Two aviation occurrences in 1999, one of them a fatal mustering incident and the other a wheels-up landing, highlight some of the potential hazards of fatigue on flying performance.

Mustering accident

A newly licensed private pilot was fatally injured at Mindaroo Station in Western Australia when mustering sheep with a Cessna 172. The accident happened late in the afternoon at the end of more than eight hours of low-level flying following nine days of intense flying activity.

During the nine days, the pilot had flown 68 (tachometer) hours. The flying was both mentally and physically demanding, involving sheep spotting and low-level mustering.

The pilot, who had no formal low-level or mustering training, had to manoeuvre the aircraft in conditions that were sometimes turbulent, and was operating under constant aircraft noise and vibration. On the day of the incident, he had taken no more than a short break, which included refuelling after about four hours of flying.

It is quite possible that he was unaware that fatigue had affected his flying performance.

The pilot had exceeded the flight duty times normally permitted for a commercial operation (dealt with in Section 48 of the Civil Aviation Orders). Although these requirements are not mandatory for private operations such as this one, they are a guide to flying limits.

In the absence of any formal duty time requirement, the pilot was responsible for determining his own daily flying limitations. This was done in conjunction with the property owners, property manager and the mustering party. A typical day started at 0700 local time and the pilot worked through the day until just before last light.

Wheels-up landing incident

In this incident, the pilot of a Cessna 210 had forgotten to re-engage the landing gear circuit breaker, which had popped during the flight.

On the morning of the incident, the pilot woke at 0530 local time and started his tour of duty at 0630. The pilot had flown an Instrument Flight Rules (IFR) check flight for 2.3 hours in the morning and his performance was considered to be above average.

The pilot departed on a Visual Flight Rules (VFR) charter towards the end of the tour of duty. The pilot had pulled the circuit breaker after it popped to prevent damage to the electric motor that had continued to run. This procedure was in accordance with the Cessna 210 Operating Handbook recommendation.

On final approach, the pilot selected the landing gear down but forgot to re-engage the landing gear circuit breaker and the landing gear did not deploy. The investigation revealed that the pilot did not recall hearing the landing gear warning horn nor did the pilot notice the status of the landing gear indicator lights.

The investigation concluded that the pilot was probably suffering from a transient fatigue-related memory lapse and, unlike the incident at Mindaroo Station, was not suffering severely from accumulated fatigue. "The pilot reported that he was very tired on the day of the occurrence and he had been for some time leading up to the incident," the ATSB report said.

During the investigation, the pilot's work and rest history for the 14 weeks before the incident was examined using a computerised fatigue algorithm developed by the Centre for Sleep Research.

The results demonstrated that the pilot probably wasn't suffering severely from cumulative fatigue. Of more significance was that the pilot had been on duty for more than 12 hours and had been awake for almost 14 hours.

Effects of fatigue

Research has shown that the effects of fatigue are similar to moderate alcohol consumption. On-the-job performance loss for every hour of wakefulness between 10 and 26 hours is equivalent to a .004 per cent rise in blood alcohol concentration. Eighteen hours of wakefulness is usually considered to be equivalent to a blood alcohol concentration of .05. A person who has been awake for this length of time will act and perform as if they have consumed .05 of alcohol.

The result is significantly delayed response and reaction times, impaired reasoning, reduced vigilance and impaired hand-eye coordination.

The article 'Pilot Fatigue and the Limits of Endurance', Flight Safety Australia (April 1999), reported that fatigue makes a pilot less vigilant and more willing to accept below par performance, and a pilot begins to show signs of poor judgement. It reported that expert research into fatigue had established that it degrades a pilot's:

• Muscular strength and coordination
• Vision and perception
• Memory
• Performance monitoring
• Error management
• Decision making
• Motivation and attitudes
• Communication
• Ability to cooperate.

But the greatest single threat is being unaware that it is happening.

Before the mustering incident at Mindaroo Station, the pilot had been talking to the ground mustering party by radio as well as flying the aircraft (possibly below 500 ft). The ATSB investigation found that he had worked very long hours in a highly demanding job in which he was inexperienced.

He had received minimal training that would help him to understand the visual illusions associated with low-level flight. The investigators considered that in the absence of specific training for low level flying operations, he was probably unaware of the appropriate techniques to safely manoeuvre an aircraft at low level.

According to the ATSB Occurrence Brief (number 199903464) a human factors report noted that the pilot had worked long hours in a job in which he was inexperienced and that he probably found this type of flying both physically and mentally demanding. The report concluded that at the time of the incident the pilot was suffering from the effects of fatigue, possibly impairing his ability to safely operate the aircraft.

According to the Centre for Sleep Research's 1999 report to the Neville Committee Fatigue and Transportation it has been difficult for researchers to determine all the factors that cause and contribute to fatigue; and "determining the relative importance of these factors under different conditions has also been problematic".

However, research had concluded that when a person works long hours, for more than say 50 hours a week, there is increasing competition between restorative sleep and the other activities of daily living.

Non-work factors contribute to overall fatigue by a reduction in the opportunity for sleep and recovery. These include social factors and domestic arrangements (for example working away from home) sleep disorders and shift work.

"For example, the same roster could have quite different effects according to social circumstances," the report stated. "A 12-hour night shift might have very different consequences for an 18-year-old single male living on his own compared to a 35-year-old single mother of two toddlers without access to 24-hour childcare facilities.

"Taken together, both employees and employers have clear responsibilities with respect to managing fatigue. The basic responsibilities of both parties relate to ensuring that adequate sleep can be obtained between shifts so that fatigue does not reach dangerous levels during shifts. Thus, lack of sleep causes fatigue and sleep allows recovery from fatigue.

"Employers have a duty of care to provide safe work schedules that permit adequate time for an employee to sleep, rest and recover as well as fulfil their social and domestic responsibilities.

"Employees also have a duty of care to use their time away from work in a safe and responsible manner to ensure that they obtain sufficient sleep and recovery in order to complete their work duties in a safe and responsible manner."

How safe are you?

There are many flying organisations operating with exemptions from the requirements of CAO 48 issued by the Civil Aviation Safety Authority (CASA).

Whether you are working to the flight and duty time guidelines under CAO 48, or under an exemption, how safe are you? Are there other factors in your life that may make you more tired than usual?

Remember, the onset of fatigue is insidious.

The final report of the accident involving Qantas B747-400 VH-OJH at Bangkok, Thailand on 23 September 1999 concluded our most important investigation of an accident involving an Australian registered jet aircraft.

The investigation was one of the most comprehensive and exhaustive ever conducted by the ATSB (or its predecessor the BASI). Investigator In Charge, Mike Cavanagh, reports on the investigation itself.

The Australian Transport Safety Bureau released its report on the Qantas B747-400 runway overrun accident at Bangkok International Airport on 23 September 1999 on 25 April 2001.

The accident occurred when the B747-400 landed well beyond the normal touchdown zone and then aquaplaned on a runway that was affected by water following very heavy rain. The crew omitted to use either full or idle reverse thrust during the landing. The aircraft was still moving at 88 kts (163 km/h) at the end of the runway and stopped 220 m later in soft turf with its nose on the airport perimeter road. A precautionary evacuation was made using emergency escape slides about 20 minutes later.

Although the flight crew and cabin crew made a number of errors, many of these were linked to deficiencies in the Qantas operational procedures, training and management processes. CASA's regulations covering contaminated runways and emergency procedures were also found to be deficient, as was its surveillance of airline flight operations. Qantas and CASA either have made, or are in the process of making, significant changes in the areas where deficiencies were identified including the development by CASA of a systems-based surveillance audit approach.

The on-site phase

As the accident occurred in Thailand, responsibility for conducting the investigation fell to Thailand in accordance with Annex 13 to the International Civil Aviation Convention. As the State of registry, Australia had the right to appoint an Accredited Representative to the investigation. On the day following the accident, a team of four ATSB investigators travelled to Bangkok with the Qantas incident response team. Thai agreement to the Australian nominated Accredited Representative was received en route.

A series of meetings was held with the Aircraft Accident Investigation Committee of Thailand over the next few days. The Committee took possession of the cockpit voice and flight data recorders, examined the aircraft, and interviewed the flight crew.

Runway 21L was closed because of the position of the aircraft in the overrun area. It was necessary to reopen the runway as soon as possible so that normal operations could resume. To facilitate this, the Committee handed custody of the aircraft back to Qantas so that recovery of the aircraft could begin. By that time, aircraft recovery experts from Boeing had arrived.

The first step in the recovery involved stabilising the aircraft to prevent further movement in the very wet, muddy soil. The landing gear was removed and a gravel road sloping down from the end of the stopway to below ground level beneath the aircraft was then constructed. New landing gear was fitted and the aircraft lowered on to the road. It was then towed backwards on to the runway. The recovery process took about seven days to complete.

In the meantime, the Committee delegated investigation of the cabin safety aspects of the occurrence to the ATSB. That enabled the ATSB investigators to conduct a detailed examination of the aircraft cabin and to speak to local sources regarding post-accident events.

The Committee retained control of other aspects of the investigation and asked the ATSB to conduct readouts of the flight recorders under the Committees supervision. Four Thai investigators attended the ATSB's Canberra facility in October 1999 and supervised the readouts. On 18 November 1999, the Committee delegated the complete investigation to the ATSB. The ATSB accepted the delegation and agreed to provide the draft report to the Committee for review in accordance with Annex 13 clause 6.9 before public release.

The investigation process

In common with widely accepted international practice, the ATSB formed an investigation team consisting of a number of groups aircraft operations, flight recorders, engineering, cabin safety, and organisational issues each under the control of an ATSB investigator reporting to the Accredited Representative who acted as investigator in charge.

The function of the groups was to collect all factual information that was relevant to the groups area of investigation. As standard practice, organisations with a direct interest in the investigation (such as Qantas, Boeing, CASA, and the flight and cabin crew industrial organisations) were invited to nominate relevant experts to the groups. In some cases, the expertise and resources available within the ATSB were not sufficient for the level and volume of information required. This meant that assistance from outside organisations was requested both as participation in a group or providing specific information to the group.

Qantas provided a very high level of cooperation and substantial expert assistance and advice regarding all facets of the investigation, especially in the areas of aircraft operations, engineering and cabin safety. This level of assistance made a major contribution to the safety benefits achieved by the investigation.

From an initial assessment of the accident and post-accident events, a logical approach to the investigation seemed to be to break the task into two segments and these were:

1. The accident flight (i.e. the approach and landing) to determine the issues relating to the flight itself that led to the overrun. Aspects to be examined included:

- weather
- air traffic control
- aerodrome/runway
- crew performance
- aircraft systems
- aircraft performance in the air and on the runway
- crew procedures and training.

2. Post accident events (i.e. from the time the aircraft touched down until the precautionary disembarkation was complete) to determine any passenger or crew safety issues. Aspects to be examined included:

- cabin damage
- aircraft emergency escape and communications systems
- flight and cabin crew performance
- flight and cabin crew procedures and training
- airport emergency response
- the evacuation process.

As these tasks progressed and the picture of events emerged, it was possible to identify areas where deficiencies might have existed. These areas then became the subject of closer and more detailed examination. Eventually, this enabled conclusions to be drawn regarding the active failures that occurred.

The next step was to look at the systems behind the active failures to see if any deficiencies existed that might have set the scene, for the active failures to have occurred. The sorts of things to be examined here included how various procedures and training programs were developed and how possible hazards were identified and risks assessed. This examination centred on Qantas and CASA.

It should be noted that the investigation groups were not involved in collecting and assessing all of the factual information. Certain types of information, such as the cockpit voice recorder, had restricted access. The organisational factors group was composed only of ATSB personnel. The analysis of the factual information was undertaken solely by ATSB investigators.

By July 2000, more than 45 files (each containing 200 documents), more than 500 photographs, and over 1100 emails of information had been collected. The next step was to draft the investigation report.

Since September 1999, three ATSB investigators had been working full-time on the investigation. A number of other investigators assisted at various stages. In total, the investigation involved six ATSB investigators.

The report and review process

Writing the report was a challenging and difficult task. It was important for the document to be reader friendly, but at the same time contain enough information to justify the conclusions of the investigation. It was felt that the recommended ICAO format for accident reports was not appropriate because of the many issues involved and their complexity. The structure settled upon involved dividing the report into a number of parts, each part covering a particular aspect and, in effect, being a report within a report.

By mid-October 2000, the draft had been completed. An extensive interested party review took place to ensure factual accuracy and natural justice. A final draft was sent to the Accident Investigation Committee of Thailand on 12 February 2001.

On April 2001 the Chairman of the Committee, Air Chief Marshal Kongsak Variana, advised ATSB's Executive director that the Committee had considered the draft report and agreed without amendment. It concluded one of the most detailed world-wide investigations of a non-fatal large passenger aircraft accident.

Read the investigation report

Runway excursion, Boeing 747-438, VH-OJH, Bangkok Airport, Thailand, on 23 September 1999

There has been an aircraft accident. Debris from the wreckage is scattered throughout a 200-metre radius. Tragically, the aircrafts crew and its passengers have been fatally injured. The sound of sirens permeates the scene as police and ambulance services attend. Soon, media representatives arrive to speculate as to its causes with cameras poised to document the wreckage.

That this could happen so suddenly and wreak such devastation strikes at the heart of many people. An occurrence like this is always associated with a sense of urgency to understand its underlying features. But aircraft accidents are commonly attributable to a complex interaction of many factors and on-scene speculation rarely resembles the final conclusion. Often, long after commotion surrounding an accident has dissipated, a team of highly skilled experts continues to investigate the reasons for its occurrence and uncover the events that preceded it.

Scientific analysis of evidence

The interpretation of evidence resulting from an occurrence can require scientific analysis. This is the role of the Technical Analysis Unit of the ATSB, which investigates, often in painstaking detail, any structural, mechanical or operational factors related to aircraft accidents or incidents.

Failures of propulsion systems, landing gear or flight control structures, fractures in crankshafts, engine rods or turbine fan blades, abnormal aircraft speeds or flying operations are just some areas of investigation undertaken by the Unit.

Because there are myriad potential causes of aircraft safety breaches, the team of specialists working in the Unit approaches each occurrence with an assumption that it is unique.

"Investigations are rarely the same" said the Units Team Leader Dr Arjen Romeyn. "There are always new issues, new understandings to be gained. What were trying to do, ultimately, is get specific answers to questions surrounding an occurrence."

Questions can include: what was the mode and sequence of failures? have all components performed to their specifications? what were the mechanical settings at the time of the occurrence? what results does analysis of the residual matter furnish?

Flight recorder analysis

To answer such questions the team uses specialist equipment and apparatus. The Unit has the capacity to download and analyse data from all civil flight data and cockpit recorders (commonly referred to as black boxes) fitted to Australian-registered aircraft. Because of its ability to establish the sequence of events prior to an accident, this undertaking can provide critical information. This is particularly so in instances where accidents have resulted in a negligible amount of recoverable aircraft wreckage or where evidence is transitory, such as occurrences involving windshear.

Even in situations where significant material evidence has been recovered an investigation can be reduced by days, or even weeks, through the retrieval of information from a flight data recorder.

Equipment for this purpose includes specialised tape decks and interfaces, and both hardware and software for signal processing and enhancing.

A radio frequency-shielded audio room, designed to prevent internal and external interference, preserves the integrity of audio analysis activities. It is also in line with the Air Navigation Act 1920, which affords protection to audio captured by cockpit voice recorders from any individuals not directly associated with its analysis, as part of an investigation.

The Unit is also equipped with advanced computer graphics software with the capacity to convert recovered data into three-dimensional animations. This capability can provide a detailed graphic reconstruction of a flight, allowing the examination of any sequence of events, from any perspective, and at any time. The benefits of this technology were demonstrated in the investigation of the much-publicised overrun of QF1 at Bangkok Airport, which occurred on 23 September 1999. Animations of the flight used for the investigation were subsequently aired on commercial television.

Materials failure analysis

Often microscopic features provide corroborating or conclusive evidence in the determination of failed components. They can also be vital to the detection of manufacturing assembly, maintenance or operational abnormalities, such as fractures in engine mechanisms or defects in airframe components.

Microscopes utilised in the Unit include: a low-power stereo microscope for general observation, which has the capacity for magnification of up to 50 times; a reflected-light microscope for the examination of the internal structures of materials, which has the capacity for magnification of up to 1000 times; and a scanning electro-microscope which magnifies from 14 to 300,000 times the actual size of an object. In addition, this microscope has an x-ray analysis facility for determining the chemistry of small material items.

The team approach

According to Dr Romeyn, while the array of equipment used in the laboratories is impressive, the Units most important assets are the highly skilled investigators who staff it.

"There is a perception that, because we work in a technical area, it's the equipment that does the work and were just operators. To do our job we need particular tools, but that's all they are. It's the understanding of what the tools allow us to see that's important", said Dr Romeyn.

Core skills necessary to undertake the work required of the Unit include a high degree of understanding in the ways mechanisms operate and their environmental affects, an appreciation of design issues, an awareness of how structures function and the ability to identify failure modes.

These skills are reflected in the academic backgrounds of the Units five investigators which comprise advanced qualifications in metallurgy, aeronautical engineering and electrical design engineering. According to Dr Romeyn, however, while knowledge of these areas is vital, it is not in itself sufficient.

"Safety investigative work is a complex system and its the depth of understanding that is important. You don't gain that just by doing a degree. It's a continual learning process and experience is an essential component of the success of our work", said Dr Romeyn.

Dr Romeyn also acknowledges the importance of contributions made from other areas of speciality. In any investigation a range of skills are applied, and this is just one skilled area. It is very important to talk to a wide range of people. Investigators with expertise in such areas as cabin safety and human performance, as well as individuals from the wider aviation industry, can be vital sources of information. It's the coming together of experience that provides the basis for fruitful investigation, said Dr Romeyn.

Often pro-active measures are initiated from work performed by the team. On 13 October 2000, while on a climb out of Hobart, a Boeing 737 experienced a dramatic malfunction in one of its engines which caused a reaction consistent with explosion. The aircraft landed safely and its pilot and passengers were unharmed. By analysing the factors surrounding the incident, the team identified deficiencies in a procedure used to repair cracks in turbine blades. Pursuant to these findings, the operator of the aircraft modified repair procedures to prevent recurrence.

According to Dr Romeyn, initiating such improvements to existing safety defences is a critical aspect of the work of the Unit.

"In the context of our work, pro-active investigations are those directed at events which haven't threatened safety directly but have the potential to do so. We know that little things can trigger big accidents. In a way, we operate as independent auditors of the aviation system", said Dr Romeyn.

Aircraft accidents and incidents can have significant, immediate and long-term affects on those involved. The determination of underlying factors takes time, and months can lapse between an occurrence and the official release of findings related to it. However, investigations into occurrences, such as those undertaken by the team of the Technical Analysis Unit, can furnish illuminating explanations as to what went wrong and how safety can be improved.

The ATSB collects and analyses data from accidents and incidents involving aircrew, ground personnel and passenger safety. In this issue of the ATSB Supplement, a selection of Australian cabin safety occurrence briefs are summarised and one from the Transportation Safety Board of Canada.

Photographs of the burnt out Saudi Arabian Airlines Lockheed Tristar at Riyadh on 19 August 1980 following an emergency landing. All 287 passengers and 14 crew on board died from smoke inhalation from a fire in the aft cargo hold which started shortly after take-off. Despite the successful landing the crew were unable to open the doors. Emergency services took 20 minutes to open one door. A serious breakdown of crew coordination was cited as one of the significant factors in the disaster.

This report from the Transportation Safety Board of Canada (A99AO046) highlights the need for continued care and vigilance in the use of ground-handling equipment to ensure safe movement to and from aircraft for passengers, aircrew and ground personnel.

In March 1999 a five-year-old child was injured during disembarkation from a B767 at a Canadian airport. The aircraft was parked on the open ramp away from an aerobridge.

After the first 10 passengers had left the aircraft a flight attendant exited the aircraft carrying an infant in a car seat. When the flight attendant stepped on to the passenger stand he noticed it was descending slowly away from the aircraft. As he turned to tell the in-charge flight attendant, the infants five-year-old brother, who was following with his mother, stepped out of the aircraft and fell between it and the stairs to the apron below. The child suffered a broken arm and lacerations to the head in the fall and was taken to hospital for treatment and observation.

The locking mechanism used to hold the upper stairs in position is a fairly simple mechanical device. The pawl that prevents the stairs from descending is held in place against the dog rail by a spring and released by energising a solenoid. In this occurrence the pawl had only partially engaged the dog rail and after several passengers had travelled over the stairs had slipped off. This allowed the upper stairs to descend away from the aircraft. According to the report it was unclear whether this was due to a weakness in the spring, a mechanical resistance in the mechanism or a combination of both. In any case proper functioning of the locking mechanism was impeded.

Investigation findings:

The locking mechanism was not functioning properly and as a consequence disengaged and allowed the upper stairs to descend away from the aircraft. There was no policy in place requiring the passenger stand operator to do a close visual inspection of the locking mechanism to ensure full engagement.

Passenger stand operators reported that they would take only a cursory look at the locking mechanism when leaving the vehicle. Any visual inspection would have been impeded because the pawl, the dog rail, and the background were all painted the same dark green colour and on this particular vehicle a support brace impeded the operators view. Operators of the passenger stand reported that they had not received formal training on the operation of the equipment.

Other contributing factors to the occurrence were the failure to follow the maintenance schedule and the absence of a requirement to visually inspect the locking mechanism of the passenger stand before use.

Safety action taken:

Since the occurrence the company has completed a comprehensive inspection of all company passenger stands. All pawl mechanisms were painted in contrasting colours to facilitate determination of the pawl position and support braces were relocated to prevent the impediment of the operators view of the pawl. All airstairs units were put on a weekly follow-up routine to ensure all checks are completed on time.

The company, the TSB and the Canadian regulator Transport Canada, have disseminated details of the occurrence to local and international air transport operators regulators and industry associations to alert other operators using similar equipment of the potential for injury and the steps that may be taken to avoid similar occurrences.

Occ No. 200100741, 22 February 2001

At top of descent to Los Angeles the cabin crew of a Boeing 747 aircraft reported smoke and fumes emanating from the cabin ceiling located in the vicinity of the rear right side (R5) emergency exit door. Smouldering paper tissues were found in an overhead light fitting. Cabin crew removed the tissues and discharged a fire extinguisher onto the light fitting, tissues and surrounding area. The cabin crew remained in the vicinity and monitored the area until passengers disembarked at Los Angeles.

The company reported that the light fitting is a night light and is always on. The light has a blue plastic cover that should always be in place, and which was not fitted on this occasion.

The investigation was unable to determine why or who placed the tissues in the light fitting.

Safety action:

The company issued an Important Information bulletin to flight attendants advising that any visible cabin light fitting must have a protective grill or glass covering the bulb.

Occ no. 200104168, 21 August 2001

During the cruise the passenger seated in 56C was warned several times for lighting cigarettes. Most cigarettes were extinguished and confiscated by the crew but one was dropped and ignited a blanket. The cabin crew members were quick to extinguish the smouldering blanket. The passenger was off-loaded in Bangkok.

Occ No. 200104464, 5 Sept 2001

During a flight between Melbourne and Sydney a smouldering fire was detected and extinguished in the waste bin of the aft toilet of the aircraft. A particular passenger was strongly suspected of smoking in the toilets during flight and the pilot in command requested that security staff meet the aircraft upon arrival in Sydney. The aircraft landed without further incident.

Occ No. 200103578, 10 July 2001

The aircraft was on climb passing FL200 when a passenger sustained a head injury from a bottle of liquor that was accidentally dropped from an overhead locker by another passenger who was removing a piece of luggage. The injury was treated immediately by the cabin crew to stop the blood flow. A paramedical team met the aircraft on arrival at Rome.

Occ No. 200103478, 15 July 2001

During disembarkation a passenger was struck on the head by a metal scooter that fell from on overhead storage bin. The passenger received a bleeding cut to the head, was given first aid and attended by the Rescue Fire Fighting Service. The passenger was later transported to a local medical centre for treatment.

Occ No. 200100393, 24 Jan 2001

During the cruise cabin crew were required to abruptly cease cabin service when the flight crew turned on the fasten seat belt sign due to severe turbulence associated with thunderstorm activity. They were not able to secure the cabin prior to landing and as a result the aircraft landed with the cabin insecure. The pilot in command reported later that he did not consider it safe to turn the sign off during the descent.

Occ No. 200103943, 8 August 2001

During the cruise a passenger seated in 20C was struck on the head by a plastic bottle full of water, which had been stored in the overhead locker by a cabin crew member. The passenger later collapsed, became ill and required medical attention. An ambulance was organised to meet the aircraft on arrival at Darwin.

Occ No. 199902180, 24 April 1999

The aircraft was cleared for take-off when the flight attendant advised the pilot that a cat had escaped from a cage in the cargo hold and was loose in the cabin. The flight attendant locked the cat in the toilet while the pilot returned the aircraft to the ramp. The cat was removed through the toilet door without further incident.

Occ No. 200102090, 3 May 2001

During the cruise the crew noticed smoke in a rear toilet. The cabin crew found a smouldering tissue box that appeared to have been used to extinguish a cigarette and then water used to extinguish the potential fire. At the time the no smoking sign was extinguished.

With the development of airborne collision avoidance systems (ACAS) and their fitment in aircraft since the mid-nineties it has become possible for pilots to know if their aircraft is on a collision course with another.

When an ACAS warning is received the pilot or crew has time to take avoiding action. Some of the systems fitted in aircraft today will advise what to do - climb or descend away from a conflicting aircraft. Future developments will also give turn advice.

The effectiveness of ACAS is totally dependent on the presence of an operating Mode C or Mode S (altitude encoding) transponder in the intruding aircraft.

ACAS can be active or passive

The two most common ACAS systems are:

• TCAS: The Traffic Alert and Collision Avoidances System (TCAS), depending on its level of sophistication, can give three levels of warning. Traffic information, where it can 'see' traffic; a traffic advisory where aural or visual warnings will alert to the possibility of conflict; and a resolution advisory, an aural alarm which will alert to impact in 20 seconds.
• TCAD: A Traffic and Collision Alert Device (TCAD) is a passive system that requires a third party to provide the response from transponders. It will identify a target aircraft if it has a transponder, which is turned on, and if an independent activator, such as a ground-based radar or an airborne active system like TCAS, has activated that transponder.

When it works well

In the following incident outside controlled airspace, the system worked and two aircraft avoided further miss-hap. It is also a good example of the enhanced 'safety net' because air traffic services did not fully appreciate the unusually busy and complex traffic disposition. On 27 January 1999 an Instrument Flight Rules (IFR) Airtrainer departed Moruya for Tamworth climbing to 6,000ft and estimated Bindook at 1842. An IFR Aerostar departed Young for Bankstown on climb to 7,000ft.

Neither aircraft was provided with traffic information when the Airtrainer elected to climb to 7,000ft which would put it into conflict with the Aerostar north of Bindook.

The crew of the Airtrainer subsequently reported having passed Bindook at 1843 leaving 6,000ft for 7,000ft. The pilot of the Aerostar reported passing abeam Bindook at 1845 maintaining 7,000ft.

The Airtrainer crew, by this time in Instrument Meteorological Conditions, had a TCAD alert, which indicated an impending conflict with another aircraft at 7,000ft about 10NM north of Bindook.

The pilot initiated a rapid descent to 5,500ft and turned away from the unknown traffic. The ATSB's investigation and radar analysis determined that at the time of the TCAD alert the aircraft were within 3NM of each other and closing with only 100ft vertical separation.

In another example, the incident again highlights the advantage to pilots of increased situational awareness while outside controlled airspace. On 5 May 1999 while on descent to Proserpine in class G airspace, the crew of a BAe 146 received a TCAS traffic advisory on a slower aircraft below and ahead of them. Although transmissions were made they were unable to make radio contact with the aircraft. The crew used the TCAS information to take appropriate avoiding action.

When things go wrong

Near Port Hedland on 23 November 1999 a Visual Flight Rules (VFR) aircraft outside controlled airspace passed a Twin Otter within an estimated 20 - 50 feet. The pilot chose to cruise the Cessna 310 at an IFR level in class G airspace, and not being subject to a directed traffic information service, no-one knew he was there.

There was no time for the Twin Otter crew to take evasive action. While maintaining the same IFR cruising level they only saw the other aircraft when it passed them travelling in the opposite direction. The aircraft was not fitted with an ACAS.

A similar incident took place inside controlled airspace. On 25 May 1999 a Boeing 737 inbound in cloud to Hamilton Island was conducting a VOR/DME instrument approach while a Cessna 182 was on climb from Shute Harbour for a parachute drop at 10,000ft. When the B737 was established on final and visual, the pilot and parachutists in the C182 sighted it in a left banking turn in their two o'clock position at the same level with less than 100m of lateral separation.

According to the ATSB report the B737 crew were unaware of the near mid-air collision with the Cessna. 'The TCAS did not alert them because the Cessna's transponder was turned off. According to the Aeronautical Publication (AIP) Australian ENR 1.6-8, the pilot of the Cessna was required to have activated the transponder on the selected code 1200,' the report said.

The message then, is pilots need to turn their transponders on in whatever airspace they are flying.

Increase in proximity warnings

Between 1 January 1993 and 19 September 1994, 47 near misses were reported to the ATSB where two or more Regular Public Transport (RPT) aircraft were involved.

TCAS Resolution Advisory occurrences

Between 19 September 1994 and 25 May 1995 there were 47 occurrences in controlled airspace where an infringement of separation standards involving aircraft not equipped with TCAS occurred. In addition there were 10 cases where TCAS was fitted and had activated and assisted crews in their decision-making.

In the same period 29 occurrences were reported outside controlled airspace where an ACAS was not fitted, and in the ATSB's opinion, an ACAS could have assisted in situations where aircraft came into conflict.

Air Safety Interim Recommendation IR19950117 of 4 May 1995 said, 'The fitment of a TCAD in some general aviation aircraft had led to three alert situations outside controlled airspace.

'In two of them, the other aircraft was sighted and avoiding action was taken. In all three cases the installation of the TCAD improved the options of the pilots and gave them timely advice for avoiding a potential near-miss.'

Since 1995 the ATSB has received almost two thousand reports of events where the proximity to another aircraft was considered to be a hazard. It has investigated more than 350 occurrences in all classifications of airspace where it considers that ACAS (or would have if fitted in the aircraft) significantly improved situational awareness for flight crews.

By the year 2000 occurrences where an ACAS would improve situational awareness outside controlled airspace had increased to 40 compared to 60 inside controlled airspace. In 1997 there were only two reported ACAS occurrences outside controlled airspace. The increase is mostly due to the fitment of ACAS in aircraft that were previously not equipped.

"The result is that crews of aircraft today have the ability to 'see' other aircraft. Consequently occurrences reported today were generally not known about prior to 1997. It also shows that a potential collision has always existed outside controlled airspace," said Bernie Rodgers, one of the ATSB Senior Transport Safety Investigators tasked with analysing air safety incident reports.

VCAs: a reality

Violations of controlled airspace (aircraft entering controlled airspace without a clearance) continue to occur in significant numbers every year. Sometimes air traffic control has been initially unaware of it.

Violations of controlled airspace 1993-2000

"An aircraft that inadvertently enters controlled airspace with its transponder on is more likely to avoid conflicting with a fully loaded passenger aircraft, which is fitted with ACAS. The frequency of VCAs has not diminished so it is reasonable to assume they will continue," said Mr Rodgers.

ACAS has already proven its worth in a VCA situation as the crew of a Boeing 737 discovered. When on approach to Melbourne at 3,000ft on 26 July 1999 they received a TCAS traffic advisory on an aircraft that had infringed controlled airspace at 2,300ft. The air traffic controller had not noticed the intrusion.

The crew were able to use the information provided by TCAS to sight the aircraft and maintain visual separation until they were clear.

The future

It is apparent given the increasing numbers of reported conflicts since the mid-nineties that near misses both inside and outside controlled airspace do occur and more often than previously thought.

This reinforces the earlier Safety Advisory Notice (SAN 941261) issued on 30 September 1994 to the former Civil Aviation Authority suggesting a timetable be introduced to mandate the fitment and use of ACAS equipment.

On September 1997 Australia was party to a regional agreement that ACAS would be fitted to all turbine-powered aircraft above 15,000kgs maximum takeoff weight with more than 30 passenger seats effective from 1 January 2000.

The International Civil Aviation Organisation (ICAO) has set the following time frame for introduction of ACAS 11 to aircraft engaged in international operations:

• 1 January 2003 for all turbine-powered aircraft with a maximum certified takeoff mass in excess of 15,000kgs and more than 30 passenger seats; and
• from 1 January 2005 for all aircraft in excess of 5,700kgs takeoff mass and more than 19 passenger seats.

The ATSB will continue to monitor developments as the Civil Aviation Safety Authority considers what further actions are necessary to increase the effectiveness of ACAS in Australian airspace.

Will you make sure that your transponder is turned on next time you fly?

Melbourne Airport photograph by Leigh Atkinson, courtesy of Airservices Australia, used in head photo-illustration.

In Australia, mountain waves are commonly experienced over and to the lee of mountain ranges in the south-east of the continent. They often appear in the strong westerly wind flows on the east coast in late winter and early spring.

Mountain waves are a different phenomena to the mechanical turbulence found in the lee of mountain ranges and can exist as a smooth undulating airflow or may contain clear air turbulence in the form of breaking waves and 'rotors'. Mountain waves are defined as 'severe' when the associated downdrafts exceed 600 ft/min and/or severe turbulence is observed or forecast.

'Breaking waves' and 'rotors' associated with mountain waves are among the more hazardous phenomenon that pilots can experience. Understanding the dynamics of the wind is important in improving aviation safety.

Glider pilots learn to use these mountain waves to their advantage; typically to gain altitude. However, some aircraft have come to grief in those conditions. Encounters have been described as similar to hitting a wall. In 1966, clear air turbulence associated with a mountain wave ripped apart a BOAC Boeing 707 while it flew near Mt. Fuji in Japan. In 1968, a Fairchild F-27B lost parts of its wings and empennage, and in 1992 a Douglas DC-8 lost an engine and wingtip in mountain wave encounters.

Mountain waves are the result of flowing air being forced to rise up the windward side of a mountain barrier, then as a result of certain atmospheric conditions, sinking down the leeward side. This perturbation develops into a series of standing waves downstream from the barrier and may extend for hundreds of kilometres over clear areas of land and open water.

Mountain waves are likely to form when the following atmospheric conditions are present:
• the wind flow at around ridge height is nearly perpendicular to the ridge line and at least 25 kts
• the wind speed increases with height
• there is a stable layer at around ridge height.

If the wave amplitude is large enough, then the waves become unstable and break, similar to the breaking waves seen in the surf. Within these 'breaking waves', the atmospheric flow becomes turbulent.

The crests of the waves may be identified by the formation of lenticular clouds (lens-shaped), if the air is sufficiently moist. Mountain waves may extend into the stratosphere and become more pronounced as height increases. Some pilots have reported mountain waves at 60,000 feet. The vertical airflow component of a standing wave may exceed 8,000 ft/min.

Rotors or eddies can also be found embedded in mountain waves. Formation of rotors can also occur as a result of down slope winds. Their formation usually occurs where wind speeds change in a wave or where friction slows the wind near to the ground. Often these rotors will be experienced as gusts or windshear. Clouds may also form on the up-flow side of a rotor and dissipate on the down-flow side if the air is sufficiently moist.

Many dangers lie in the effects of mountain waves and associated turbulence on aircraft performance and control. In addition to generating turbulence that has demonstrated sufficient ferocity to significantly damage aircraft or lead to loss of aircraft control, the more prevailing danger to aircraft in the lower levels in Australia seems to be the effect on the climb rate of an aircraft. General aviation aircraft rarely have performance capability sufficient to enable the pilot to overcome the effects of a severe downdraft generated by a mountain wave or the turbulence or windshear generated by a rotor. In 1996, three people were fatally injured when a Cessna 206 encountered lee (mountain) waves. The investigation report concluded, "It is probable that the maximum climb performance of the aircraft was not capable of overcoming the strong downdrafts in the area at the time".

Crossing a mountain barrier into wind also reduces the groundspeed of an aircraft and has the effect of keeping the aircraft in the area of downdraft for longer, while an aircraft flying downwind on the upwind side of a mountain range is likely to initially encounter updrafts as it approaches rising ground. Rotors and turbulence may also affect low level flying operations near hills or trees. In 1999, a Kawasaki KH-4 hit the surface of a lake during spraying operations at 30 feet. The lack of sufficient height to overcome the effects of wind eddies and turbulence was a factor in the accident.

Research into 'braking waves' and 'rotors' or eddies continues but there is no doubt that pilots need to be aware of the phenomenon and take appropriate precautions. Although mountain wave activity is usually forecast reasonably well by the Bureau of Meteorology, many local factors may effect the formation of 'breaking waves' and 'rotors'. When planning a flight, a pilot should take note of the winds and the terrain to assess the likelihood of waves and rotors. There may be telltale signs in flight, including the disturbances on water or wheat fields and the formation of clouds, provided there is sufficient moisture for cloud to form.

Prudent flight planning may include allowing for the possibility of significant variations in the aircrafts altitude if updrafts and downdraughts are encountered. A margin of at least the height of the hill or mountain from the surface should be allowed, and consideration given to the need to adopt a manoeuvring airspeed appropriate to the circumstances. Ultimately, it may be preferable for pilots to consider diverting or not flying, rather than risk flying near or over mountainous terrain in strong wind conditions conducive to mountain waves containing 'breaking waves' and 'rotors'.

Illustration of mountain wave and associated turbulence

Further Reading
Bureau of Meteorology. (2007). Manual of Aviation Meteorology. Second Edition, pp 59, 60, 68. Airservices Australia.

Bureau of Air Safety Investigation Journal. (1991, September). Downslope winds are dangerous. BASI Journal, 9, pp 38-39.

Jorgensen, K. (undated). Mountain flying: A guide to helicopter flying in mountainous and high altitude areas. Westcourt, QLD: Cranford Publications.

Lester, P. F. (1993). Turbulence: A new perspective for pilots. Englewood, CO: Jeppesen Sanderson.

McCann, Donald W. (2006). Diagnosing and forecasting aircraft turbulence with steepening mountain waves. National Weather Digest, pp 77-92.

New Zealand Civil Aviation Authority (2006), Good Aviation Practice, Mountain Flying booklet.

Welch, John, F. (Ed.). (1995). Van Sickles modern airmanship (7th Ed). New York, NY: McGraw-Hill.

Woods, R. H., & Sweginnis, R. W. (1995). Aircraft accident investigation. Casper, WY: Endeavor Books.

Revised: 29 October 2009.

ATSB's air safety investigator, Mike Watson, in his unique style, discusses the insidious dangers of carburettor icing.

The aircraft was on short final for runway 29L when the pilot made a brief Mayday call. The aircraft was then observed to land in a car-yard, short of the runway. Both occupants managed to evacuate without injury.

The pilot later reported that the engine did not respond when an increase in RPM was required, as the aircraft was undershooting the approach. The aircraft subsequently collided with a fence, short of the runway.

Weather conditions at the time were conducive to severe carburettor icing at descent power. It is likely that carburettor icing occurred during the low power descent and precluded the engine accelerating above idle power on the final approach.

If I were to stuff a gag forcibly down your throat, you would not be able to get air into your lungs, and after quite a short time, your body would stop working. The same is true of aircraft engines: if I were to block their air intakes, they would also stop working.

The easiest way to block an engines air intake is to freeze water and simply choke the engine, so that it can no longer breathe.

Can this happen to my aircraft? Yes. Let us look at how water can find its way into the air intake when we least expect it. To do so, we need to examine how water is carried in the atmosphere and how it can choke a carburettor.

Water is dissolved in the air that both we and our engines breathe, in much the same way as sugar can be dissolved into a cup of tea. It is much easier to dissolve sugar into a hot cup of tea than a cold cuppa, and likewise it is easier to dissolve more water in warm air than into cold air. Water that has been dissolved into the atmosphere is actually a gas that you cannot see, and it is always present in the atmosphere.

Let us take a hot cup of tea, stir in as much sugar as we can, and then put the cup in the fridge. Once the tea has chilled, you will see that some of the sugar is no longer dissolved in the tea but has formed crystals of sugar in the cup.

In the same way, if you take a cup of warm, humid air, (lots of water dissolved in it), and cool it, you will see that some of the water that was dissolved in the air as a gas will change back into a liquid. Normally, this can be seen is as tiny droplets like those found in a cloud. Many clouds are formed in exactly this way: as humid air rises and cools, it cannot hold all its dissolved water, and some of the water condenses into a cumulus-type cloud.

How will this affect the engine in your aircraft? When air passes through the carburettor on the way to the engine, fuel is evaporated into the carburettor. This chills the air, in just the same way as evaporating water chills a swimmer leaving the ocean for the beach. If this chilled air was previously humid, then some of the water dissolved in the air will immediately change into cloud-type water droplets. If the chilling effect of the fuel was sufficient to cool the carburettor below freezing level, then when these water droplets hit the sides of the venturi (the part where the air passes through), or the throttle valve, the water droplets will freeze in place. This will start the process of choking the engine. Eventually, if the process is allowed to continue, it will no longer be able to breathe, and the engine will stop.

The problem will be more pronounced if the engine is operating at a low power setting. In this case, the airflow through the carburettor will be partially impeded by the throttle valve. This valve not only provides more area for the ice to form: it also increases the partial vacuum downstream of the valve, and that will cause a further chilling of the air and the water droplets.

It is interesting to note that although fuel does act as a refrigerant in a carburettor, it is also needed to keep the engine running. When your aircraft is flying in cruise, the engine should normally be leaned with the mixture control. If this is not done, then not only are you using more fuel than you need to, but you are also putting more refrigerant into to the carburettor airflow, thus increasing the likelihood of carburettor icing. This is yet another good reason for using correct procedures when controlling the engine!

Even at temperatures exceeding 25 degrees Celsius, air passing through a carburettor may form ice that can choke your engine. The more humid the air in which your aircraft is flying, the more likely it is that ice will form in the air-intake system.

Following a normal climb, the pilot dropped two parachutists over Hamilton Island. A power-off descent to circuit height followed. The pilot did not select carburettor heat during the descent. When on a long final approach, the pilot attempted to arrest a high descent rate with the use of engine power. The engine failed to respond. The pilot found that the aircraft was outside gliding range of the runway. Engine trouble checks failed to restore power to the engine. The aircraft was ditched in shallow water and after a successful escape from the cabin, the pilot was picked up by an island launch.

Bureau of Meteorology data showed that the relative humidity at ground level was 65 per cent. A carburettor icing-probability chart showed that serious icing at descent power was to be expected at such a humidity level.

How do I recognise the start of this problem? The best solution is to be on the lookout for carburettor icing at any time the air temperature is less than 30 degrees Celsius. If an engine is being choked by ice, then its power will be reduced. However, this is not always easy to detect in the early stages, particularly if the engine is operating at reduced power settings or if the air is humid.

Application of carburettor heat for a short time will melt any ice, and when the carburettor heat is turned off again, you will see an increase in engine power for the same throttle setting. If this happens, then apply the carburettor heat, and leave it on!

Textron Lycoming, the engine manufacturer, point out that a pilot should expect a delay of 30 seconds to several minutes while ice is melted after carburettor heat is applied. During this time, rough running and a further reduction in power can be expected. It is much better to experience a small reduction in power because of the application of carburettor heat, than to experience a large reduction in power because of the engine being throttled by ice!

If you are flying a carburetted engine with a constant speed propeller, such as a Cessna 180 or 182, then you will not detect the onset of carburettor icing by a change in RPM. The manifold air pressure (MAP) is measured between the carburettor and the engine air inlets, so if the inlet is being blocked by ice and the engine is still trying to suck-in air, there will be an increased vacuum in the inlet manifold. This can be seen as a decrease in the manifold air pressure indication, when there is no other good reason for it happening.

It's a bit like your lungs being the engine, your lips the carburettor, and your cheeks the manifold air pressure gauge. If you breathe normally through your mouth past your lips, the air pressure in your mouth is nearly the same as atmospheric pressure. As an analogy, think of when your friendly neighbourhood murderer sneaks up behind you, and puts his hand over your lips in an attempt to suffocate you. He is doing the same to you as the block of ice in the carburettor is to the engine. There will be a significant vacuum in your mouth, (sucking in of the cheeks) as you desperately try to suck in your last breath, like your aeroplanes engine desperately trying to suck in the air it needs past the ice blockage to the carburettor.

Carburettor icing can sneak up on you when you are cruising along. In my case, I've found that can happen as dusk approaches, and the air cools, making the atmosphere more humid. It's always worth carefully setting the correct power setting, and noting it, so that if the RPM or the MAP starts to slowly reduce, and there's no other good reason for it, like climbing, then you can immediately suspect icing and do something about it. It's best to keep an eye out for it.

Carburettor icing a contributing factor? How does ATSB know if carburettor icing is a contributory factor of an accident? This is often difficult to answer because ice melts, it leaves no evidence. It is usually a case of elimination: if the engine is OK, there is plenty of fuel and all the controls are in the right place, then the investigators will look at weather conditions at the time. All we can usually say is that there was no other good reason for a loss of power. It always seems a shame to come across such a case, where everything was working fine, only to find that an aircraft has been downed for such an easily preventable phenomenon.

During a test flight, on short final approach, the aircraft encountered windshear. The engine failed to respond to throttle application. The aircraft landed heavily, ran into a fence and overturned.

Post-accident inspection of the engine did not reveal any mechanical reason for the lack of response to throttle application. Information from the Bureau of Meteorology showed that conditions were conducive to the formation of serious carburettor icing at any power setting. The pilot thought that because carby heat was only applied for about 10 seconds, carburettor ice was the only reasonable explanation for the loss of power.

Am I likely to experience carburettor icing? Provided with this article is a chart that will help you to work out the likelihood of experiencing icing, based on information from your forecast. You will need to find the temperature and dewpoint, and these can be found in a meteorological aviation report (METAR), or a SPECI, or a TTF type forecast. Plot the dew point depression against the temperature on the chart, and you will see an indication of the likelihood of experiencing carburettor icing.

Remember, note the air temperature: the most severe icing will occur at temperatures up to around 20 degrees Celsius, and the severity will decrease slowly as the temperature increases. The other major factor is the humidity in the air. If the air feels muggy, it is humid; if perspiration does not dry rapidly off your body, it is humid; if a breeze does not cool you on a warm day, it is humid.

When you are flying, remember that the air gets cooler with an increase in altitude, and this can increase the humidity. If you are flying near clouds, then the air is likely to be humid, (the relative humidity in a cloud is normally 100 per cent).

If you aren't sure, check for carburettor icing by applying full carburettor heat for a short while, and checking for an increase in power after it is removed.

Prevent carburettor icing at the first indication, rather than leave it until the engine is choked by ice!

How do I obtain a METAR? Use NAIPS, or the pilot briefing page on the website www.airservicesaustralia.com/(Opens in a new tab/window), but you will need to arrange yourself a username and password first. Some aerodromes have an aerodrome weather information service (AWIS), which is available from ERSA.

A full listing of METARs is also available, by State or Territory from the Bureau of Meteorology on the internet at Aerodrome Weather Reports (METAR/SPECI) (bom.gov.au)(Opens in a new tab/window) and using the user ID and password provided on that page.

In 1991 ATSB's predecessor (BASI) published a research report titled Limitations of the See-and-Avoid Principle. This report concluded that 'the see-and-avoid principle in the absence of traffic alerts is subject to serious limitations'. Unalerted see and avoid has a 'limited place as a last resort means of traffic separation at low closing speeds' and is 'completely unsuitable as a primary traffic separation method for scheduled services'.

Nevertheless, operations in a number of types of airspace currently require the application of see-and-avoid techniques by the pilots of both visual flight rules and instrument flight rules aircraft operations. In areas such as mandatory broadcast zones, pilots should be assisted by radio calls from all other aircraft to provide an 'alerted' see-and-avoid environment. However, the final level of protection is provided by pilots being able to see potentially dangerous traffic in time to take avoidance action.

The report highlighted the fact that 'many of the limitations of see-and-avoid are associated with physical limits and human perception' and encouraged pilots to be 'made aware of the limitations of the see-and-avoid procedure, particularly the factors which can reduce a pilot's effective visual field'.

Each year ATSB investigates incidents where aircraft have come perilously close whilst operating in weather conditions well above the visual meteorological conditions minima. Some of these incidents occur in the circuit area, where pilots should have had an acute awareness of the position of all traffic at all times. Incidents also occur where aircraft were established in an en-route cruise. Given that there indeed is a lot of sky out there, there is often an understandable tendency during the cruise to be less assiduous in maintaining a lookout. The following paragraphs address the issue of detecting other aircraft during an en-route cruise by examining some of the problems of lookout or visual

The Avgas contamination event that happened over Christmas 1999 caught everyone by surprise. It had not been seriously considered as a potential hazard to aviation anywhere in the world, therefore the consequences had not been considered. The reasons behind why the fuel became contaminated were unexpected. Mike Watson, one of a team of transport safety investigators who had the task of sifting through an overwhelming amount of data and publishing the final report, gives some insight.

No one was hurt as a result of contaminated aviation fuel, and there were no accidents that could be attributed to a loss of power caused by fuel contamination. At the time of the crisis the fuel refiner responded immediately and recalled all Avgas that had been manufactured at the refinery, and CASA grounded all Avgas powered aircraft that could have been contaminated until it was known that they were safe to fly.

The chemical contaminant is now known to have been ethylene diamine. At the time of the event, there was a concerted effort to define what the contaminant was (concentration in the Avgas was low); how the contaminant had got there; and what the contaminants behaviour would be in an aircraft fuel system.

In the initial response a method to guarantee aircraft would be safe again was developed, and a testing process to detect ethylene diamine was also developed in a number of weeks. Components for the test kits were sourced from all over the world.

The ATSB's investigation looked at what had happened. It looked at what could have prevented it from happening and why it didn't. It also looked at lessons that could be learnt and applied to other aviation systems. This included what would have happened if a similar contamination event occurred in a large turbine-engine passenger aircraft operating with contaminated jet fuel.

The main defence against any safety-critical system failure in an airliner is to have backup, or redundant, systems for any system that is essential for safe flight. The problem with fuel storage and supply systems in an aircraft is that they simply don't have a redundant backup. If fuel is contaminated, the contaminant will be supplied to all an aircraft's engines at the same time and could make them all unreliable at the same time.

As the primary defence of a redundant system isn't available to protect against the safety critical problem of fuel quality, we could reasonably expect there to have been a number of fuel quality related accidents in the recent past; however, that was not so. This can only be attributed to a highly reliable system for manufacture and distribution of aviation fuels, with a well-managed quality control processes.

Despite this, it is clear that complacency on the part of any group that has a responsibility towards maintaining fuel quality, be they refiner, distributor, regulator or consumer, can have catastrophic consequences.

This Avgas contamination event must be seen as a clarion call to highlight an aspect of the system of safe aviation that is more vulnerable to abuse or neglect than most other safety critical aviation systems.

Avgas contamination investigation report released

The Australian Transport Safety Bureau (ATSB) released its report on the contaminated aviation gasoline (Avgas) investigation at a media conference on 30 March 2001. The investigation followed the grounding in January 2000 of thousands of piston engine aircraft across eastern Australia when a black gunk was found in fuel systems.

The investigation found that a very small amount of an anti-corrosion chemical that was not removed in Mobil's Avgas refining process in late 1999, and not detected by the usual tests, led to the safety problem.

The ATSB made 24 separate recommendations as a result of its investigation that included recommended safety actions for Mobil Oil Australia, US and UK fuel standards bodies, the Civil Aviation Safety Authority, and other Australian regulatory organisations.

ATSB Executive Director Kym Bills told the media that the scale of the Avgas contamination was an unprecedented event anywhere in the world and was unexpected in such a mature industry as fuel refining. As a result, it caught the refiner and regulators by surprise and also revealed deficiencies in international fuel standards.

The investigation found that a temporary variation in the production process at Mobil's Altona refinery in late 1999 involving problems with reduced caustic wash and increased acid carry over, led to an increased dosage of an alkaline anti-corrosion chemical by a contractor. This was not totally removed from the final Avgas. The normal tests for the quality of Avgas did not pick up the very small concentration of the chemical contaminant in the Avgas that was sufficient to react with brass in aircraft fuel systems and form a black gunk that clogged them.

Mr Bills said it was not the ATSB's role to ascribe blame to any party. The task was to uncover the facts including all of the significant contributory factors (including weaknesses in defences), and then to publish findings and recommendations in a report.

Accordingly, it was important that relevant parties learnt from the identified safety deficiencies and acted promptly on the 24 recommendations made to reduce the chances of a recurrence, either with Avgas or jet fuel.

On 3 July 2009, the Australian Transport Safety Bureau (ATSB) was notified that a SAAB Aircraft Company 340B (SAAB), registered VH-ZLW, had commenced its take-off roll along the runway 25 left edge lights at Sydney Kingsford Smith Airport, New South Wales. This was one of three occurrences over the previous 2 years that involved aircraft commencing take-off on the runway edge lighting.

In addition, within the previous 2 years the ATSB investigated two other occurrences involving pilot misidentification of runway alignment cues or lack of those cues during take-off. All five Australian misaligned take-off and landing occurrences involved aircraft with weights greater than 5,700kg and three of the six occurrences involved scheduled regular passenger transport (RPT) operations. The remaining two occurrences involved charter operations.

This research investigation examined each of these occurrences and relevant international occurrences to identify the common factors associated with misaligned take-off and landing occurrences.

After reviewing the Australian and international occurrences, eight common factors were identified that increased the risk of a misaligned take-off or landing occurrence. The factors included: distraction or divided attention of the flight crew; confusing runway layout; displaced threshold or intersection departure; poor visibility or weather; air traffic control clearance/s issued during runway entry; no runway centreline lighting; flight crew fatigue; and recessed runway edge lighting.