FlyDubai B737 accident report at Rostov throws up more questions than it answers

Flight profile
Flight pattern over Rostov

The final report of the Fly Dubai B737 accident at Rostov 19 Mar 2016 has been released. The report is primarily focussed on the go around procedure and what transpired thereof.

There is a strong possibility that the crew were fatigued and committed to land. The Captain could have switched off or not used the head up display (HUD) during the second approach in order to disregard a possible predictive wind shear warning which had prevented them from landing on the 1st approach.

The crew had spent over 6 hrs in the air including 2 hrs in the hold. A diversion would have been a sort of failure to complete the task therefore the crew could have decided to take the risk of attempting to land despite the adverse weather conditions prevailing at the airport at Rostov.


During the first approach (RWY22) the crew relayed the presence of “wind-shear” on final to the ATC (as per the aboard wind-shear warning system activation). From the altitude of 1080 ft (330 m) above runway level performed go-around.

Further on, the flight proceeded at the holding area, first at FL080, then at FL150.

After almost 2 hours of holding, the crew requested descent for another approach. It was an ILS approach. The A/P was disengaged by the crew at the altitude of 1900 ft , and the A/T at the altitude of 1700 ft.

In the process of another approach the crew made the decision to initiate go-around, from the altitude of 830 ft above the runway level.

After reaching the altitude of 3350 ft above the runway the aircraft transitioned to a steep descent and impacted the ground with a pitch of about 50⁰ and IAS about 340 kt at ~0040hrs on 19th March 2016.

The investigation report is very comprehensive but skips certain key issues as a result does not provide the reader with the data to form their own opinion. The report throws up more questions than the answers it is able to provide.

The final report can be accessed here


The TAF report of the Rostov-on-Don destination aerodrome of 18.03.2016:

TAF URRR 181401Z 1815/1915 25007G13MPS 9999 SCTC010 SCT020CB TEMPO 1815/1821 25012G18MPS 3000 -SHRA BR SCT005 BKN020CB FM182100 26010G17MPS 3000 –SHRA BR SCT005 BKN020CB TEMPO 1821/1906 28017G25MPS 1000 SHRA BR SCT003 BKN020CB TEMPO 1906/1915 30011G17MPS.

Information Bravo of 00:02: «…wind 230⁰ 11 m/s gusts 14 m/s, visibility 7 km, light shower rain, cloud: scattered at 570 m, broken cumulonimbus at 1200 m, temperature 6⁰, dew point 3⁰, QFE 741 mm/988 hPa, QNH 998 hPa, moderate turbulence from GL up to 1000 m, moderate icing in cloud from 900 up to 1500 m; temporarily: surface wind 250⁰ 15 m/s gusts 20 m/s, visibility 1000 m, shower rain, mist, cloud scattered at 90 m, broken cumulonimbus at 600 m…».

At 00:20 the tower control unit officer relayed the weather information to the aircraft:

«Weather as of 00:20: visibility 5 km, cloud base 630 m, wind 230⁰ 13 m/s gusts 18 m/s, light shower rain, mist, on final the severe turbulence and moderate windshear».

At 00:22 the approach control unit officer relayed to the aircraft: «…as of 22…230-14 gusts 18, visibility 6 km, scattered 480 m, correction 630 m, meteorological office is not reported about windshear on the RWY».


  1. The fuel data of the flight is not available anywhere in the report. The only mention is that the flight was planned with adequate fuel to hold and cater for diversion. The flight held at Rostov for little over 02 hrs. How much was the fuel on board when the final approach was initiated? From the cognitive viewpoint this could play a vital role.
  2. Why did the flight hold for 02 hrs in probably heavy icing conditions, over Rostov whereas the alternate Mineralnye Vody was at a flight time of 44min. The first approach at Rostov was discontinued due to wind-shear and the weather was similar till the final approach was attempted the second time.
  3. After the first approach due to wind-shear and holding for 02 hrs, the second approach was attempted. With similar wind conditions, why did the Captain disconnect the auto-pilot at 1900ft and Auto Thrust at 1700ft above aerodrome?
  4. On the second approach, why did the PIC carry out a go-around, since there was no wind-shear warning unlike the first approach. Landing gear and flaps were retracted as per a normal go-around procedure?
  5. On the second go-around why did the PIC not manually trim the aircraft using manual STAB trim for almost 45 seconds, to counter for the pitch up movement due to high engine thrust?
  6. Why does the FDR data show a mismatch between the aircraft pitch flown and the flight director indication?
  7. Why did the crew fail to monitor the rapidly increasing speed?
  8. The report concludes that the potential occurrence of the somatogravic “pitch-up illusion” did not have crucial importance as far as the onset of the accident situation is concerned.
  9. Last and not the least, why is the data released in snippets, especially the CVR transcript.

mindFly analysis

2nd Approach Pitch Vs FD guidance
1st approach go-around pitch vs FD guidance
  • The flight director(FD) guidance during the go-around was very shallow. The FD pitch was close to +5 degrees. The aircraft pitch flown was above the FD guidance. Is it normal for the FD to guide at 5degrees during a go-around.
  • During the 1st approach there is even a FD pitch command of -14.1 degrees during the go-around phase.
  • The HUD was serviceable and the Capt. was expected to fly the guidance provided by the HUD.
  • The recreation of the HUD for the 1st and 2nd approach can be seen below.
Recreated HUD for the 1st approach wind shear warning
1st approach Green Dot wind-shear guidance and the dashed TO/Ga pitch target line
Recreated HUD for the 2nd approach/ Go-around.
  • Refer to the 2nd approach HUD recreation. The pitch is just below the target pitch line of ~19degrees. At this stage the FDR FD command is +0.7degrees.
  • Why was there a discrepancy between the HUD command and primary flight display FD pitch command?
  • After the 1st go-around the Capt commented that he had the runway the runway in sight and could control the aircraft but for the predictive wind-shear warning, they had to go-around.
  • It was unreasonable, from the point of view of the PIC, not to take the chance to perform another approach. Particularly given that to his mind he could have completed even the first approach if «GO-AROUND, WINDSHEAR AHEAD» warning were not activated and brought him to make the decision to perform go-around.
  • There could be a possibility that the Captain did not use his HUD for the 2nd approach where in he had decided to disregard the predictive wind-shear warning if it came again.
  • During the 2nd go-around , the pitch attitude is closer to the FD guidance and very shallow. The first officer gives a call to maintain 15degrees is when the Capt pulled up the nose and reached a max of 18.5 degrees before returning to a lower pitch angle.
  • In order to maintain the shallow pitch angle with high thrust and low aircraft weight, the Capt decided to use the STAB nose down trim and gave a long burst of 12 seconds.
  • STAB is a very large moving surface and with flaps selected it moves at a higher speed. The extreme nose down was cause due to the excess nose down trim after almost 40 seconds into the go around initiation.
  • An approach flown without automation of auto pilot and auto thrust especially in adverse weather cannot be explained.
  • Longitudinal acceleration was experienced when the throttles were set to TOGA. Somatogravic illusion should have set in during this phase but there is no apparent nose down pitch down movement.

An operational decision to divert and wait on ground for weather improvement could have been decision instead of holding for 2 hrs in air. This puts additional pressure on the crew to complete the task.

The role of the pilot monitoring is not clearly defined by the Operator/Manufacturer. Therefore its left to their experience and presence of mind to figure out whats happening and assist the pilot flying with call outs and advice.

About Capt. Amit Singh

I think therefore I am Airlines Operations and Safety balance expert. A former head of operations/training and safety of successful LCC's in India. An experienced member of the startup teams of these airlines has hands-on experience in establishing airlines systems and processes.

2 Responses

  1. Vilas Shinde

    B737 with full power requires effort to prevent excessive pitch up. Here the attitude initially is shallow and very erratic and then with full elevator and excessive stab trimming is in a steep dive from which recovery was not possible. It’s total disorientation triggering primal emotion. Where the result of pitch input was not checked on the PFD or HUD. Something like AF447 but opposite. The GA was triggered by excessive speed at 900ft as unstable approach. The Tartarstan 737 crash is similar.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.