know he was 7 nm away from the inbound course.

Now, at 185kts max (as per the approach plate), let's say he was going 180 kts, 3nm per minute. Pulling a 180° turn takes you a minute, and you have gone half way around a circle. The circumference of your circle is two minutes in time; 6 nm. Let's say 6.28.. i.e. 2 times pi. Then your radius is about 1 nm and your diameter 2 nm, so you are 2 nm left of inbound course. How on earth could you get 7 nm away, even if you are sloppy with the turn?

So that's why I am more inclined to go with Hurriyet's 253°. But that explanation does have them coming from Afyon, not on UW91.

PBL

[LuckyStrike]

PBL,

Agreed on the explanations that you have given quoting my reply... In the first part I was also trying to stress that what is plotted on the charts just can't be reflecting what happened.

And just to clarify, I am well aware of no mirror situation.:E However I was referring to the inbound leg not the outbound; so you should be seeing the airport lights. (i.e. flying at 0<x<43°)

Well anyway FDR will solve the many questions in my mind...

[PBL]

LuckyStrike,

understood; thanks for the clarification.

This emphasises the point about proper procedures. Even if they had some wrong heading set, if they had entered holding first with a one-minute outbound after station passage and then a turn back inbound, as one is supposed to do, then they would have noticed visually that they were not aligned with the HIALS, which is almost aligned with the inbound course. At that point they likely would have been prompted to check and correct.

Still speculation, not confirmed by definitive course data, but looking likelier based on indirect info.

PBL

[pls8xx]

Sometimes it can be useful to take a theory and review it. If found to violate a known fact, it can then be eliminated as a possibility.

With that in mind, I'll take up the supposition that the outbound leg was flown in error at heading 253 instead of the correct 223.

In the graphic below I have moved the graphics from the chart, rotated 30 degrees to the 253 heading and scaled to the google image. Another copy of the raceway is then overlaid at the turn-back position with the correct heading shown.

We might now consider what the crew expected to see near the end of the turn. Some time has passed since the outbound error was made, and a referral to the chart may have induced the crew to pull out of the turn at the chart heading of 043. They would have expected to see the runway dead ahead.

On a heading of 043, the town of Keciborlu is straight ahead at about 12 miles. The 05 runway is 30 degrees right at 12 miles. As the turn is completed, if lights at Keciborlu are visible they will be seen before the lights at the runway.

http://i9.photobucket.com/albums/a93/pls8xx/TURN.jpg

A closer look at the line of sight to these two points might be helpful. There is also the possibility that some mist has formed over the higher elevations at this time of night. Though general conditions are clear, a low line of sight over peaks may be less than perfect.

In the graphic below I show the two lines of sight that I will profile.

http://i9.photobucket.com/albums/a93/pls8xx/PHOFILE.jpg

The profile below is that to the 05 runway. A grid was selected for vertical and horizontal scale. The ground line was taken from google earth data along a line with 1 mile way-points shown. The descent profile from the chart was imported and scaled to the graphic. The descent profile was refined from the data and shown as a light blue line.

Note that the line of site from the point of turnback is just over a ridge about 3 miles from the runway. The runway will be seen from an angle, not straight on, and very early in the descent they will be lost completely, as seen by the red line.

http://i9.photobucket.com/albums/a93/pls8xx/PROFILE1.jpg

The profile from the turnback to the town of Keciborlu is a bit more interesting. Once again the lights are seen over the tops of peaks though any mist that was there. High ground near the town blocked sight of the left part of town and created a hard straight edge to the lights on that side. Lower ground slightly to the right of profile allowed some of the lights to be seen. Once again as descent progressed, all the lights would have been lost (red line). Note also the crash site suggested by Machaca lies only 0.6 mile left of this profile and the way a standard descent pattern might intersect the ground line near this point if the crew tried to pull out. This scenario would also be compatible with my earlier conclusion that the heading at impact was NE and likely close to the chart heading for final.

http://i9.photobucket.com/albums/a93/pls8xx/PROFILE2.jpg

The question now becomes, is there a known fact that would eliminate this possibility. If not, one must remember that there may well be other possibilities that also would fit the facts now known.

[LuckyStrike]

Please check the aerial photo and use hybrid setting (http://www.indexmundi.com/z/?lat=37.9413889&lon=30.3011111&t=p&r=16180&p=keciborlu&cc=tu&c=turkey)

pls8xx, that's real good piece of work you have done. However I can't hold myself down from asking this question:

After turn to final, Keciborlu lies straight ahead and along the line are 6 villages (Cukuroren, Kuyucak, Yesilyurt, Yenitepe, Senir, Kozluca together with the runway and also Keciborlu). Now, if you examine the aerial photo from a distance and the chart, you'll see that on the final leg you shouldn't be encountering any kind of light as you are flying over the lake. However if you fly the path to Keciborlu you'll be encountering more than one village with lights. That would give you a hint that you are not over the lake of course assuming no mist, no fog clear skies. At least to the co-pilot who was ex-AF...

Still, I must say I find your theory reasonably well organised.

[pls8xx]

LuckyStrike (http://www.pprune.org/forums/member.php?u=120281), you make a very good point about lights where there should only be black water. I would think this would apply even more so on the outbound leg. Could be the outbound on 253 is plain wrong. Still, I believe it was stated the pilots were new to the area?, and the villages appear to be small. I wonder how many lights there would be at that time of night. Then there is some water where you will see the light from a boat. A pilot's past experience might dictate whether an occasional light over what he thought to be water would raise a red flag or not.

A long day, late night, and the last runway in sight. Sometimes the mind just seems to shut down.

[piesupper]

Doubtful if there would be any boats out on that lake, especially at night. Lake Burdur is alkaline and I don't think has much in the way of fish in it. Its not the tourist season and anyway this is not a particularly tourist area so there are unlikely to be any pleasure boats out either.

[Towerman]

Can anyone confirm whether the Turkish authorities have introduced an age restriction on aircraft coming onto the Turkish register or operating from Turkey?

[MaxBlow]

German investigators confirmed that the CVR was u/s.
After the first techlog entry it should have been repaired within 72hrs / 3 sectors but has been flying for 9 days like this.

[Green Guard]

Can anyone confirm whether the Turkish authorities have introduced an age restriction on aircraft coming onto the Turkish register or operating from Turkey? may be a naive question. Can you find any country in world with younger MD83-s then in Turkey ?
You should rather think about MaxBlow post above.
:}

[MaxBlow]

CNN reported that the EGPWS (sent to Honeywell for inspection) has also been u/s.

[SeattlePilot]

There are also reports that the FMS was not up to date or hasn't been updated for ages.

[sinanerdem]

I totally agree with you. But, What I noticed so far is that, accident investigation reports are not reaching to the individuals and public without being destorted or manupulated. I hope this time it works as you expect.

[sinanerdem]

After Ankara radar transfered local control unit no radar facility. Pilots are expected to apply full procedures.

[Southernboy]

sinanerdem (http://www.pprune.org/forums/member.php?u=228114)

You say reports are distorted. Is there a report available - distorted or not -in English do you know?

[sinanerdem]

There is no report so far. I said it in general. But, By the time I reach any report, I'll share it with you.

[Southernboy]

That would be great thanks. This crash shows just how important full and open reporting is for all sides involved.

[BoeingMEL]

PBL, it pleases me to note that you have belatedly acknowledged CFIT as the most likely cause of this accident. (Exactly as some of us did over 3 months ago) May I respectfully remind you that you disparaged me when I took Skydrol Leak to task for his ludricous analysis (stall, low-level impact, engine-out etc) on November 30th and that you diagnosed "loss of control" in your post of December 4th? You suggested in an early post that those of us who suggested CFIT and dismissed some outrageous suggestions, should "have the grace to apologise." Looks like I won't have to eat humble pie for a while. Cheers, bm:ugh:

[punkalouver]

At this point we still don't know for sure what happened, so an open mind is of utmost importance. CFIT is one of many possibilities.

[PBL]

A message to me! What fun!

PBL, it pleases me to note that you have belatedly acknowledged CFIT as the most likely cause of this accident. (Exactly as some of us did over 3 months ago)

BoeingMEL, you'll get no credit from me simply for guessing right. Indeed had you said "It was CFIT" when first hearing the words "commercial airplane crash" you would historically have been right about, oh, 40% of the time IIRC.

May I respectfully remind you that you disparaged me when I took Skydrol Leak to task for his ludricous [sic] analysis (stall, low-level impact, engine-out etc) on November 30th and that you diagnosed "loss of control" in your post of December 4th?


Well, technically SL is still more or less right, and if you want to contrast your guess with his then I still think SL's was better. I think the wreckage suggests he stalled it in. The details of the track seem to show, though, that he pointed it at the mountain on what he thought to be an approach, which makes classifying it as CFIT more sensible than as LOC. It seems the radar tracks are all anyone's got, so I doubt we'll hear much more about this accident.

You suggested in an early post that those of us who suggested CFIT and dismissed some outrageous suggestions, should "have the grace to apologise."

Not because of your beliefs, obviously, but because you were unnecessarily rude. And I am glad to see you are still smarting from the reprimand. :)

PBL

[punkalouver]

The unfortunate thing is that people speculate when they have no idea what they are talking about.

[esenuslu]

At least the official report was submitted. It was kept secret, but today Turkish press reported it extensively.
a couple of days ago the Minister gave a press conference and claimed that the accident was "an ordinary CFIT." However the continued press digging in. The 361-page report laid bare serious misgivings such as:
1) Both CVR and FDR had not been working for eight days before the accident. And plane continued flying in that state!
2) EGPWS was faulty, the audible warning was not working, and it was recorded by the pilots before the flight!
3)The pilots' training and experience with MD83 were very limited, below standarad, and they were flying to Isparta for the first time.
4) The approach data was not keyed into the FMS
5) During final approach the pilots experienced a flap/slot problem.
6) That event created loss of spatial awareness and the pilots made a 30 dgr error on the final turn, heading 253 instead of 193, and flew into the mountain.

Let us press on to avoid yet another whitewash with "pilot error" pretext. This whole event was a corporate manslaughter/murder case where the liability goes high into the the state authorities that were supposed to oversee civil aviation.

[MaxBlow]

ensenuslu,

do you know if this report is also available in English ?
I could only find the Turkish version.

[esenuslu]

MaxBlow

The report was prepared in English. However I was not able to see the full report niether in Turkish nor in English.

The press reported that the Accident Report was submitted to the Public Prosecutor investigating the incident. So far as I see, the case is considered sub judice and the Accident Report is part of the evidence, so they did not disclose it.

Yet another example of lack of transperancy and public scrutiny over the civil aviation matters in Turkey.

If I somehow get a copy of the report, I will post it.

[esenuslu]

Today press reported that a copy of the Accident Report was delivered to one of the Defence Lawyers.

That Lawyer said that as the report was highly technical, the Defence team get it studied in detail before they give copies to the victims' relatives.

Maybe a real chance at the end to get hold of a copy...

[MaxBlow]

You can download the accident report in Turkish at airporthaber.com. I scrolled thru the pages understand the pictures but not the text.

[Storminnorm]

Ref post 281, why are German investigators looking into
the accident? Was it a leased aircraft?
Seems to me that there is a total lack of any responsibility
for the state of the aircraft maintenance that is a major
contributor to this awful combination of events.
It doesn't surprise me somehow. I wonder why?