Accident on 24 December 2000 at Tahiti Faaa Airport to the DC10-10 registered N132AA operated by Hawaiian Airlines REPORT TRANSLATION

SYNOPSIS

Summary

While landing on runway 04, in a storm, the airplane touched down about halfway down the runway. It overran the end of the runway and came to a stop resting on its engines, its nose in the lagoon.

Consequences

ORGANISATION OF THE INVESTIGATION

The initial work in the investigation was undertaken by the BEA investigator on duty at Le Bourget and the Field Investigator at Tahiti Faaa. An Investigator-in-Charge (IIC) was then nominated to take over. The NTSB was informed and invited to nominate an Accredited Representative on behalf of the USA as State of Registry and State of Manufacture of the airplane.

The Field Investigator carried out the preliminary gathering of evidence. Examination of the tires and the life jackets was performed by technicians from the airline.

The flight recorders were taken to the BEA by a judicial police officer. They were opened in his presence and the recordings were then read out, in coordination with the NTSB.

A two-man team, led by the IIC, then went to the headquarters of Hawaiian Airlines in Honolulu to collect information and to analyze parameters with the assistance of NTSB investigators and advisers from FAA and the operator.

Calculation of the airplane's performance was performed by Hawaiian Airlines, under the supervision of the NTSB.

1 - Factual Information

1.1 History of Flight

On Sunday 24 December 2000 at 4 h 38 min, the DC10 registered N132AA took off from Honolulu bound for Tahiti with a hundred and thirty-nine passengers and fifteen crew members on board. This was Hawaiian Airlines scheduled international flight HAL 481. The crew consisted of a Captain, a co-pilot and a flight engineer. No notable events occurred on the flight until the approach.

Between 9 h 24 min 35 s and 9 h 26 min 57 s, the Captain , Pilot Flying (PF), performed the arrival briefing and mentioned the following items: frequency, description of missed approach. He then asked the co-pilot to find out about the weather. The controller transmitted the 9 h 00 information, mentioning a 080°/5 kt wind, occasionally 340°/15 kt, with gusts to 25 kt. He also mentioned rain showers, some cumulonimbus and indicated that the runway was wet.

During the descent, numerous thunderstorm cells were observed by the crew on the track and around the aerodrome.

At 9 h 32 min 19 s, flight HAL 481 passed the ARONA point at 9,000 ft. At 9 h 46 min 30 s, the tower controller asked them to descend towards 2,500 ft QNH and to report back when passing OVINI. He stated that there were showers at the aerodrome. Twenty-two seconds later, the co-pilot announced that they were passing OVINI.

At 9 h 49 min 04 s, the Captain armed the spoilers. At 9 h 49 min 09 s, he asked for the flaps to be extended to 22° then, twenty-five seconds later, to the 35° position.

At 9 h 49 min 46 s, the controller cleared flight HAL 481 to land on runway 04. He gave the wind as 060°/10 kt, gusting to 14. The Captain noticed changes in the wind and the co-pilot announced, based on the airplane's GFMS, a wind from 280° at 28 kt.

At 9 h 51 min 24 s, the controller transmitted new meteorological information: 330°/18 kt, gusting to 28. Thirty-two seconds later, the crew had the airfield in sight and a final wind reading was given to them: 330°/18 kt, gusting to 29.

At 9 h 52 min 11 s, the autopilot was disconnected. Nine seconds later, the airplane passed under the approach path and the GPWS "sink rate" warning sounded. The pilot rejoined the descent path with the aid of the PAPI and continued his approach using external visual references. At that time, the airplane was following a track parallel and to the right of the approach path.

Between 9 h 52 min 32 s and 9 h 52 min 38 s, the radio altimeter call-outs between fifty and ten feet began and continued at a rate of one per second. Power reduction began five seconds after the ten feet call-out.

At 9 h 52 min 45 s, the wheels of the main landing gear touched the runway, to the right of the centerline. Five seconds later, which was two seconds after the nose gear touched down, the thrust reversers were deployed. The parameters indicate that reverse thrust on each of the engines was adjusted and regulated without it ever reaching its maximum value.

At 9 h 52 min 53 s, eight seconds after the main landing gear touched down, the spoilers were deployed manually by the flight engineer.

At 9 h 52 min 59 s, the copilot's "Centerline" call-out suggested the Captain rejoin the runway centerline, from which the airplane was moving away to the left.

At 9 h 53 min 21 s, the airplane crushed the localizer antennae, continued along the runway extended centerline and then came to a stop past the end of the runway, its nose in the lagoon.

The Captain , after checking on the safety situation around the airplane, ordered the evacuation via door 2R.

1.2 Injuries to Persons

No injuries were sustained during the accident or the subsequent evacuation.

1.3 Damage to Aircraft

The airplane was damaged by contact between the main landing gear and the outboard engines with obstacles, as well as by the immersion of the forward part of the airplane.

1.4 Other damage

The localizer aerials were completely destroyed.

1.5 Personnel information

1.5.1 Captain

• Male, aged 56.
• Valid Commercial Pilot's License.
• Last medical checkup on 4 December 2000.
• Last line check or equivalent on 25 September 2000 (recurrent, simulator).
• Total flying hours: 18,905.
• Flying hours on DC10: 4,860 (all as Captain).
• Flying hours in three previous months: 214 (all on DC10).
• Type ratings DC-10, L-1011, DC-9, DHC-7.

Before joining Hawaiian Airlines in 1977, the pilot had been employed by the US Navy until 1975 , by Panorama Air Tours and by Royal Hawaiian Air Service.

1.5.2 Copilot

• Male, aged 35.
• Valid Commercial Pilot's License.
• Last medical checkup on 16 May 2000.
• Last line check or equivalent on 25 January 2000 (recurrent, simulator).
• Total flying hours: 7,142.
• Flying hours on DC10: 526.
• Flying hours in three previous months: 179 (of which 121 on DC10).
• Type rating: DC-8.

• DC-10 Flight Engineer.

Before being employed by Hawaiian Airlines as a Flight Engineer from 1998, this pilot had been employed by Island Air then by the Air National Guard.

1.5.3 Flight Engineer

• Male, aged 37.
• Valid Flight Engineer's License.
• Last medical checkup on 6 June 2000.
• Last line check or equivalent on 13 December 2000 (simulator).
• Total flying hours: 4,133 (of which 600 as Flight Engineer).
• Flying hours on DC10: 613 as Flight Engineer.
• Flying hours in three previous months: 259.

Before being employed in 1999 by Hawaiian Airlines, the Flight Engineer had worked for Corporate Air.

1.6 Aircraft Information

1.6.1 Airframe

• Manufacturer: MCDONNELL DOUGLAS CORP .
• Type: DC10-10.
• Serial n°: 47827.
• Airworthiness certificate issued on 9 November 1979.
• Total service time on 24 December 2000: 64,841 flying hours.
• Last overhaul: 29 November 2000.
• Flying time since last overhaul: 152 hours.

1.6.2 Engines

Manufacturer: GENERAL ELECTRIC.

Type: GE CF6-6 K.

Thrust under standard conditions: 42,000 lbs or 18,670 daN .

1.7 Meteorological Conditions

1.7.1 General situation

Surface situation:

An active convergence line was crossing the island with numerous stratocumuli and scattered squalls.

Situation at altitude:

Above the lower layer convergence, a notable altitude divergence was observed over the area, a factor that particularly amplified the instability.

Figure 1 - Infra-red satellite photo taken by GOES 10 on 24/12/00 at 10 h 52

1.7.2 Forecast provided to crew

Forecast included in flight dossier:

PPT: 232100Z 240024 34015KT 9999 VCSH FEW010 FEW012CB BKN016 BKN045 TEMPO 0024 34015G25KT 5000 SHRA SCT008 SCT010CB BKN015 OVC035 PROB40 TEMPO 0024 34025G40KT 1000 ?TSRA SCT006 BKN008CB OVC015

ACARS update of METAR on 24 December at 4 h 00:

PPT: 240400Z 36012KT 9999 VCSH FEW008 FEW010CB SCT013 BKN040 27/25 Q1007 TEMPO 34015G25KT 4000 SHRA SCT010CB BKN015 OVC035

ACARS update of METAR on 24 December at 9 h 00:

PPT: 240900Z 08005KT 9999 SCT015 SCT016CB BKN050 26/24 Q1009 TEMPO 34015G25KT 4000 SHRA SCT010CB BKN015 OVC035

1.7.3 Meteorological conditions at Faaa during the landing

1.7.3.1 Meteorological observations

METAR at 10 h 00:

Wind 290° / 18 kt, heading  variation 240° to 360°, visibility 2,000 m, air temperature 26°C, dew point 24,3°C, 4/8 of cumulonimbus at 1,600 ft, QNH 1009, lightning notified by observer.

SPECI at 9 h 52 min:

Wind variable 08, gusts to 27 kt, heading  variation 290° to 200°, visibility 2,000 m, light rain showers, 3/8 to 4/8 of cumulonimbus at 1,600 ft, air temperature 26°C, dew point 24°C, QNH 1009. Trend: tempo 340° 15 / 25 kt, visibility 4,000 m, moderate showers, 3/8 to 4/8 of cumulonimbus.

1.7.3.2 Wind measuring equipment

The control tower has analog wind indicators at the control positions and a SIGMA terminal which receives wind, atmospheric pressure and temperature information from the thresholds of runways 04 and 22 and at about mid-runway. The SIGMA terminal gives:

CVR

• Make: L3-COM
• Type: A100A
• Type number: 93-A100-80
• Serial number: 50223

The CVR had a magnetic tape with thirty minutes recording time.

Both recorders arrived at the BEA on 8 January 2001. They were in good condition and readout could commence immediately.

1.11.2 FDR readout

Hawaiian Airlines did not possess any conversion documents allowing the raw binary data to be transformed into engineering values. American Airlines, the airplane's owner, provided two different conversion documents, not knowing which of the two corresponded to the airplane. After having selected the most likely document, the investigators met with conversion problems and validation of certain parameter values (Radio Altitude, acceleration and Glide Slope and Localizer deviations) had to be done based on recordings of previous flights. In addition, the operator was unable to provide up-to-date and exact documentation on the evolution of the parameter acquisition and recording system. Thus, the FDR readout was slowed down and some lack of precision may remain.

The graphs are in the appendices.

Of note, during final approach:

• The autopilot was disconnected at 9 h 52 min 10 s; the airplane was then at a radio altimeter height of 328 feet and its speed was 155 kt.
• Between autopilot disconnection and main gear touchdown, to the right of the runway centerline, the positions of the aileron and elevator controls changed rapidly, which indicates significant inputs on these controls.

Some parameters are shown in the following table for various phases of the landing.

1.11.3 CVR

The CVR transcript is in appendix 1.

The following points are of note:

• The landing briefing was carried out twenty-eight minutes before the landing. Despite the meteorological information received via the ACARS, which was identical to that which was provided by Faaa ATC one minute after the briefing, the crew did not mention the presence of cumulonimbus and the instability that dominated the area round the aerodrome. They planned to perform a VOR DME ILS approach to runway 04.
• During the descent, there were a lot of exchanges about the storm situation on the airplane's track and around the aerodrome. Twelve minutes before the landing, i.e. six minutes before the OVINI IAF, the crew mentioned the presence on the onboard meteorological radar of some thunderstorm cells.
• The spoilers were armed during the approach and the extension lever was activated seven seconds after main gear touchdown.
• A GPWS "Sink Rate" alarm was recorded on short final.
• The synthetic callouts of fifty feet down to ten feet occurred between 9 h 52 min 32 s and 9 h 52 min 38 s.
• The airplane went off the runway thirty-six seconds after main gear touchdown.  The co-pilot asked for assistance on the Tower frequency while the Captain ordered the evacuation.

1.12 Wreckage and Impact Information

Source: Tahiti Faaa Air Transport Gendarmerie Brigade

The visible damage on the airplane, from forward to aft, was as follows:

• Impact marks between the nose and nose gear.
• The nose gear was marked by the impact with the localizer aerials, one of which was jammed inside. The anti-shimmy was destroyed. No serious structural damage was visible.
• The air intake on the right engine was punctured by a rock and that on the left engine showed signs of scraping along the ground. Some fan blades on the left engine were slightly damaged. The pylons of both outer engines were twisted.
• The left main landing gear had struck the localizer installations. Numerous contact marks with the antennae were visible.
• The left inner flap was damaged on its lower part. The right inner slat was deformed by contact with the pylon of engine three.

The day after the a ccident, the airplane was towed to the north ramp to be repaired.

Note: the localizer antennae, located seventy-two meters from the end of the runway, were destroyed.

1.13 Medical and Pathological Information

The investigation brought to light no evidence of any medical anomalies that may have altered the crew's abilities.

1.14 Fire

There was no fire.

1.15 Survival Aspects

Since the airplane was in landing phase, the cabin preparation had been carried out and the passengers and crew were seated and strapped in.

After the airplane stopped:

• The Captain and the Flight Engineer carried out the emergency evacuation procedure.
• The Chief Flight Attendant switched on the emergency evacuation lighting.

Given the sloping position of the airplane, the evacuation was carried out via the right center door emergency slide (2R)  which was the only one considered to be usable. The slide at the left center door  (2L) was also deployed but it seemed that the depth of the water was about fifteen meters. The lighting conditions made it difficult to evaluate the situation and advice was taken from the firefighters not to evacuate the passengers via that door. The end of the right slide was tied up to the sea wall located at the end of the runway so as to give the passengers the shortest walk possible through the water, which was shallow on that side.  The passengers were greeted by the firefighters, who had taken care to remove the barbed wire coils that were protecting the aerodrome against intruders from that side.

On exiting the airplane, the safety of the passengers was ensured by the cabin crew who grouped them together on the runway while awaiting the arrival of transportation. All of the passengers were equipped with life jackets, but putting on the jackets had posed some problems (see 1.16.4).  Some of the life jackets were found partially opened on the cabin floor.

Some of the passengers were French-speaking. Some of them stated that, in the course of the evacuation, they had some communication problems with the cabin crew who were speaking to them in English.  The operator explained that for flights bound for Tahiti, some cabin crew members were selected according to their ability to speak French. Of the twelve cabin crew on the flight, two spoke French and translated the safety and evacuation announcements into French, though the loss of the PA system meant that it was impossible for them to be heard throughout the cabin.

1.16 Tests and Research

1.16.1 Airplane's track on the runway

The airplane's track on the runway was calculated from the accelerations, the localizer spacing and the airspeed recorded on the FDR, which allowed the main gear wheel touchdown position and time to be correlated.

The landing roll, compared with the airplane's magnetic heading, indicates that the airplane started sliding to the right for eleven seconds from touchdown of the main gear, then slid to the left, between 9 h 53 min and 9 h 53 min 13 s.

Ground track determined from recorded FDR data

Note: The winds indicated above correspond to those at 9 h 53 min averaged over the last two minutes at thresholds 04 and 22 and the meteorological station. Threshold 04 is not included due to the scale chosen.

1.16.2 Wet runway and crosswind

On a wet runway and with a crosswind, the wind's effect on the fuselage and the tail tend to align the airplane into the face of the wind and to move across with the runway wind. This tendency of the airplane to align itself into the face of the wind and to be pushed off the centerline is increased by the application of reverse thrust. 

In addition, since adhesion is low on a wet runway, the effectiveness of braking is reduced. To limit this effect, a solid contact between the main gear wheels and the runway through a firm landing allows the wheels to rotate more rapidly, thus reducing skidding.  Setting the nose gear down quickly and a nose down movement on the control column also help to keep the airplane in good contact with the ground.

1.16.3 Examination of tires

A visual examination of the main gear tires was performed so as to determine the presence of any traces of possible local melting[3] on the tires that would be characteristic of hydroplaning. No such traces were observed.

Note: these traces are not systematically present when an airplane slides on a wet runway. If, for example, the layer of water is thin and lubricates the runway, the tires can skid and there are no signs of vulcanization. This type of hydroplaning is denoted as viscous.

1.16.4 Spoilers

1.16.4.1 System description

By reducing lift on the wings, the spoilers push the airplane onto the ground and assist deceleration and improve braking.

In auto mode, their extension depends on the rotation speed of the main gear wheels. Under the following conditions, they cannot be extended:

• crosswind,
• wet runway,
• low vertical speed on touchdown.

During the landing roll, a roll input on the control column leads to the spoilers on the opposite wing to the input direction being retracted.

1.16.4.2 Use

The spoilers must be armed by the captain before landing. Spoiler arming is not checked during the landing checklist. The captain must ensure that they have extended after touchdown and extend them manually in case they do not extend automatically. The flight engineer must also check their extension and, if they do not extend when the thrust reversers deploy, he must extend them manually, which means he must lean forward.

1.16.4.3 Effect on landing distance

Without the spoilers, deceleration during the landing roll is not in accordance with the performance in the manufacturer's documentation. In case of a known unavailability of the system, an additional two hundred meters roll must be allowed for.

1.16.5 Examination of life jackets

The life jackets, manufactured by Hoover Industries, are packed by W. H. Brennan Inc. Some passengers had difficulties in using them. Some tests were performed on life jackets taken at random from the airplane and which had not been used during the evacuation.

On some of the life jackets, when the passenger put it on and pulled the "Pull to tighten" tag, the elastic band installed during packing to maintain the adjustment slack was placed in the male part of the attachment mechanism, thus preventing it from locking into the female part.

If the wearer managed to attach the male and female parts by forcing them, he could not pull the excess strap through the locking buckle and thus adjust the life jacket.

1.16.6 GFMS

The electronics bay, where the GFMS computers are located, was flooded with water during the runway excursion and the computers were damaged and rendered unusable. Further, it was impossible to read out their non-volatile memories so as to be able to get some parameters not recorded on the FDR, such as ground speed and wind.

1.17 Information on Organizations and Management

1.17.1 Flight preparation

Hawaiian Airlines' flight preparation is computerized and outsourced. The main parameters that had been determined for the landing were as follows:

The approach speed, indicated by the flight engineer about twenty-five minutes before the landing based on the operations manual [4], and equal to V_ref  , plus half of the wind speed, plus a gust value, was 156 knots in the case of flight HAL481.

1.17.2 Crosswind procedure

The maximum value of the demonstrated crosswind component is thirty-one knots and does not constitute a flight manual limitation. On a wet runway, this value must be reduced to twenty knots.

Note: these values are based on an established  wind speed.

1.17.3 Crosswind and wet runway procedure

If the airplane tends to deviate under the runway wind (see 1.16.3), the operations manual advises releasing pressure on the brakes and canceling use of the thrust reversers in order to regain directional control of the airplane. It states that an increase in thrust can allow the airplane to be re-aligned more easily, while calling attention to the limitations imposed by the length of the runway.

1.17.4 Use of flaps

The operations manual specifies:

• Use of the flaps at 35° when the runway is dry and longer than 7,000 feet,
• Use of the flaps at 50° in the following cases:
• When the runway length is less than 7,000 feet,
• When the runway is wet or slippery and when braking quality is lower than normal or in the case of a tail wind.

Note: the case of a wet runway with normal braking is not covered.

1.17.5 Crew training

Landing on a wet runway and hydroplaning are covered by Hawaiian Airlines during the pilots' line oriented flight training. They receive reminders on micro-bursts and wind shear during recurrent training courses.

Use of the flaps at 50° is dealt with during regular simulator training but the operator's common practice is to use them only at 35°. The operator justified this practice by citing the lower stresses on the airplane structure. However, the choice of the landing configuration remains at the Captain's discretion.

1.17.6 Calculation of landing  and landing roll distances

Taking into account a flaps configuration of 35°, the graph relative to a landing at 335,000 lbs on a wet runway at sea level, with water level between three and six millimeters with full reverse thrust on all three engines, gives a  required landing distance [5] of 5,800 feet, so LD _35° = 1,740 meters. The approach slope at Faaa for QFU 04 is 5.24 %. From a strictly performance perspective, by passing above the runway threshold by fifty feet, the corresponding computed distance between theoretical wheel touchdown and complete airplane stop is 1,450 meters [6]. The data drawn from the FDR showed that the distance between the touchdown and the runway excursion was about 1,800 meters.

It should be noted that the direction and strength of the wind, the runway slope, the temperature and  the runway surface adhesion are also factors that influence the landing distance and that are not taken into account for this calculation.

1.18 Additional Information

1.18.1 Controller's testimony

The controller thought that the airplane was on the slope and the path with a significant drift. At the flare, he knew that the touchdown would be late since the crew had to correct the drift. At the moment the airplane passed abeam the tower, the controller thought that it was fast and only a slight nose up attitude. He noticed that it touched down at about the level of the cross, after taxiway G. He heard the noise of the reversers a little after the touchdown and it seemed to him that this application of reverse thrust was done in a progressive manner.

The controller alerted the emergency services. He diverted an Air New Zealand

plane to Rarotonga then, later the Air France flight to Los Angeles

1.18.2 Crew testimony

According to the crew, the landing gear lever was activated during the glide capture and the landing checklist was carried out shortly after. The spoilers were armed, the flaps extended to 35° configuration and the speed reduced to V_ref. The approach was carried out in clouds, without any particular turbulence. At six hundred feet, the crew saw the approach lights, the PAPI and the runway. It was raining and the windshield wipers were on.

The crew kept the autopilot on until about six hundred feet. Before landing, there was a "Sink Rate" alarm. The Captain then checked the PAPI indication and noticed that the airplane was a little low in relation to the glide slope. He corrected this and landed the airplane slightly to the right of the runway centerline. He selected reverse thrust and braked. He thought he recognized some hydroplaning. The Flight Engineer extended the spoilers manually after the "no spoiler" callout by the co-pilot.

The Captain wanted to bring the airplane back onto the runway centerline. The airplane was to the left of the centerline and the co-pilot called out "Centerline". The Captain then corrected to align the airplane on the centerline while continuing to brake and to use the thrust reversers. He thought that by using too much thrust reverser power he would have more difficulty in keeping the airplane on the runway.

The Captain did not see the distance indications on the runway, but, about three hundred meters from the end of the runway, he noticed a metal barrier across the end. He selected more thrust reverser power and increased pressure on the brakes. It seemed to him that he was on an ice-covered runway. The airplane hit the metal barrier and left the runway.

Fearing a fire, the Captain and the Flight Engineer carried out the evacuation procedure rapidly. The engines were shut down and the all the items in the procedure were carried out normally. The Captain used the interphone to reassure the passengers.

The co-pilot informed the control tower then went into the passenger cabin to evaluate the situation in relation to the airplane's position. He decided not to use the aft doors for the evacuation, the airplane's tail being too high in relation to the ground. Doors 2R and 2L were open with slides deployed. He saw the water and the barbed wire and decided to wait for the Captain to arrive in the cabin before beginning the evacuation. The emergency services removed the barbed wire and attached a rope to the slide to bring it closer to the land. The evacuation took place calmly and very rapidly.

Captain's additional statement

The Captain stated that he had used reverse thrust because he wanted to control the airplane's direction in relation to the runway centerline.  However, since page 10-4 of the Jeppesen document for Tahiti, noted as "Noise abatement Procedure", stipulated that reverse thrust should not be used between 5 h 00 and 16 h 00 UTC, except in case of emergency, he did not use it fully.

The airplane was not equipped with an automatic braking system. The Captain remembered that his action on the brakes seemed ineffective.

Flight Engineers additional statement

The Flight Engineer did not remember what distance the airplane had landed from the threshold. He remembered that there was neither any lighting indicating remaining runway length nor runway centerline lighting at Tahiti Faaa.

2 - Analysis

2.1 Scenario

The Honolulu-Papeete flight encountered no particular problems until the initial approach.

2.1.1 The approach

During preparations for the landing, the crew did not take into account the presence of cumulonimbus around the aerodrome. During the descent, the extremely changeable meteorological conditions progressively made the crew more aware of the particular meteorological conditions on arrival. Nevertheless, they did not ask the controller how the situation had evolved over the aerodrome. Nor did they envisage delaying the landing, while a hold at OVINI was possible. Arriving on short final, they entered an area of rain and turbulence associated with the passage of a storm over the airfield at that moment.

2.1.2 The flare

The recordings of the intensity and direction of the wind along the runway show variations in all directions over time. The turbulent nature of the wind explains the Captain's significant inputs on the flight controls in order to maintain lateral control of the airplane's track, perhaps at the expense of control of the descent path on short final.

Thrust reduction was late: seven seconds passed between the radio altimeter "ten feet" call-out and touchdown of the main landing gear. This may be explained by the Captain's focusing on lateral control, and could be intentional, as a high thrust level allows for better control of the track. In any event, the high thrust led to an increase in glide distance. This increase was accentuated by a sudden headwind component, as is shown by an increase in computed airspeed four seconds before touchdown of the main landing gear.

The accumulation of these factors pushed back the touchdown zone by about nine hundred meters towards the middle of the runway.

On landing, the airplane veered to the left of the runway due to the crosswind, the wet runway and the use of the thrust reversers.

Although armed, the spoilers did not extend automatically after the wheels touched down. This phenomenon, linked to the main landing gear wheels starting to rotate, can occur during a landing in crosswind, on a wet runway or when the vertical contact speed with the ground is low. All of these conditions were present. The Captain, absorbed by controlling the airplane's track on the runway, did not notice the non-extension. It was noticed by the flight engineer eight seconds after touchdown. During this time, the airplane was slipping and did not benefit from optimal aerodynamic braking; the effectiveness of the brakes was also reduced. This extended the airplane's ground roll.

Note: It is regrettable that the physical measurements of runway friction are not available, since this would make it possible to correct any deficiencies where necessary. In addition, they would also allow the pilot to select the most appropriate braking mode.

2.1.4 The runway excursion

When the Captain noticed the localizer antennae at the end of the runway, he increased inputs on the thrust reversers, now ineffective due to the airplane's speed, and on the brakes. These events occurred in the runway 22 touchdown area, an area whose surface is covered with tire rubber and where adherence is worse than on the rest of the runway. The airplane began to slide and the braking action was virtually nil, as the Captain indicated in his testimony. Nothing could then prevent a runway excursion.

The poor conditioning of some life jackets had no effect on safety since there were more life jackets than passengers and the evacuation of the airplane took place calmly and with no rush. It is probable that under other, less favorable, conditions, the discovery by the passengers of the unavailability of some life jackets could have been a factor leading to panic.

Although two of the cabin crew spoke French, some francophone passengers did not understand the safety instructions given during the emergency evacuation. This situation appears hard to avoid under emergency evacuation conditions, in particular where the public address system is inoperative, with the added stress and the obvious impossibility of grouping the passengers together according to their language. Under certain conditions, it may even be impossible to avoid this with passengers who speak the crew's language perfectly. Only attentive listening to the safety announcement presented at the beginning of the flight and the reading of the documentation available to each passenger can at least partially compensate for this difficulty.

2.2 Strategy for performing approaches

According to the regulations, an approach can only be initiated if the aerodrome is accessible, that is to say if the horizontal visibility is greater than the minimum value specified for the planned approach. The ceiling thus gives an indication of the chance of being able to perform the approach. Further, other parameters such as crosswinds, windshear and precipitation are taken into account by the crew. In practice, these checks are performed during the briefing and the crew generally has a sufficiently precise idea at that time on how the approach will go up until the landing or, at least, until the minimum descent altitude (or height in the case of precision approaches).

In the case of the accident flight, the visibility conditions were clearly greater than the minima and the ceiling did not present any obstacle to a landing. The crew observed deterioration in the meteorological conditions during the approach, but this did not cause them to reassess the decisions that they had taken previously. Even the indications on the radar, which they qualified as "serious" did not stimulate any particular reaction on their part. A calling into question of the approach strategy decided on during the briefing was in fact difficult, since the performance of the approach left them little time available and the image of the approach had in part been fixed in the pilots' minds through the actions defined during the briefing.

Consideration of the changes in meteorological conditions announced by ATC, the weather radar information and the conditions observable outside could have prompted the crew to reconsider the approach strategy decided on during the briefing. Furthermore, the crew had to take into account the possible presence of windshear around the airfield, which could have made a go-around difficult once the final approach was started. Under these conditions, during the approach, the crew should have considered holding at OVINI so as to allow the stormy weather to pass over the aerodrome, especially since the diversion airfields for Tahiti Faaa are a long way away.

Such a change in strategy during the approach is, however, unusual. It is likely that a crew would have recourse to this particularly when it has been mentioned during the preparation for the approach. This point could be emphasized during crew training so as to sensitize crews to the risks inherent in passing through a storm during landing. The probability of such simultaneous events is low and pilots almost never encounter them, which leads to an under-estimation of the risk.

2.3 Faaa runway Infrastructure

Centerline lighting could have helped the pilot to determine the airplane's position, both laterally and in relation to the runway end.

2.4 Documentation used by the crew

Most operators use Jeppesen documentation and crews are used to doing so. However, although Jeppesen receives NOTAM's, the chart of the Tahiti airfield had not been updated. Such a failure could lead to dangerous situations for more restricted airfields, since it could lead to erroneous performance calculations during flight preparation.

Nevertheless, it should be noted that it is the responsibility of the operator to ensure that documentation used is up to specifications when it differs from the official documents.

2.5 Flight recorders

In contrast to the European JAR's, the American regulations do not oblige operators to perform a systematic flight analysis. This may help to explain the lack of follow-up in the management of conversion documents and in the maintenance of the parameter measurement and acquisition chain.

If it had not been possible to validate the parameters from previous flights, the parameters of the accident flight, although recorded on the FDR, might not have been usable.

3 - Conclusions

3.1 Findings

• The crew possessed the licenses and qualifications required to undertake the flight.
• The airplane possessed a valid Certificate of Airworthiness.
• The operator did not have the conversion documents for the recorded FDR parameters.
• The JEPPESSEN aerodrome chart for Faaa did not take into account the modifications in runway landing distances notified by  NOTAM.
• The possibility of a hold, given the evolution in the meteorological conditions, was not considered by the crew during the preparation of the approach.
• During the final approach, the airplane entered an area of rain and turbulence associated with the presence of a storm.
• The runway was wet, the wind was blowing from the left at the threshold of runway 04 and was between 18 and 28 knots.
• Due to the difficulties in controlling the airplane, thrust reduction was late.
• The touchdown of the main landing gear wheels occurred about one thousand three hundred meters after the displaced threshold of runway 04.
• Landing with a crosswind on the wet runway with a low vertical speed did not allow the spoilers to be deployed automatically.
• The flight engineer extended the spoilers manually eight seconds after touchdown of the main landing gear wheels.
• The airplane touched down to the right of the runway centerline, then its track veered to the left before coming back towards the center.
• The airplane left the runway along the extended centerline then came to a stop with its nose in the lagoon.

3.2 Causes of the accident

The accident was caused by the failure, during the preparation for the approach, to take into account the risk of a storm passing over the airfield at the time of landing.

The following factors contributed to the accident:

• The crew focusing on lateral control of the airplane's track, due to a strong crosswind that was changing in strength and direction, and late thrust reduction, resulting in a glide and a  long touchdown;
• The late manual extension of the spoilers, which increased the length of the landing roll;
• The presence of water on the runway, the low vertical speed during contact with the ground and perhaps the slipperiness of the runway, which made the airplane slide, in particular in the threshold 22 wheel touchdown area.

4 - RECOMMENDATIONS

Performing an approach in meteorological conditions with local tropical storms over the aerodrome guarantees neither the landing nor a go-around. Once the decision is taken to perform the approach, however, it is difficult for a crew to reconsider, in this case as a result of deteriorating meteorological conditions, if they have not planned to do so when they develop their landing strategy during the arrival briefing.

Consequently the BEA recommends that:

• operators ensure that crews are made aware of the importance of specifically planning, during the arrival  briefing, for circumstances that would lead to a modification in the approach strategy, where the meteorological situation warrants it.

The only information available concerning the characteristics of the runway at Tahiti Faaa aerodrome is qualitative, which is inadequate for a precise evaluation of the influence of these characteristics on airplane braking.

Consequently the BEA recommends that:

• the DGAC measure the adherence characteristics of the runway at Tahiti Faaa aerodrome.
• The absence of runway centerline lighting could have contributed to increasing the crew's difficulties in positioning the airplane laterally and in relation to the far end of the runway.

Consequently the BEA recommends that:

• the DGAC study the possibility of equipping all aerodromes on French territory used for public transport with runway centerline lighting.

The regulatory documentation for French aerodromes is published by the French Aeronautical Information Service (IAC and AIP charts). Updates to this documentation can be the subject of NOTAM's.  Crews, for their part, frequently use the Jeppesen documentation and erroneous information in the latter can have negative consequences in terms of safety.Consequently the BEA recommends that:

• operators systematically ensure that the documentation used by aircrew is in accordance with the relevant national regulatory documentation.

Readout of the flight data recorder was complicated by and could have been jeopardized by the absence of conversion documents. Hawaiian Airlines did not possess these documents and those obtained from American Airlines, the owner of the airplane, contained errors.  In France, operators are required to deposit these documents with the oversight authorities for airplanes included in the fleet list. Consequently the BEA recommends that:

• the FAA ensure that American operators possess up-to-date conversion tables for onboard data for airplanes used for public transport.

Appendices

APPENDIX 1: CVR TRANSCRIPT

FDR graphs

 

 

Recording of winds in the hour of the accident

OVINI-VOR/DME-ILS approach chart for runway 04 on 2 November 2000

Telecommunications transcript

---

[1] Except where otherwise noted, the times shown in this report are expressed in Universal Time Coordinated (UTC). Ten hours should be subtracted to obtain the time in Tahiti on the day of the accident or one hour added to obtain the applicable time in Paris on the day of the accident.

[2] Related to the vertical distance between the pilot's eye and the airplane's wheels in approach configuration (D_or).

[3] Airplane hydroplaning usually causes the water at the contact point to boil, which melts the rubber, leading to the obliteration of the tire tread.

[4] The Hawaiian Airlines operations manual is the same as that of American Airlines.

[5] The landing distance (LD) is the distance between the passage over 50 ft and complete stop of the airplane.

[6] The same calculation with a flaps 50° configuration gives 1,020 meters. It should, however, be noted that this configuration would likely have increased the airplane's glide distance and, perhaps, the wind effect.