B738, Pristina Kosovo, 2016

B738, Pristina Kosovo, 2016


On 2 May 2016, a Boeing 737-800 veered off the 2,500 metre-long landing runway near its end at speed following a night non-precision approach flown by the Captain. It then stopped on grass having sustained damage to both the left engine and landing gear. The Investigation noted that a significant but allowable tailwind component had been present at touchdown and found that the approach had been unstable, the approach and touchdown speeds excessive and that touchdown had occurred beyond the touchdown zone after applicable operating procedures had been comprehensively ignored in the presence of a steep authority and experience gradient.

Event Details
Event Type
Flight Conditions
Flight Details
Type of Flight
Public Transport (Passenger)
Flight Origin
Intended Destination
Take-off Commenced
Flight Airborne
Flight Completed
Phase of Flight
Location - Airport
Approach not stabilised, Copilot less than 500 hours on Type, Landing Flare Difficulty
Authority Gradient, Procedural non compliance
Intentional Veer Off Runway, Excessive Airspeed, Late Touchdown, Significant Tailwind Component
Damage or injury
Aircraft damage
Non-aircraft damage
Non-occupant Casualties
Off Airport Landing
Causal Factor Group(s)
Aircraft Operation
Safety Recommendation(s)
None Made
Investigation Type


On 2 May 2016, a Boeing 737-800, (TC-JFY) being operated by THY Turkish Airlines on a scheduled international passenger flight from Istanbul to Pristina as THY97A did not touch down until almost the middle of the destination landing runway after an approach conducted in night VMC. As the end of the runway neared, the aircraft was intentionally steered off the side of the runway onto soft ground and eventually came to a stop. The left engine and both main landing gear assemblies were damaged but there were no injuries to the 157 occupants. 


A Serious Incident Investigation was carried out by the Kosovo Aeronautical Accident and Incident Investigation Commission (AAIIC). Relevant data from both the SSFDR and the SSCVR were successfully downloaded. It was noted that the 48 year-old Captain, who was PF for the flight, had a total of 11,099 flying hours of which 8,090 hours were on type and the 30 year-old First Officer had a total of 615 flying hours of which 335 hours were on type. 

What Happened

The flight crew were operating the first flight of their duty. It was found that there had been no proper approach briefing from the Captain. It was established that Pristina Radar had provided radar vectors to enable the flight to establish on the VOR/DME P offset approach to runway 35. This requires a 3.47° nominal descent angle if flown as a CDA and the track is offset by 8° until approximately 0.5 nm prior to the runway threshold. The approach was commenced with the AP engaged and with VS mode selected. After transfer to TWR, the landing clearance given was accompanied by a wind check of 150°/6 knots with advice that the 2,501 metre-long landing runway was wet. The ELW was 64,515 kg which meant that the applicable VREF was 141KIAS with landing flap 40° and autobrake 3 selected. 

As the aircraft passed 1000 feet aal, the FDR recorded the rate of descent was 1,900 fpm and the flight path angle was 5.5°. Nearing the MDA with the AP still engaged, the VS mode changed to ALT ACQ and the vertical speed began to reduce. The AP was disconnected at a recorded 815 feet aal and the rate of descent increased in order to regain the procedure vertical profile. This increased rate of descent was still present as the aircraft reached 180 feet agl which led to a two second EGPWS ‘SINK RATE’ activation. This was followed by a significant reduction in both vertical speed and pitch attitude.

The runway 35 threshold was subsequently crossed at 80 feet agl and the flare was initiated at 20 feet agl but it was still a further five seconds before touchdown followed at VREF + 4 knots approximately 1,030 metres beyond the runway threshold leaving approximately 1,500 metres of runway ahead. Autobrake was activated immediately and reverse thrust was selected without delay, although after about 600 metres, manual braking was commenced.

During the ground roll, both pilots reported “noticing that the aircraft did not decelerate as much as expected” and as the end of the runway got nearer, the Captain attempted, despite a ground speed of around 50 knots, to begin the left turn onto taxiway ‘C’, the final exit point at the end of the runway but this resulted in the aircraft crossing the taxiway near its junction with the runway and continuing onto the adjacent grass. Reverse thrust was cancelled as the aircraft left the runway and it subsequently came to a stop approximately 36 metres to the left of the runway - see the ground track illustration below. Although there were no injuries to those on board, six passengers asked for medical assistance “due to disturbance and fear”

As a result of the landing gear sinking into “mud and wet grass” following the excursion, the left engine was subsequently found to have sustained enough  damage - to the nacelle, thrust reverser, fan and inlet cowls and bleed air cooler - to require its replacement. Both main landing gear wheel assemblies and brake units were also replaced.

The aircraft ground track during the landing roll. [Reproduced from the Official Report]


A number of observations were made in respect of the conduct of the approach and landing:

  • The non precision approach to runway 35, although requiring less track miles than positioning for and completing an ILS approach to runway 17, involved accepting a significant tailwind approach onto a wet runway, albeit circumstances which were within the required landing performance criteria. It was noted that the Captain’s decision to use runway 35 had not been preceded by any assessment of the relative risk represented by these two alternatives.
  • Although a constant tailwind component was present throughout the approach, it was found from the CVR data that there had been no discussion of this on the flight deck.
  • The required landing distance on the Pristina runway taken from the operator’s 737-800 QRH whether dry or ‘wet but with good braking action’ (the latter was the condition measured shortly after the excursion) was 2,065 metres. It was raining prior to and during the landing.
  • Despite both pilots being aware that they were above the procedure vertical profile, the Captain did not take the opportunity to discontinue the unstabilised approach at or prior to the MAPt or when the excessive rate of descent below the MDA led to an EGPWS ‘SINK RATE’ caution being annunciated. 
  • CVR evidence indicated that CRM during the approach had been poor overall as well as not in accordance with the aircraft operator’s SOPs with the Captain’s conduct of the approach not eliciting any comment or adequate situational awareness prompting from the low experience First Officer in the presence of what appeared to be a steep authority gradient.
  • Applicable aircraft operator SOPs explicitly and independently required that a “long flare, as occured in this event,  should be avoided.  
  • The maximum speed at which the aircraft type was supposed to taxi off the runway other than when using an RET was 30 knots. 
  • The THY Turkish Airlines’ OM Part ‘A’ was found to have no clear guidance on the conduct of a go around if a touchdown cannot be made within the “first 1000 feet of the runway” as required. 

The Cause of this Serious Incident was formally recorded as the aircraft crossing the runway threshold above the standard crossing height combined with a long flare and a late touchdown beyond the touchdown zone. 

Two Contributory Factors were identified as:

  • Crew Resource Management was not performed according to the aircraft operators SOP. 
  • The flight crew didn´t initiate a timely go-around procedure. 

The Final Report was completed in July 2018 and subsequently released in a single three-language format, the definitive English language version together with translations into Albanian and Croatian.  

Related Articles

SKYbrary Partners:

Safety knowledge contributed by: