At 1718 ESuT On 5 November 2002, VH-TJG, a Boeing 737-476 aircraft, encountered turbulence 1718 ESuT during the landing flare on runway 35 at Canberra International Airport. The aircraft was operating a scheduled fare-paying passenger service from Melbourne, Victoria to Canberra, ACT. The pilot in command was the handling pilot for the flight.

At 1700, the wind direction and speed at Canberra was 280 degrees T at 18 kts, gusting to 23 kts. At 1730, it was 280 degrees T at 18 kts, gusting to 26 kts. Runway 35 was aligned on magnetic heading 348 degrees, which was equivalent to 360 degrees T.

The automatic terminal information service (ATIS) at Canberra airport provided information on the prevailing weather conditions. At the time of the occurrence, information "Sierra" was current. It included information that runway 35 was in use, and that the wind direction and speed was 270 degrees M, with a minimum speed of 15 kts and maximum speed of 25 kts.

The aircraft was equipped with a solid-state digital flight data recorder (SSFDR). The flight data plots revealed that the pilot in command applied left control wheel to achieve a left wing low attitude of about 3 degrees as the aircraft descended through a radio altitude of about 60 ft. At about 6 ft radio altitude, the aircraft suddenly rolled left to a left wing low attitude of about 6 degrees, and the pilot in command rapidly applied right control wheel input to arrest the roll to the left. The aircraft landed about one second later in a slightly right wing low attitude.

The landing was completed without further incident, and there were no reported injuries to any of the 34 occupants of the aircraft.

The pilot in command subsequently reported that the turbulence encountered during the landing flare appeared to have resulted from a hangar located adjacent to, and to the west of, the touchdown zone of runway 35.

Construction of the hangar was completed in April 2002. The airside (eastern) face of the hangar was located 283.5 metres from the centreline of runway 35, and the roof height on the airside face of the hangar was 21.7 metres.

Turbulent wake eddies may be generated downwind of obstacles by wind flowing over and around them. The turbulent effects will depend on the size and location of an obstacle or group of obstacles, such as a cluster of buildings, as well as the direction and speed of the wind. If obstacles are located close to runways, turbulent wake eddies from those obstacles may have the potential to affect safety of flight if they result in aircraft experiencing difficulties during takeoff or landing.

The International Civil Aviation Organization (ICAO) has published standards and recommended practices that relate to aircraft, personnel, airways and auxiliary services. Those standards and recommended practices are contained in various Annexes to the Convention on International Civil Aviation, which was signed at Chicago on 7 December 1944 (the Chicago Convention). Australia is a contracting State to the convention. It is obliged under Article 37 of the convention to conform to standards and to endeavour to conform to recommended practices unless a difference has been filed with ICAO.

Annex 11 contained the standards and recommended practices that relate to the provision of air traffic services to the aviation sector. Paragraph 4.3.7 of Annex 11 detailed the information to be included in ATIS messages, and paragraph 4.3.7 k) required that messages contain:

`…other essential operational information.'

Paragraph 4.3.7 s) required ATIS messages to contain:

`…any available information on significant meteorological phenomena in the approach, take-off and climb-out areas including wind shear, and information on recent weather of operational significance.'

ATIS "Sierra" contained no information regarding the likelihood of turbulence in the touchdown zone of runway 35 at the time of the occurrence.

Canberra International Airport is a licensed aerodrome, and particulars about the aerodrome are required to be published in the Aeronautical Information Publication (AIP) Enroute Supplement Australia (ERSA).

The Civil Aviation Safety Authority (CASA) publishes Civil Aviation Advisory Publications (CAAPs) which provide guidance on the preferred method for complying with the Civil Aviation Regulations. CAAP 89O-1 (2), issued in November 2000, related to publishing aerodrome information and reporting changes to that information. That information included any event that affected the safety of aircraft using the aerodrome.

At the time of the occurrence, both the ERSA and the location briefing material for Canberra aerodrome contained no information to caution pilots of the likelihood of turbulence from the hangar located adjacent to, and to the west of, the touchdown zone of runway 35.

There have been two similar events of building-induced turbulence in the UK. One resulted in a B747 leaving the runway at London Heathrow airport, and was described in the UK Air Accident Investigation Branch (AAIB) Bulletin No. 5/2002. The other involved an A300 that sustained a podstrike at London Gatwick airport, and was described in AAIB Bulletin No. 6/2002.

The National Transportation Safety Board (NTSB) of the USA has had no reported occurrences of building-induced turbulence, nor has the NTSB identified that condition as a safety concern.

There are presently no building codes or standards in Australia that address the phenomena of building-induced turbulence with respect to proposed buildings to be located on or adjacent to aerodromes.

The co-pilot of the BAe146-300 reported that he noticed both cockpit side windows were open as he entered the flight deck to commence his pre-flight preparations and almost immediately detected the presence of fumes. He stated that these fumes had a distinctive odour that he recognised as being consistent with the contamination of the cabin air supply by lubricating oil from the aircraft's engines. This typically occurs when a faulty bearing seal allows lubricating oil to escape and contaminates one of the sources of cabin air for the aircraft.

The auxiliary power unit (APU) was operating and air conditioning Pack 1 and 2 were supplying air to the flight deck and passenger cabin. As Pack 1 supplies most of the air to the flight deck and because fumes were not apparent in the passenger cabin, the co-pilot immediately suspected that the contamination was associated with Pack 1. He de-selected Pack 1 and noticed an immediate improvement to the quality of the air.

The co-pilot briefed the pilot in command of these observations when he arrived on the flight deck. The crew used engine bleed air to provide cabin air conditioning as the aircraft taxied for take-off and Pack 1 remained de-selected. This was on the basis of the tailwind conditions encountered while taxiing, to reduce the possibility of ingesting fumes from the APU's exhaust into the aircraft cabin and the co-pilot's observations of air quality from Pack 1. The take-off was performed with all air conditioning packs selected off and number 4 engine supplying bleed air for pressurisation control of the aircraft cabin.

The co-pilot was the handling pilot for the sector from Perth to Karratha. After take-off, the crew selected engine bleed air sources from all engines and Pack 1 and 2 were used to provide air conditioning to the passenger cabin and flight deck. No fumes were evident in either the passenger cabin or the flight deck and the flight proceeded normally.

About 10 minutes prior to the top of descent, the co-pilot recalled starting to experience symptoms of a headache. He was able to continue his duties as handling pilot and completed the descent to the Karratha circuit area. The aircraft was on final approach to land, when strong oil-type fumes were again detected on the flight deck. As the aircraft cabin had already stabilised at sea-level atmospheric pressure and the source of fumes appeared to be either the air conditioning packs or one of the engine bleed air supplies, all air conditioning packs and all sources of engine bleed air were selected off. The intensity of the fumes quickly dissipated and the co-pilot completed the landing. Fumes were not detected in the passenger cabin by any of the flight attendants. The co-pilot reported that supplemental oxygen was not used on this occasion due to the critical phase of flight (short final approach to land) and the prompt action taken to isolate the source of fumes.

During the taxi to the passenger terminal, the co-pilot became aware that he was experiencing symptoms of an unusually strong headache, nausea and irritated eyes, nose and throat. His symptoms quickly improved as he conducted the external turnaround duties and he felt capable of performing his duties on the return sector as pilot not flying.

The pilot in command was satisfied that the source of fumes experienced on short final was associated with contamination of the cabin air supply by engine lubricating oil. The return flight was conducted without using APU air and with Pack 1 de-selected to minimise the possible recurrence of fume contamination. The flight was completed without incident.

Subsequent to the incident, the co-pilot recalled that during the outbound sector his bottled drinking water, which had been poured and left standing in his cup, had acquired a rank swampy, slightly metallic taste. The pilot in command was also reported to have made a similar comment about the taste of his drinking water. The cups were resting in the flight deck cup holders adjacent to the air outlet vents. This was noticed prior to the co-pilot reporting symptoms of a headache and fresh cups of water from the bottle tasted normal with no apparent sign of contamination. The ATSB were not able to test the water for contaminants as it had been discarded following the flight.

Following entry of the fume report in the aircraft's defect log, company engineering personnel applied the requirements of Airworthiness Directive AD BAe 146/86, issued 30 March 2001 that required inspection of various components associated with the aircraft's cabin air supply. This inspection revealed no apparent defects or source of contamination to the cabin air.

The operator received subsequent Operating Crew Reports associated with poor cabin air quality on 23 and 26 October 2002. The aircraft commenced a period of heavy maintenance on 28 October 2002. During this maintenance the APU was removed from the aircraft, cleaned, inspected (with nil defects found) and refitted. Air conditioning Pack 1 was also cleaned. The operator received subsequent reports of flight deck odours with respect to this aircraft on 5 and 8 December 2002 and the requirements of the Airworthiness Directive were again applied. On this occasion the inspections revealed slight leakage of engine lubricating oil from a bearing seal on the number 4 engine.

History of the flight

The pilot of the de Havilland Beaver floatplane registered VH-BVA was conducting a charter positioning flight from Hamilton Island Marina to Chance Bay, Whitsunday Island. He had landed at Chance Bay seven times in the previous two days. Weather conditions in the area were good. At 1700 Eastern Standard Time the Hamilton Island automatic weather station recorded a 7 - 10 knot wind from the northwest. Witnesses in Chance Bay said that the surface wind in the bay was 2 - 5 knots. The water surface in Chance Bay was smooth, but not glassy.

The pilot said that he commenced a straight-in approach to Chance Bay but elected to go around due to the increased number of vessels moored in the bay since the previous flight. He flew a left circuit at 500 feet and assessed that the area for landing was adequate. He said that on final approach, the flight path was higher than he would have preferred. His intention was to touchdown before passing abeam the vessels. He recalled that the floatplane speed shortly before touchdown was about 80 knots, rather than the target speed of 70 knots. He said that his response at this time was consistent with flying a landplane in that he reduced the back pressure on the control column and allowed the floatplane to contact the water at a lower nose attitude, and at a higher speed, than was ideal. Upon touchdown, the floatplane yawed sharply left 50 - 60 degrees and headed directly towards the anchored ketch 'Seark', about 300 m away. The pilot said that the water rudders (at the rear of each float) were retracted, so all the yaw control he had available was via the conventional aerodynamic rudder. As the aircraft yawed, it felt as though the rudder was stuck at full left deflection, but he thought that this was due to hydrodynamic drag. When the floatplane was an estimated 100 m from the 'Seark', it swung right so that it was heading slightly to the east side of the 'Seark'. However, the outer portion of the floatplane's left wing subsequently collided with the rear mast of the 'Seark'.

A video recording of the event showed that the floatplane touched down with a lower nose attitude than was ideal. It also showed that the left float touched the water first. The sharp left yaw followed immediately. The aircraft became airborne momentarily, shortly after initial touchdown.

The left wing of the floatplane and the rear mast of the ketch were substantially damaged. There were no injuries to the pilot or the three occupants of the ketch.

Landing area

Witnesses estimated that landing floatplanes (including this accident floatplane prior to touchdown), were passing about 50 m abeam the anchored yachts.

Civil Aviation Advisory Publication (CAAP) 92-1(1) 'Guidelines for aeroplane landing areas' stated that "a minimum width water channel of 60 m ... is recommended" for single-engine and centreline thrust floatplanes not exceeding 2,000 kg maximum take-off weight (MTOW). There were no diagrams or other guidance material provided for floatplanes greater than 2,000 kg MTOW or for multi-engine floatplanes. The MTOW for the accident floatplane was 2,313 kg. The Transport Operations (Marine Safety) Regulations (Queensland) 1995 s95 (1)(a)(ii) required that vessels operating at speeds greater than 6 knots must not approach within 30 m of a moored vessel.

Float alignment

Another company floatplane pilot reported that the aircraft required greater than normal right rudder input to maintain a constant heading during flight. During an inspection after the accident, maintenance personnel established that the floats were aligned slightly left of the aircraft's longitudinal axis. The aircraft manufacturer commented that the reported flying characteristics were consistent with the float alignment. The aircraft had been flown 23.8 hours since the last maintenance inspection, including about 10 hours by the accident pilot. No record had been made in the aircraft maintenance release regarding the 'in-flight' or 'on-water' handling characteristics.

Floatplane stability on the water

The ideal landing attitude for a floatplane is nose high, so that the rear portion of the float contacts the water first. If a floatplane lands at too high a speed (lower nose attitude than ideal), the point of contact of the floats with the water (that is, the centre of rotation of the floatplane) is at a position on the floats that is forward of the aircraft's centre of gravity and the aircraft's directional stability is reduced. If the loss of directional stability is not too severe, the pilot may be able to regain control if nose-up elevator is applied very rapidly to move the centre of rotation aft, behind the centre of gravity.

Pilot information

The pilot was the chief pilot of the company. Of his 11,256 hrs total flight time, 34 hrs were in floatplane operations, and included 79 water landings.

The pilot commenced floatplane endorsement training in October 2001. The endorsement was issued on 12 February 2002, after four sessions of training in Cessna 206 floatplanes involving 4.1 hrs and 16 water landings. Between 15 March and 25 July 2002, the pilot conducted four sessions in command under supervision in de Havilland Beaver floatplanes, involving 7.3 hrs and 23 water landings. On 25 July 2002 the company floatplane training pilot authorised the pilot to conduct solo commercial operations to specific destinations but noted that he was to be closely monitored in marginal conditions until he was more experienced.

The accident flight was the last flight of the day for the pilot. He had completed five flights with a total of 3 hrs flight time and 10 hrs duty time before the accident. The day before the accident the pilot had completed 10.2 hrs duty, including 4.8 hrs flight time during eight flights. The day prior to that was a rostered day off. He reported that neither fatigue nor any other personal issues had impaired his ability to safely operate an aircraft on the day of the accident.

The pilot's work/rest history for the 14 days prior to the accident was examined using a computerised fatigue algorithm developed by the Centre for Sleep Research, University of South Australia. The results indicated that the pilot was not experiencing significant levels of fatigue in the week leading up to, and on the day of, the accident.

Organisational information

At the time of the accident, the company employed an experienced floatplane pilot who conducted all the accident pilot's floatplane endorsement and in command under supervision training. The authority, duties and responsibilities of this floatplane pilot were not established or formalised in either the Civil Aviation Safety Authority (CASA) issued instruments of approval for key company personnel, or in the company operations manual.

The company operations manual stated that pilots with less than 250 water landings were subject to the direct supervision of the chief pilot who was to take into account the wind strength, wind direction, turbulence, tide and sea state at the origin and destination of the flight before approving a flight. There was no provision in the operations manual for when the chief pilot did not have the minimum 250 water landings experience. (Other companies required up to 300 water landings in command under supervision and 50 hours total floatplane flight time before solo commercial flight operations were permitted.)

For VFR charter in single-engine aircraft, CASA required that a pilot hold a commercial licence and the applicable aircraft special design feature endorsement, for example float alighting gear. CASA did not specify additional minimum experience requirements on the aircraft type or the special design feature.

Chief pilot appointment

Civil Aviation Order 82.0 required that a chief pilot hold licences, endorsements and ratings that permit command of all company operations. CASA had approved the pilot's appointment as chief pilot for the operator on 19 April 2002. At that time, he had obtained a float alighting gear endorsement but had not been authorised by the company to conduct solo commercial floatplane operations.

While on approach to land at Perth, the crew of the Boeing 727 notified air traffic control (ATC) that they required runway 21 instead of 24. This was in consideration of the wet runway conditions and as a possible precaution should they experience antiskid problems. They also advised ATC to expect a normal approach.

After parking the aircraft, the crew reported the loss of the inner left main wheel and requested rescue fire-fighting services to attend while the aircraft was secured (see Figure 1.). An inspection of the Perth runway did not find any debris.

Figure 1. Left main landing gear missing number 2 wheel.

A search conducted at Melbourne airport, the point of departure, resulted in the recovery of debris from runway 34. The wheel had impacted the perimeter fence past the runway 34 overrun area and was recovered from that location.

Inspection of the landing gear revealed that the outer bearing had failed, which allowed the wheel assembly to migrate over the locked retaining nut and depart from the axle. Landing gear damage was confined to the sacrificial alloy axle sleeve (see Figure 2).

After completion of the Boeing Conditional Inspection following bearing failure, the alloy sleeve and main wheel were replaced. As a precaution, the operator also replaced the remaining three wheels of the main landing gears. As the nose landing gear wheels were from a different overhaul facility, it was not considered necessary to replace them prior to the aircraft returning to service. The line engineers who removed the main wheels verified that bearing dust seals were present on all the remaining wheel bearings.

Figure 2. Left main landing gear axle damage.

The inner bearing cone of the lost wheel was recovered from the aircraft landing gear and, along with the wheel bearings from the remaining three wheels, was sent to the ATSB for further technical analysis. The separated main wheel and bearing cone debris recovered in Melbourne were also forwarded to the ATSB.

Main wheel assemblies

The operator had recently acquired the aircraft on lease after it had been in long term storage in the USA. During the process of transfer to the Australian register, the American owner sourced a serviceable set of main wheel assemblies to standardise the aircraft with the rest of the Australian operator's fleet. The wheels carried release to service documentation after overhaul from an approved American maintenance facility.

The incident wheel (number 2) had been overhauled in accordance with approved data and released for service with the correct documentation on 14 November 2000. The wheel was subsequently fitted to the Boeing 727 on 31 August 2002. The wheel was in storage for all the intervening period between release to service and fitment to the aircraft. The remaining wheel (number 1) on the left landing gear had been overhauled in accordance with approved data and released for service with the proper documentation on 7 December 2001. The wheel was subsequently fitted to the aircraft on 31 August 2002.

Wheel numbers 3 and 4 on the right landing gear were overhauled in accordance with approved data and released for service with the proper documentation on 17 and 22 August 2001 respectively. The wheels were subsequently fitted to the aircraft on 29 August 2002. Although still a lengthy storage period, wheel numbers 1, 3 and 4 had been in storage for a significantly shorter period than wheel number 2.

The quality assurance inspector for the overhaul facility stated that, although not mandated, the company recommended to customers to re-inspect stored wheels every twelve months. He also stated that, after overhaul, all wheels were returned to their owners, and the company did not hold any wheels in long-term storage. As part of the shipment preparation, the overhaul company routinely packed the wheel bearings in grease and sealed them in plastic bags before they were placed in the wheel assemblies. The plastic bags were to be removed before the wheel re-entered service. With no wheel assemblies stored at the overhaul facility, the company had no requirement for a repetitive storage inspection procedure to be in place. This meant that the wheels fitted to the aircraft had been returned after overhaul to their owner (the aircraft owner) and were placed in storage at a site other than the overhaul facility. The aircraft owner had then supplied the wheels to the Australian operator for fitment to the Boeing 727.

At the time of the incident, the aircraft operator's maintenance control manual did not include a section relating to wheel storage and tracking procedures, which specifically mandated a storage life expiry interval and re-inspection procedure for wheels. The maintenance personnel who fitted the wheels to the aircraft stated that the wheels were received from the USA with the bearings in place. There were no reports of the bearings being received in plastic bags as described by the overhaul facility. There was also no paperwork supplied to the Australian operator to suggest that the wheels had been fitted to another aircraft after overhaul.

ATSB laboratory component examination

While damage to the failed bearing prevented the determination of specific factors that contributed to the failure of the particular unit, the evidence found during the investigation suggests that the development of corrosion damage over the contact surfaces of the wheel bearings was a major factor contributing to the failure. This was evident in the bearings examined from the remaining wheels. The corrosion led to insipient contact fatigue spalling, which is well understood can lead to catastrophic bearing failure consistent with the circumstances of this occurrence. The bearing lubricating grease was found to be congealed, hardened and in a dry state. A foreign material identified as brake dust was also found in the grease. The inclusion of this foreign material may also have contributed to the failure.

History of the flight

The pilot of the Cessna 182B Skylane (C182), with one passenger on board, departed Leongatha aerodrome in Victoria at about 1530 Eastern Standard Time (EST). The pilot planned the private flight in accordance with the visual flight rules to track coastal to Moruya and then via the Araluen Valley to 'Turalla', a private property located approximately 3 km northwest of Bungendore, NSW. The aircraft carried sufficient fuel for the flight.

Witnesses reported that, at about 1755, the aircraft overflew a property belonging to a relative of the pilot 3 km to the southeast of Bungendore, at about 500 ft AGL. That relative had been nominated as the responsible person to hold and cancel the nominated SARTIME of 1820. The aircraft was then observed to track west towards the Kings Highway, at a low level, and make a right turn to join a wide left base for the grass airstrip located at 'Turalla'. The airstrip was aligned approximately 305 degrees M. The aircraft was observed to turn left onto final approach at about 50-80 ft AGL. Witnesses reported that all turns were made using about 45 degrees angle of bank. Witnesses also observed the aircraft to be buffeted by gusting winds.

The weather at the time was reported by witnesses to be clear conditions with some upper level cloud. Winds were strong and blustery, from the west and northwest. A Bureau of Meteorology assessment of the weather indicated gusty north-westerly winds backing to the west after the passage of a front, which had occurred earlier in the day. The general wind structure lent itself to at least moderate turbulence and the strong possibility of lee waves and strong downslope winds.

While on short final, the aircraft was observed to make a steep climb towards a downwind position, turning to the right using 60-80 degrees angle of bank. It was then observed to lose altitude rapidly. It turned through about 295 degrees before impacting the ground approximately 300 m east-north-east of the airstrip threshold. The aircraft struck the ground at a 60-80 degree nose down, left wing low, attitude on a heading of about 240 degrees M. The aircraft was destroyed by impact forces and post-impact fire.

Injuries to persons

The passenger was fatally injured. Autopsy and toxicological tests conducted on the passenger revealed a low level of carbon monoxide in the blood. Medical opinion indicated that death occurred prior to the commencement of the post-impact fire. The pilot was seriously injured and survived the accident for 68 days before succumbing to the effects of his injuries.

Annex 13 to the Convention on International Civil Aviation, Aircraft Accident and Incident Investigation, defined a fatal injury as an injury that resulted in death within 30 days of the date of an accident.

Wreckage and impact information

A post-impact fire consumed the cockpit area. The extent of the fire damage precluded a productive examination of the cockpit, controls and instruments. Many of the alloyed components had been reduced to a molten state. Examination of the wreckage indicated that at the time of impact the aircraft was not configured for landing. The wing flaps were set to the retracted position. The engine was removed from the accident site for technical examination. The examination determined that the engine was capable of normal operation prior to impact and that it was producing power at the time of impact. Nothing was found during the investigation to suggest that mechanical failure of any part of the aircraft could have contributed to the accident.

Personnel Information

The pilot held a current private pilot's licence, was endorsed on the aircraft type and familiar with the landing area. A review of the pilot's personal flying logbook indicated that he had accumulated in excess of 340 hours total time, most of which were accrued on the occurrence aircraft. Four days prior to the accident, the pilot had satisfactorily completed a biennial flight review. The pilot held a valid Class 2 Aviation Medical Certificate. The pilot had just completed a short hiking holiday and there was no evidence to indicate any physical or psychological conditions that may have adversely affected his ability to pilot the aircraft.

Landing area

The landing area at Turalla was inspected after the accident. With the exception of a centrally located windsock, there were no aerodrome markings to detail the exact location and dimensions of the airstrip. There was evidence of recent aircraft use on the grass strip. The landing distance available was approximately 800 m, orientated approximately 305 degrees M. There was no evidence to indicate that livestock had been grazing on the paddock that contained the airstrip in the past month. A north-south ridgeline was located approximately 1 km to the west of the airstrip. At the time of the accident, the sun was setting in the west and was visible above the horizon.

Previous occurrence at the airstrip

One of the pilot's brothers and a sister were fatally injured in a separate aircraft accident (Bureau of Air Safety Investigation Report No. BO/199603734) in the vicinity of the same airstrip in 1996, when similar wind conditions were experienced. On that occasion, the Cessna U206F (C206) aircraft was climbing on departure from Turalla when the pilot turned left onto downwind. The C206 stalled at a height from which it was not possible to recover.