Lufthansa Flight 2904

A check flight arrives as the weather changes

Lufthansa flight 2904 flew from Frankfurt to Warsaw on 14 September 1993. The Airbus A320-211, registered D-AIPN, carried sixty-four passengers and six crew members. Its flight deck contained two captains. The pilot in the left seat was flying under a check after a period away from flying, while the instructor captain in the right seat monitored and commanded the flight. Earlier sectors that day had been uneventful.

The flight departed Frankfurt at 14:27 UTC and cruised over Poland at flight level 330. The crew listened to Warsaw's automatic terminal information and descended towards an instrument landing system approach to runway 11 at Okecie. Warnings about windshear began reaching the cockpit. Meanwhile, an atmospheric front was approaching the airport from the west, bringing changing winds and heavy rain.

The available account is a third-party transcription of the Polish Main Commission's report, prepared by Peter B. Ladkin. It explicitly omits the original figures and appendix, and explains that formatting and Polish diacritics were altered. This sequence follows the report text reproduced there. It cannot substitute for inspecting the missing original illustrations or meteorological appendices, and detailed measurements should be read with that source limitation in mind.

The weather transition was important because information describing one moment could quickly become unrepresentative of the landing conditions. The aircraft approached runway 11 while the front passed the airport ahead of it. The crew had to make decisions about approach speed, runway length and landing technique in conditions that were changing during the final minutes.

Windshear warnings and an increased approach speed

The tower warned flight 2904 that an earlier aircraft had reported windshear on the final approach to runway 11. Another aircraft subsequently reported severe windshear after landing. The commission's analysis states that the crew heard three warnings about windshear in the approach area. The pilots continued the approach with the landing gear down, full flaps selected and manual control of thrust and the flight controls.

To respond to the windshear threat, the crew increased the approach speed. The report discusses an allowance in Lufthansa's manual to add up to twenty knots, provided the available landing distance permitted it. That qualification linked the protection against an airspeed loss to a different problem: an aircraft arriving faster requires more distance to stop. The commission found that the crew did not adequately assess whether the runway could accommodate the additional energy.

The wind measured or displayed aboard the aircraft also differed from the wind information being passed by air traffic services. The report found that the crew did not make effective use of the wind vector on the electronic flight instrument system or the difference between airspeed and ground speed. In its conclusions, it identified a tailwind component beyond the aircraft's operational limit and a continuation of landing despite the excessive ground speed.

Heavy rain accompanied the front. The runway became covered by an uneven layer of water whose thickness could not be firmly established. This affected the braking conditions after touchdown, but the problem was already developing before the wheels reached the pavement. Increased approach speed, a tailwind and the eventual touchdown position would leave less runway in which to reduce speed. The report's criticism concerns that combined situation, rather than the speed addition considered on its own.

The first wheel contact was not a completed landing

At approximately fifty feet above the runway, the recorded calibrated airspeed was 158.8 knots and ground speed was 172 knots. The aircraft carried its higher speed into the flare. The commission's analysis found that increased thrust had been maintained to a very low altitude and that the aircraft was distinctly above the instrument landing system glide path after crossing the threshold. The throttles were brought to idle at approximately six feet.

The first recorded contact with the runway occurred on the right main landing gear, approximately 770 metres beyond the threshold. The aircraft was banked slightly right as part of the response to the crosswind. That contact compressed the right shock absorber sufficiently to register on that side, but it did not establish the ground condition needed for immediate deployment of the principal deceleration devices. The left main landing gear did not register its corresponding touchdown until nine seconds later.

During those nine seconds, the aircraft travelled to approximately 1,525 metres from the threshold. Its ground speed was still about 154 knots when the left gear contact was recorded. The right-gear contact had given the crew an impression that the landing was complete. The pilot attempted wheel braking and, finding the expected response absent, asked the right-seat pilot to help.

The commission treated this interval as part of the touchdown sequence, rather than a normal completed landing followed by a sudden failure of every brake. Residual lift and the asymmetric initial contact delayed the conditions used by the aircraft's systems to recognise landing. By the time both main gear legs registered the necessary contact, the aircraft had consumed a substantial portion of runway 11. The distinction between physically touching the runway and satisfying the systems' ground criteria explains much of the delay that followed.

Why the deceleration devices responded late

The A320's ground spoilers and thrust reversers did not simply deploy whenever the pilots selected them. The report describes conditions intended to prevent their use before a valid landing. Ground spoilers required either compression of both main landing gear shock absorbers or sufficient wheel speed on both main gears. The thrust reversers required compression of both main gear shock absorbers. Until the applicable conditions were met, a selection from the cockpit did not produce immediate deployment.

The wheel brakes had separate activation logic. The system description states that normal braking depended on wheel speed relative to a reference computed by the braking control unit; the alternate mode could be enabled by switching off antiskid and nosewheel steering. It is misleading to describe all three devices as having one identical weight-on-wheels switch. The report's rollout account associates delayed wheel-brake action with wheel spin-up, alongside the delayed availability of spoilers and reverse thrust.

Once the left main gear registered touchdown, the ground spoilers deployed and the thrust reversers began operating. Engine speed increased as reverse thrust developed. Wheel braking followed later, with the report describing an additional delay of about four seconds. The delay was particularly costly because the aircraft had landed fast, with a tailwind, on a rain-covered runway.

The commission found that the crew could not override the lockout of the spoilers and thrust reversers in an emergency. Its recommendations asked the manufacturer to analyse possible emergency use independent of the existing logic. That was a recommendation for analysis, not a statement that unrestricted deployment would be safe in all circumstances. For this landing, the recorded result was that the devices became available only after much of the useful stopping distance had been lost.

The runway ends before the aircraft stops

Only approximately 1,275 metres of runway remained when the left main gear contact was recorded. Heavy rain and standing water reduced the effectiveness of the braking that was now available. The commission described aquaplaning during the rollout and found that deceleration became worse over the last 180 metres. In those conditions, the remaining distance was insufficient to bring the aircraft to a stop.

As the runway end and the obstacle beyond it approached, the pilot attempted to turn the aircraft right. The report explains that the aircraft rotated, but its centre of gravity continued on a path that did not turn sufficiently to avoid the obstruction. It left the paved runway at a ground speed of approximately seventy-two knots.

After travelling about another ninety metres, the aircraft struck an embankment with its left wing, passed over it and damaged the localizer antenna mounted there. The landing gear and left engine were also destroyed. The aircraft stopped beyond the embankment and caught fire. The right-seat instructor captain died in the impact, and one passenger also lost their life.

The embankment became a separate investigation issue. The commission found that it was located on the runway extension and was not described in the aeronautical information publication. It recommended that the obstruction and the localizer installation be documented. This feature affected what happened after the aircraft overran; it did not remove the earlier findings about approach decisions, touchdown and delayed deceleration. Each stage of the accident sequence left fewer options for the next.

Evacuation during a developing fire

The four cabin crew members began an evacuation as fire developed. The report credits the evacuation with rescuing sixty-three of the sixty-four passengers. Passengers helped one another leave the damaged aircraft. Two cabin attendants who had not suffered major injuries opened doors, deployed slides, directed the evacuation and assisted injured occupants away from the fire-threatened area.

An attempt to save the remaining injured passenger was unsuccessful. The report records that the body of the instructor captain could not be evacuated. In addition to the two fatalities, the injury table lists serious injuries to two crew members and forty-nine passengers. The severity of the injuries shows that survival did not mean an uneventful escape from the wreckage.

Five airport fire vehicles reached the site within three minutes of the emergency call. Their crews extinguished the external fire, but the fire inside the aircraft was more difficult to reach. The commission described fuel and oxygen contributing to rapid fire growth in the cabin. Firefighting agents could be directed inside after part of the roof was opened by an explosion.

The evacuation and rescue took place after a violent impact, in a damaged aircraft with an expanding fire. That sequence is documented in the available report text without requiring invented descriptions of individual passengers' thoughts or conversations. The established outcome is that most occupants escaped, many with serious injuries, while the accident destroyed the aircraft and killed two people.

The commission conclusions

The commission identified incorrect flight crew decisions and actions after receiving windshear information as the cause of the accident. It placed those decisions in the context of a front passing the airport with intensive changes in wind and heavy rain. It also found that aircraft design features limited the availability of deceleration devices and that the operations manual gave insufficient information about the increased landing distance associated with a higher approach speed.

The crew continued an approach with excessive tailwind and speed, touched down far along the runway and did not abandon the landing while the commission considered a go-around still feasible. Aircraft logic then delayed the deceleration response in the particular touchdown conditions, and the wet runway further reduced effectiveness. The report identifies these linked factors rather than attributing the accident to a simple mechanical brake failure.

Recommendations addressed landing-distance information, possible emergency use of spoilers and reverse thrust, meteorological reporting and publication of the runway-end obstruction. The transcription records those proposals at the time of the report. It does not establish their later implementation status, and its missing figures and appendices remain a limitation when examining the investigation in detail.