Yeti Airlines flight 691 crash in Nepal

20 Jan 23 16 Comments

On the 15th of January 2023, Yeti Airlines flight 691, a scheduled domestic passenger flight from Kathmandu, crashed on approach to Pokhara, killing all 72 souls on board.

The aircraft was a fifteen-year-old ATR 72-500, registered 9N-ANC. The ATR 72 is a French twin-engined turboprop popular for short-haul services. The “72” refers to the standard seating capacity for 72 passengers. That day, there were 68 passengers and four crew on board. This is the deadliest ATR 72 accident since the aircraft was introduced in 1989.

Yeti Airlines ATR-72 Landing in Pokhara Airport (9N-ANC) photographed by TMLN123 in May 2022 CC BY-SA 4.0.

Yeti Airlines ATR-72 Landing in Pokhara Airport (9N-ANC) photographed by TMLN123 and released under CC BY-SA 4.0.

The aircraft departed Tribhuvan International Airport at Kathmandu at 10:33 local time for a scheduled flight to the resort town of Pokhara, 200 kilometres to the west.

Pokhara is a popular destination for tourists interested in the Annapurna Circuit, a hiking trail in the Himalayas, and the gorges of the Seti Gandaki River.

The flight crew consisted of a senior captain supporting a captain on her familiarisation flight to the new international airport. Yeti Airlines has confirmed that after this flight, she was to be cleared to perform as a solo captain. She was in the left-hand seat and in command of the flight. With over 6,000 flight hours, she was not, as some press reports would have it, an inexperienced first officer.

That day, the mountains were clear, visibility was good and there were no issues with the wind or weather.

Pokhara Regional International Airport (VNPR) had opened just 15 days before the crash. Previously, domestic flights from Kathmandu, Jomsom and Manang went to Pokhara Airport, established in 1958. The new airport has better facilities and a 2,400 metre (8,200 ft) concrete runway, twice the length of the old one, oriented to 120°/300° (12/30).

The land for the airport was acquired in 1976 but the project stalled until 2009. The foundation stone was placed in 2016 to mark the start of the building, with a plan to start operations in 2021. Various delays, including COVID issues, delayed the opening under much pressure. The official opening date was the 1st of January this year, despite the lack of a fuel depot: fuel was transported by road from the old Pokhara Airport to support the traffic.

Yeti Airlines flight 691 was inbound on runway 30, a straight-in approach from Kathmandu. However, the flight crew requested runway 12 during the final descent. ATC cleared the flight to land on runway 12 and cleared the aircraft for landing.

“We were not sure why,” a senior air traffic controller told the Kathmandu Post. “Permission was granted and accordingly, the aircraft started its descent.”

The Aviation Safety Network reports that the inbound flights from Kathmandu earlier that day and in fact almost every Yeti flight that year, had all landed at runway 30. One flight, YT677, landed on runway 12 on the 12th of January, flying north of the airport before turning left on base and final near the Pokhara VOR.

I can only presume the winds at Pokhara are very calm, as no one seems to have seen it as significant that the crew wanted to land from the opposite direction. Before the opening, an airport official explained that the new airport would have a “single approach”, with international flights would land on on runway 12 and domestic flights from “both sides,” that is, either runway 12 or 30. This is not a system that I have heard of before and I half wonder if the official was confused about how the runway functioned.

Google Maps screenshot showing both airports in Pokhara

There are two videos of the last moments of the aircraft. Both show detail of a tragic event, so I am going to advise viewer discretion in viewing these, especially the second. If you are reading this in your inbox, you may need to click through to the site to see the videos.

The first was taken with a smartphone by a local resident who enjoyed watching the inbound flights. Although he noticed that the aircraft was turning later than he had seen before, he thought it was a normal landing until the aircraft banked sharply before entering a stall.

“I saw that and I was shocked,” he said to local media. “I thought today everything will be finished here after it crashes, I will also be dead.”

There was no mayday call or reference to distress.

The second video, hosted on Facebook, shows the cabin of the aircraft. A group of four Indian friends were visiting Nepal and one laughingly filmed their descent, livestreaming it to his Facebook account. It is extremely disturbing to watch and hear the people moments before their death. However, the footage shows the wing through the window as they pass the football stadium and some have argued that this view shows the flaps at 15° instead of the 30° one would expect on final approach. I’m not convinced that there’s a clear enough footage to be able to tell. There’s no sound of increased power before the crash to signal an attempt to recover from the stall. After the initial impact, the phone lies among the flames and continues to stream for another half a minute.

I should note that The Aviation Herald had initially reported that the cabin video was a fake, based on a wrong understanding of how live streaming works, the belief that the view out the window showed an approach to the old airport and the mistaken belief that the phone with the footage was discovered intact at the crash site. Simon Hradecky has since posted that he may have seen a different version and that the actual cabin-video is credible. Multiple friends of the victim have confirmed that the video shows him and his four friends and quite frankly, this would be a difficult version to fake. Usually, fake footage is from a film or from another crash, which is clearly not the case here.

The crash site is on the bank of the Seti Gandaki River.

As of writing, 71 of the 72 on board have been recovered from the wreckage of the worst air crash in Nepal in three decades. The data recorders have been recovered and appear to be intact. Time Magazine reports that the Cockpit Voice Recorder will be analysed locally and the Flight Data Recorder has been sent to the Bureau d’Enquêtes et d’Analyses (BEA) for analysis. The BEA is also participating in the investigation representing the state of the design and manufacture of the aircraft.

The Civil Aviation Authority of Nepal (CAAN) is leading the investigation. Nepali airlines are banned in EU airspace as a result of concerns to regional training and maintenance standards. The Nepali National Aviation Safety Plan has already designated Aircraft Accident and Incident Investigation as the area with the utmost priority, following the most recent ICAO USOAP audit. Thus, there will be serious pressure on this investigation to supply a thorough final report.

Kathmandu Post reports that although Nepal adopted the Montreal Convention in 2018, which establishes airline liability in the case of death or injury to passengers, the bill establishing a minimum compensation for $100,000 has been stalled for over three years. The current minimum compensation for a passenger death on a domestic flight is $20,000.


  • A tragedy, but we will not know what exactly caused this crash until more data are released.
    I have flown the ATR42 (the -72 was just about to go into production) and found it a delightful aircraft to fly.
    Of course, the 72 is a larger version but does that affect the handling? I don’t know of course. What I do seem to remember is that the 70 passenger version of the BAC 1-11 was a lot easier to handle, especially with respect to landing it. Very positive touch-downs were common on the -500 series. But that probably does not apply to the ATR.

  • With phone cameras and live-streaming becoming increasingly popular, I expect to see many more of these sad and disturbing video clips.

    The tragedy of this is that we may be tempted to seek them out just for their sensationalism and shock value.

    The blessing is that some of them may provide vital clues to the ensuing investigation.

  • I wonder how difficult it is to assess the value of a pilot’s experience, as measured in flight hours, in that kind of extreme mountain environment. How much does safe flying depend on extremely specific local knowledge of each airport and its approaches, rather than overall flying skill and time in the cockpit?

    If the “local resident” is correct in saying that the aircraft turned later than usual and then banked sharply, it suggests that the crew missed the normal turning point and then over-corrected (Get-There-Itis?). However, that is just speculation based on one person’s subjective impression, so it will probably be wrong.

    • An overtight 2nd turn falling into a stall is an obvious answer — do flight recorders capture enough detail to show this? — but the cause is more uncertain; if they hadn’t made that tight turn, would they have been too close to surrounding terrain? I’ve read here and elsewhere of how difficult the mountains make approaches in Nepal.
      The pilot was a Nepali (who went into flying after her husband crashed, according to some stories); I would expect her to have had a lot of experience with the landscape in general. However, this was a “familiarization” flight for her and the airport was only recently opened (good detail — I hadn’t seen this in any other story), suggesting that she wouldn’t have known any visual markers that would help define where to start a turn and that even the supervisor wouldn’t have been very familiar with them. In such cases a localizer (horizontal component of ILS) \might// be helpful — but there’s been a report that the ILS wasn’t on that day.

      This has been a bad week for aviation, between this and the runway incursion (fortunately without a collision) at JFK.

  • Is it possible that the runway 12 request was for training/checkride purposes. If its a visual approach, downwind, base etc will look very different based on whether you’re shooting 30 or 12. So 12 might be a harder/unfamiliar approach than 12 given that 30 was the norm. Maybe the training captain wanted to throw in a challenge and asked her to do 12 instead?

    • We will get to know when the voice recordings are published, but this is something which I deem most probable.

    • It is, except that it was a little bit late for such a request if it were part of the training. But that should hopefully come clear from the conversation in the cockpit.

  • As the satellite picture of the airport shows, only the Western end of the runway has a parallel taxiway. Therefore, it makes sense that heavy aircraft land on 30 and take off on 12. Consequently, the ILS is only to be installed on 30:

    Runway 12 has Non Precision Approach (NPA) and Runway 30 has ILS System Precision Approach CAT-1 (PA1) facility.

    I believe that’s the reason for the operations’ lack of symmetry.

  • I’ve said it before and I’ll say it again!. Hundreds of single engine pilots over the years have died attempting to do a circuit to land after an EFATO. However many thousands of hours they might have, some are obviously not aware of the fact that IF YOU ATTEMPT A STEEP TURN WHEN FLYING LOW AND SLOW YOU ARE DEAD MEAT!
    They assume (presumably due to lack of training) that they only need to apply opposite aileron to level the wings! This is intuitive but fatal. They do not realise that in this situation the drag caused by opposite aileron merely stalls the inside wing promoting a spin.
    Google Fairchild B52 crash for a perfect example. You can see the pilot applying opposite aileron (spoilers) as the plane, at 90deg bank spins into the ground.
    The key question with regard to the Yeti disaster is, did either the PF or the captain have recent training on incipient spins? Either on a simulator or an aircraft?
    All of this assuming of course that there was no defect on the aircraft.

  • Apparently the preliminary report is out. Here’s what seems like a good and measured video about it:

    Still, not a lot of what was said in the cockpit is included (in the report and therefore the video) so it’s still a bit of a puzzle. Sylvia, are you going to write about this or wait for the final report? I can see either making sense.

    • the instructor inadvertantly feathered the propellers and never caught on despite the PF saying repeatedly the engines weren’t producing power

      it wasn’t the tight turn so much as they were already going too slow

      I can only assume the instructor wasn’t really used to flying from the right-hand seat? I hope the final report sheds more light on that (human factors).

      The report also suggests that flying the approach at more altitude is safer. The problem here is that landing on 12 requires entering the older airport’s airspace.

Post a comment:

Your email address will not be published. Required fields are marked *


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