Safety

Jeju Air 7C2216 and the runway end: what investigators and regulators have said

June 6, 2026Updated September 29, 202611 min readFor pilots

Leer en español · Lire en français · Ler em português

In this article

What the preliminary report says
The missing four minutes
The engine update that was never released
From missed deadline to restart
The embankment
What the rules say
What Korea changed
Bird strikes and the go-around: what is and is not known
What it means for pilots

On December 29, 2024, Jeju Air flight 7C2216, a Boeing 737-800 from Bangkok, belly-landed at Muan International Airport, overran runway 19 and struck an embankment that carried the ILS localizer antennas. Of the 181 people on board, 179 died. Two crew members survived with serious injuries.

Almost two years later there is no final report, and the investigation has effectively restarted under a new parent agency. This article sets out what official bodies have published or stated, flags what came from leaks and press reports, and compares the Muan structure with ICAO and FAA rules. It does not try to explain the cause, because the investigators have not done that yet.

What the preliminary report says

Korea's Aviation and Railway Accident Investigation Board (ARAIB) issued a six-page preliminary report on January 27, 2025. Its sequence, in local time:

08:54:43 First contact with Muan tower; cleared to land runway 01.
08:57:50 Tower advised the crew to be cautious of bird activity.
08:58:50 Both the CVR and the FDR stopped recording. At that moment the airplane was at 161 kt and 498 ft pressure altitude.
08:58:56 The crew declared "Mayday" three times, reporting a bird strike during a go-around. The report says this time was converted from the CVR waveform.
The airplane flew over the left side of runway 01, turned right, lined up with runway 19 and belly-landed without the landing gear extended.
09:02:57 Impact with the embankment. The last 4 minutes 7 seconds were not recorded.

The report states that the pilots saw a group of birds on approach, that a security camera filmed the airplane close to a flock during the go-around, and that feathers and blood were found on both engines. DNA analysis identified Baikal teal, a migratory duck. The weather was benign: wind 110 degrees at 2 kt, 9,000 m visibility, few clouds at 4,500 ft. The captain had 6,823 hours (6,096 on the 737); the first officer had 1,650 (1,339 on type).

The report does not say when the strike occurred, why the gear was not extended, or why both recorders stopped.

The missing four minutes

The NTSB examined the recorders between January 7 and 11, 2025, and the investigation board, then part of the transport ministry, said the CVR stopped recording at 8:58:49 (The Korea Times).

Bloomberg reported that a 2008 FAA rule required new airplanes, from April 2010, to have backup power that keeps the CVR recording for about 10 minutes after main power is lost. HL8088 was delivered to Ryanair on September 4, 2009, according to the preliminary report. Bloomberg also reported the ministry saying more than half of the 101 Boeing 737-800s flying in Korea lacked the feature. ARAIB has not published an explanation of why the recorders stopped.

The engine update that was never released

In July 2025 ARAIB planned a press briefing on its findings. The families were briefed first, rejected the findings as premature, and the briefing was canceled (The Korea Times). The Jeju Air pilots' union said: "We firmly reject the malicious framing that seeks to shift the blame onto the pilot."

Reuters obtained the five-page update, dated July 19, 2025. According to Reuters, it said the left engine was less damaged than the right but was shut down 19 seconds after the bird strike. The right engine surged and emitted flames and black smoke, but investigators wrote that it "was confirmed to be generating output sufficient for flight." Reuters noted the update gave no reason for the crew's actions.

ARAIB never published it. Treat it as a leaked draft, not a finding.

From missed deadline to restart

December 8, 2025. The transport ministry, which had defended the structure as legal, acknowledged in material disclosed by an opposition lawmaker that the localizer did not comply with airport safety standards, AeroTime reported. A cross-party parliamentary inquiry began on December 22.
Late December 2025. Reuters reported that ICAO rules call for a final report within a year or a public interim statement on each anniversary, and that ARAIB postponed its update while legislation to reform the board was pending. Korea's anti-corruption commission had found that the embankment violated the requirement for such structures to be frangible.
February 28, 2026. Revised legislation took effect moving ARAIB from the transport ministry to the prime minister's office (Aju Press). Prime Minister Han Seong-sook later said the move was meant to ensure the board's independence and fairness (The Korea Times).
July 20, 2026. Choi Ki-young, an aerospace engineering professor, was appointed ARAIB chair. Chosun Ilbo reported that the new board plans to give some work, including analysis of possible aircraft defects, to foreign organizations to avoid claims of bias. It also plans to examine why the airplane touched down at about 380 km/h (about 205 kt), which the paper described as 1.5 times normal landing speed, and engine, power and control anomalies after the bird strike.
July 2026. SisaIN reported that ARAIB had missed its earlier target of a final report by June 2026, and that police had referred 34 people to prosecutors, who said they cannot decide on charges until ARAIB finishes.
August 6, 2026. Choi told reporters that results from the foreign contractors could slip into 2027, because writing the specifications and choosing vendors would take time (Yonhap via Daum).
August 11, 2026. Police searched the transport ministry, Korea Airports Corporation and the Busan Regional Office of Aviation over suspected professional negligence, after earlier raids in March and June (Aju Press).

The search for remains also had to be reopened: between April 13 and July 21, 2026, teams found 1,582 suspected human remains, the prime minister's office said (SBS).

As of late September 2026, we found no final report and no interim statement from the reconstituted board.

The embankment

On January 7, 2025, the transport ministry described the structure to The Korea Times: a mound about 2 m high, 264 m beyond the runway end, holding 19 concrete pillars, topped by a 30 cm concrete slab added during work in 2023 and 2024. The ministry gave Muan's runway end safety area (RESA) as 199 m and argued the structure was legal because it sat outside it.

Korea's Board of Audit and Inspection disagreed. Its March 2026 report, Reuters reported, found the ministry built a 2.4 m concrete embankment to mount the localizer in part to reduce earthwork costs, without adequate safety review. It counted 14 non-compliant localizer installations at eight airports and found safety approvals wrongly certified for up to 22 years. Bloomberg reported the audit found that ministry officials between 2008 and 2024 falsified documents to describe the structure as breakable, and that the ministry shortened the RESA in 2013 rather than remove the wall. Aerospace Global News reported the auditor told the ministry to extend the RESA out to the localizer at nine airports, Muan among them, and to discipline three officials.

A government-commissioned simulation by the Computational Structural Engineering Institute of Korea concluded that all 179 would likely have survived if the structure had been frangible or absent (AeroTime). That is a simulation, not an ARAIB finding.

What the rules say

ICAO Annex 14 measures the RESA from the end of the runway strip, not the runway end. An ICAO presentation on Annex 14 sets out the lengths below. Annex 14 paragraph 9.9.2, quoted in a 2025 ICAO regional working paper, says equipment that must sit in a RESA "shall be frangible and mounted as low as possible." ICAO's Aerodrome Design Manual (Doc 9157, Part 6) goes further on localizers, as quoted in a 2026 ICAO regional working paper: the transmitter housing should never be in the RESA or its extension within 300 m of the runway end, and existing non-frangible structures within 300 m should be replaced or moved.

	ICAO (Annex 14, code 3 or 4)	FAA
Length beyond the runway	Standard: 90 m from the end of the runway strip. Recommended: 240 m, or less with an arresting system	Typically 1,000 ft (305 m) beyond each runway end
Width	At least twice the runway width; recommended equal to the graded strip	Typically 500 ft
Objects inside	Frangible and mounted as low as possible (9.9.2)	Minimal mass and height, as low as possible, on frangible connections
Localizer guidance	Doc 9157: existing non-frangible structures within 300 m of the runway end should be made frangible or moved	Localizers typically on fuse bolts; frangibility point no more than 3 in (76 mm) above grade
Arresting systems	Allowed to reduce recommended length	EMAS on 118 runways at 70 airports

FAA figures come from its runway safety fact sheet (February 2025) and AC 150/5220-23A, Frangible Connections. The fact sheet says the Runway Safety Area Program improved more than 1,000 runways at 500 airports and that a standard EMAS can stop an overrunning aircraft at about 80 mph. The AC asks for foundations flush with grade and supports that "break, distort, or yield" on impact.

Set against these, the Muan structure sat 264 m from the runway end by the ministry's own figure: inside the 300 m in ICAO's localizer guidance and inside the FAA's 1,000 ft.

What Korea changed

January 22, 2025. The ministry named seven airports with problem localizer structures (Muan, Gwangju, Yeosu, Pohang Gyeongju, Gimhae, Sacheon and Jeju) and planned to remove concrete mounds and raised foundations and replace a rigid H-beam structure (Korea Herald).
February 6, 2025. Bird strike measures: thermal cameras at all 15 airports (four had them), bird detection radar at all airports, and at least two wildlife control staff on duty at all times. The ministry found Muan had one on duty on December 29 (Korea Herald). It also planned rules to move bird-attracting facilities away from airports (UPI).
April 30, 2025. An aviation safety plan converts localizers to ground-level, lightweight breakable structures, with six airports due by end of 2025. It requires a 240 m RESA at all airports by 2027, or EMAS where that is impossible, and a bird radar trial at Muan before a wider rollout from 2026 (BusinessKorea).
March 2026. After the audit, the ministry said it was still working on embankments at five airports (Bloomberg).

Muan was still closed to commercial traffic in August 2026. Sisa Journal reported that rebuilding the localizer and other safety work would take five to six months once recovery work ends.

Outside Korea, we found no regulator that announced rule changes in response. The ICAO Asia-Pacific aerodrome task force discussed papers on non-frangible navaids in strips and RESAs in February 2025 and again in February 2026, the first urging aerodromes to remove or correct such hazards. Neither paper mentions Muan.

Bird strikes and the go-around: what is and is not known

The preliminary report establishes the bird advisory, the crew's sighting, the flock near the go-around path, Baikal teal remains in both engines, and a gear-up landing on the opposite runway.

No published finding establishes when the strike happened, why an engine was shut down and which one, why the gear was not extended, why the recorders stopped, or why the crew chose to land on runway 19. ARAIB has issued no safety recommendations on bird strikes or go-around decisions that we could find. Any lesson drawn today comes from commentators, not investigators.

What it means for pilots

The overrun area is a design choice you never see from the flight deck. At Muan, a rigid concrete structure stood 264 m past the runway end and passed safety approvals for years, and Korea's auditor found 14 non-compliant localizer installations at eight airports. The rules exist. Compliance is another matter, and nothing on the airport chart tells you whether the field you are landing at meets them.

Identifying the damaged engine after a multiple bird strike happens under time pressure. The leaked draft puts 19 seconds between strike and shutdown. Investigators have not said what happened in that cockpit, so the lesson for now is the one already in your company's procedures for confirming the affected engine.

Older airframes may have no independent power for the CVR. HL8088 was delivered in 2009, before the US requirement took effect. If you fly airframes built before 2010, find out whether yours has it.

Treat a bird advisory as a threat, not background. At Muan the tower's caution came one minute before the recorders stopped.

The families of the 179 people who died are still waiting for an answer. When ARAIB publishes, we will update this article.

Sources

This is general information, not operating guidance. Your company's procedures, your aircraft's manuals and the regulations come first.