On 4 May 2024, a Boeing 737, registered VH‑VYK, being operated by Qantas Airways Limited as QF520, departed Sydney, New South Wales on a scheduled flight to Brisbane, Queensland. The aircraft departed Sydney at 1202 local time and was scheduled to arrive in Brisbane at 1335. On board was the flight crew, comprised of the captain and first officer (FO), a customer service manager (CSM) supported by 3 cabin crew members, and 143 passengers. 

Prior to departing Sydney for Brisbane, the captain recalled briefing the CSM about leaving the seatbelt sign on for the departure from Sydney due to weather, however neither the captain nor the CSM could recall any specific details regarding the weather conditions in Brisbane.

Descent into Brisbane

At about 1300, the flight crew commenced their descent into Brisbane while approaching Lismore, New South Wales. As the aircraft continued descent towards Coolangatta, Queensland, the captain visually observed cloud over Moreton and Stradbroke islands, but recalled no weather radar indications that identified precipitation normally associated with increased turbulence. 

After passing 30,000 ft above mean sea level (AMSL) at 1303, the captain, who was pilot monitoring, recalled performing the ‘prepare cabin’ public announcement (PA), and the cabin crew commenced securing the cabin for landing. About 6 minutes later, the captain contacted the CSM to enquire about the amount of time the cabin crew required to complete their cabin preparations. At about 15,000 ft AMSL, the aircraft entered a thick layer of stratiform[1] cloud with minimal turbulence observed initially by either the flight or the cabin crew. The CSM advised the captain the preparations would take about 2 minutes. About 2 minutes later, the captain turned on the seatbelt sign and announced the ‘seatbelts PA’.

Immediately after the captain turned on the seatbelt sign, the CSM and the 2 cabin crew members in the aft galley completed their duties as per the procedures. The CSM made a PA while standing in the forward galley, then checked and locked their assigned lavatory. The 2 cabin crew members in the rear galley got out of their jump seats to verify that the lavatories were vacant before locking them.

At about this time, the captain observed a cumulus[2] cloud embedded in the stratiform layers. However, there was no radar return consistent with increased turbulence, and the cloud did not appear to be overly concerning in terms of turbulence risk from their shape or size.  

After descending below 12,000 ft AMSL at 1311, the aircraft encountered a severe turbulence event less than one minute after the seatbelt sign was illuminated, while 3 cabin crew were unrestrained. The CSM recalled observing the right 2 primary (R2P) cabin crew member at the rear galley rising off the floor and colliding with the aircraft's ceiling. They immediately fell back to the floor, landing on their right ankle. The R2P felt a crushing sensation as they landed on their ankle and experienced intense pain and was unable to move.

Events in the cabin

The CSM contacted the captain to inform them that the R2P was injured and that some passengers were standing. The captain responded with instructions that all passengers and cabin crew must remain seated. Following this interaction, the CSM made a public announcement to remind passengers to stay seated and to request assistance from the cabin crew if needed. The captain contacted the operator’s Brisbane airport coordinator to advise that there was an injured cabin crew member and medical assistance would be required on arrival in Brisbane. Queensland Ambulance service records showed that a request for an ambulance was received at 1315.

Although being instructed to remain seated, the CSM immediately moved to the rear of the aircraft to assist the injured R2P in the aft galley. There, they observed the injured R2P lying on the floor while a passenger was holding their leg. Another passenger, who identified themselves as a doctor, offered to assist with providing first aid. Meanwhile, another cabin crew member, the left two primary (L2P), was supporting the R2P’s head. 

At this point, the CSM advised the L2P and the passengers that the captain had instructed everyone to return to their seats. However, they were unwilling to leave the R2P unattended. The CSM instructed the L2P to advise the captain of the situation. The CSM then retrieved the physician's kit from the front of the cabin. At this point, the passenger seated in 3F advised the CSM that they were a travelling cabin crew member and were able to assist. 

The L2P contacted the captain to advise that all the occupants located in the rear galley were still unrestrained. However, the captain did not recall receiving requests for additional time to address the situation. The captain reiterated that all uninjured occupants must return to their seats as the aircraft was in the final stages of the approach and would be landing soon.

The CSM subsequently returned to the rear galley with the physician’s kit and the off‑duty cabin crew from 3F, who subsequently relieved the passenger who was holding the R2P’s leg. While the CSM was attempting to provide first aid and preparing a splint with the assistance of the travelling doctor, a passenger seated in 30D yelled, ‘we’re about to land’. Shortly after at 1322, the aircraft landed in Brisbane with 4 unrestrained passengers and cabin crew in the rear galley. The flight crew taxied the aircraft to its assigned gate, arriving at 1328, with paramedics in attendance at 1338.

Context

Flight crew information

Captain

The captain held an Airline Transport Pilot (aeroplane) Licence with an instrument rating and a Class 1 aviation medical certificate. They had 23,177 flight hours, including 15,005 hours on the Boeing 737, and had logged 165 hours on the 737 in the last 90 days. 

The captain reported sleeping 7 hours the night before the occurrence. They were awake for 8 hours and 45 minutes at the time of the occurrence and reported feeling ‘responsive, but not at peak’.

First officer 

The first officer held an Airline Transport Pilot (aeroplane) Licence with an instrument rating and a Class 1 aviation medical certificate. They had 10,163 flight hours, including 1,717 hours on the Boeing 737, and had logged 152 hours on the 737 in the last 90 days. 

The first officer reported sleeping 7 hours the night before the occurrence. They were awake for 8 hours and 45 minutes at the time of the occurrence and reported feeling ‘somewhat fresh’.

Cabin crew

The cabin crew on board was comprised of a complement of 4 members, with their assigned jump seats located in the forward and aft galleys. Each cabin crew member was responsible for one of the 4 main cabin doors (Table 1) during critical phases of flight, with their assigned jump seat (Figure 1) located immediately next to their assigned door. The cabin crew were under the supervision of the customer service manager (CSM) who was responsible to the captain for administration of in‑cabin service and liaison with the crew for all service and safety related matters. 

Table 1: Cabin crew door assignment

Figure 1: 737 Cabin layout, doors and assigned jump seats for cabin crew

Source: Qantas, annotated by the ATSB

Cabin crew injuries

R2P

The R2P had been in the process of taking their seat when the turbulence occurred. They rose into the air during the turbulence, struck their head on the ceiling, and landed heavily on their feet. R2P immediately fell to the galley floor and told L2P that they were injured, possibly with a broken bone. They were later diagnosed in the hospital with a fracture involving 2 breaks in the ankle and another break in the leg, which required surgery.

CSM

The CSM sustained minor injuries due to striking aircraft fixtures while standing unsecured during the turbulence. The CSM self‑assessed their injuries and applied first aid the following day after noticing minor pain, including discomfort in their lower back and right shoulder blade. They also became aware of facial pain 2 days after the event.

L2P

The L2P sustained a head injury, due to striking the ceiling or other aircraft fixtures during the event but did not initially believe they were injured. They had several rostered days off after the event. They returned to work on 11 and 12 May and were made aware by co‑workers that they were displaying symptoms of possible injury. On 16 May, 12 days after the turbulence event, they were diagnosed with concussion after a consultation with their general practitioner.

Aircraft

The aircraft was registered as VH‑VYK in Australia on 11 January 2006, serial number 34183. The Boeing 737‑800 is a twin‑engine, narrow‑body commercial aircraft in the 737 Next Generation series, used for short to medium‑haul routes. It had a seating capacity of 174 passengers, with Qantas configuring its aircraft with 12 business class seats in a 2‑2 layout and 162 economy class seats. It is powered by two CFM56‑7B turbofan engines. 

The passenger address system broadcasts announcements throughout the cabin, and the interphone facilitates communication between the flight and cabin crew. 

Weather radar

The weather radar system fitted to the Boeing 737 detects and locates various types of precipitation bearing clouds along the flight path of the aircraft and gives the pilot a visual indication in colour of the cloud’s intensity. The radar antenna sweeps a forward arc of 180°. The radar indicates a cloud’s rainfall intensity by displaying colours contrasted against a black background. Areas of heaviest rainfall appear in red, the next level of rainfall in amber, and the least rainfall in green (Figure 2). 

Figure 2: Example depiction of 737 weather radar returns on pilot's navigational display

Source: Qantas, annotated by the ATSB

The turbulence mode displays normal precipitation and precipitation associated with turbulence. When the radar detects a horizontal flow of precipitation with velocities of 5 or more metres per second toward or away from the radar antenna, that target display becomes magenta. These magenta areas are likely associated with heavy turbulence.

The captain did not recall identifying areas of turbulence on the weather radar on descent or report that they were experiencing any difficulties operating the weather radar. Neither the captain nor the operator reported that the weather radar fitted to the aircraft was unserviceable during this occurrence.

Post‑event maintenance

Aircraft data was collected for use by the operator’s maintenance operations control, and engineering. The aircraft health monitoring (AHM) section collected and analysed data from aircraft components and systems post‑event to assess their condition and identify any potential overstressing of components. The operator’s AHM and flight data analysis showed that the event did not exceed tolerances, and no additional inspections were required.

Meteorological information

Brisbane Airport weather 

The TAF for Brisbane Airport predicated winds from 100° at 10 kt with visibility more than 10 km in light rain showers and scattered cloud at 3,500 ft from 1200‍–‍2000 local time. Table 2 shows the automatic terminal information service (ATIS)[3] report at the time of departure. 

Table 2: Brisbane Airport automatic terminal information service (abridged)

Gold Coast Airport weather

The TAF for Gold Coast Airport predicted winds from 100° at 10 kt with visibility more than 10 km in light rain showers and scattered cloud at 2,000 ft. There was a significant intermittent variation from the prevailing conditions from 1300‍–‍2200 local time. For up to 30 minutes at a time during this period, the visibility was forecast to reduce to 4 km in rain showers with broken[5] cloud at 1,500 ft.

Graphical area forecast

The flight transited through a region contained in the graphical area forecast for Queensland south, covering subdivisions A and A2 (Figure 3). Table 3 shows the forecast conditions for the duration of the descent into Brisbane Airport which was issued at 0815 local time.

Figure 3: Graphical area forecast Queensland south

Source: Australian Bureau of Meteorology, annotated by the ATSB

Table 3: Graphical area forecast Queensland south

Turbulence reporting 

In accordance with the requirements of regulation 91.675 of CASR (Civil Aviation Safety Regulations 1988) and instructions contained in the Aeronautical Information Publication Australia, a special air report must be provided to air traffic control whenever turbulence meeting the following specifications is encountered:

• Moderate: Changes to accelerometer readings of between 0.5 g and 1.0 g at the aircraft’s centre of gravity. Moderate changes to aircraft attitude and/or altitude may occur but aircraft remains under positive control. Usually small changes in airspeed. Difficulty in walking. Loose objects moved about.
• Severe: Changes to accelerometer readings greater than 1.0 g at the aircraft’s centre of gravity. Abrupt changes to aircraft attitude and/or altitude may occur; aircraft may be out of control for short periods. Usually large changes of airspeed. Loose objects tossed about.

The flight crew could not recall whether a special air report was provided to air traffic control and the turbulence event was classified as moderate in a post‑flight conference call between the captain and Qantas personnel. The operator defined moderate turbulence as ‘causing rapid bumps or jolts without appreciable changes in aircraft altitude or attitude.’ The description elaborated that unsecured objects are dislodged, and walking is difficult. 

Recorded data

The operator’s internal investigation report detailed flight data and showed that over a 4‑second period, at an altitude of approximately 11,100 ft, the aircraft recorded:

• vertical G went from +1.2G to -0.06G (negative) to +1.35G to +0.61G to +1.59G.
• pitch attitude changed from -2.4deg to -0.7deg over 1 sec.

Operator procedures

Dispatch weather briefing

For the operator’s domestic sectors, a flight plan, textual weather, and notice to airmen (NOTAM)[7] were provided in a briefing package to flight crew via an electronic application installed on their iPads. The briefing information included graphical area forecasts, additional weather, turbulence information and weather radar overlay imagery. Figure 4 shows a weather radar overlay issued at the time of departure. For operator domestic flights under 90 minutes, the crew did not receive a flight watch service[8] from the operator’s flight dispatch during the flight and consequently, any hazard alerts or amended terminal forecasts (TAF)[9] were only provided by air traffic control.

Prior to departure from Sydney, the flight crew conducted their pre‑flight planning and reviewed the weather briefing package. The crew recalled that the weather briefing package contained forecast light showers and some cloud in the vicinity of Brisbane and that the forecast weather for departure at Sydney was worse. 

Figure 4: Brisbane Airport weather radar overlay on 4 May at 1200 local time (time of departure)

Source: Qantas, annotated by the ATSB

Cabin preparation for landing 

The Qantas Flight administration manual (FAM) specifies that the flight crew make the ‘prepare cabin’ public announcement (PA) for cabin crew to commence cabin preparations at 20,000 ft or no lower than 10,000 ft above the destination airport. This PA can be performed at a higher altitude when considering descent profile, arrival procedures, weather and workload management. The customer service manager will then confirm receipt of the PA by directly contacting the flight crew. 

The Qantas Cabin crew operation manual (CCOM) states the timing of this announcement should provide the cabin crew at least 10 minutes to secure the cabin and occupants for landing prior to the illumination of the seatbelt sign. After receipt of the ‘prepare cabin’ PA, all activity by the cabin crew shall be safety‑related only and no new service duties may be initiated.

Additionally, the FAM stated, ‘should contingencies occur that impact on the planned preparation time, every effort should be made to advise the cabin crew of these changes.’

In the case of the occurrence, the CSM recalled that the captain advised them that they would prepare the cabin earlier for a ‘bit of weather’. The prepare cabin PA was performed after passing 30,000 ft and the seatbelt sign was illuminated less than 10 minutes later following a further discussion with the CSM relating to the progress of the cabin preparations. 

Anticipated turbulence on descent 

In the event that turbulence is anticipated on descent, the Qantas FAM stated that:

Where turbulence is anticipated during descent, the flight crew should consider the 10 minutes requirement to prepare the cabin for landing. Cabin crew are to be alerted to anticipated turbulence as early as possible to enable them to complete their duties. 

Any time the seatbelt signs are illuminated for turbulence, a PA must be made by the crew. 

Securing cabin for landing

Cabin secured for landing

The CCOM specified that when the seatbelt sign is illuminated at the conclusion of the 10 minutes, the CSM will make the following PA:

All customers and crew must now be seated for landing with their seatbelt fastened

The cabin crew will subsequently perform the following procedures:

• ensure passengers are seated

• verify lavatories are vacant

• conduct a simultaneous galley secure check

• return to jump seats and secure within one minute

• after one minute, the CSM initiates a callback from their jump seat to the other cabin crew, to advise that the cabin was secure

• perform silent [safety] review.

Cabin unsecured for landing

Approximately one minute after the seatbelt sign is illuminated, the CSM will initiate a callback from their designated jump seat. Each cabin crew member will respond with ‘door number, name, cabin secured for landing’. If the cabin is not secured for landing, the CSM will inform the flight crew. In this case, the CCOM stated the ‘the CSM will assess the situation and inform the flight crew regarding the status of the cabin.’ This allows alternative actions before the no contact period, which commences when the landing gear is extended for landing. The cabin crew are not to contact the flight deck under any circumstances during this period. 

In this case, the captain stated they would have ensured that passengers and cabin crew assisting the injured member were seated before landing had they been informed the cabin was not secure.

Use of cabin secure notifications in Australian airlines

The ATSB also reviewed the cabin secure procedures of 7 similar Part 121 passenger air transport operators in Australia. Of those, 4 operators employed a positive signal to confirm cabin security during normal operations, while 3 did not. 

In the past 20 years, the ATSB identified 26 occurrences involving unrestrained occupants on landing, based on historical occurrence data. None of these resulted in fatalities, serious injuries, or minor injuries. Of these 26 occurrences, only 4 could potentially be linked to the absence of a positive cabin secure signal. 

Anticipated turbulence procedure

If the flight crew anticipates turbulence, the following procedure from the CCOM will apply:

When the flight crew become aware of anticipated turbulence, they will liaise with the CSM advising the time and likely duration of the anticipated turbulence. 

The CSM will relay this information to the members of the cabin crew to enable them to prioritise their duties by securing carts, galleys, items of service equipment, cabin and galley curtains, and the passenger cabin, based on time available. 

A PA may be made by the flight crew to the passengers and cabin crew advising that cabin service is to cease as there is a likelihood of turbulence and the following actions will be initiated (Table 4):

Table 4: Crew actions during anticipated turbulence

Unanticipated turbulence

Unanticipated light turbulence

If it is deemed necessary to illuminate the seatbelt signs for unanticipated light turbulence, the flight crew will perform the ‘seatbelts’ PA announcing, ‘all passengers and crew must be seated and fasten seatbelts.’ 

The CCOM states the following will then apply: 

Cabin crew are to prioritise their duties by securing carts, galleys, items of service equipment and the passenger cabin and initiate crew actions for anticipated turbulence…..

When the captain illuminated the seatbelt sign prior to the turbulence encounter, they performed the ‘seatbelts’ PA. They stated they did this as a precautionary measure as this was their normal practice if there is the possibility of some turbulence. 

Unanticipated turbulence posing an immediate hazard

In the event of unanticipated turbulence that poses an immediate safety hazard, the flight crew must select the seatbelt signs on and announce the ‘turbulence’ PA: 

All passengers and crew be seated and fasten seatbelts immediately. 

Following this announcement, the CCOM issues the following instructions: 

Cabin crew will lock carts in position and secure themselves in the nearest seat or wedge themselves in the aisle. 

If circumstances permit, the CSM will initiate the call back procedure, ascertain the condition of the cabin and relay this information to the flight crew. 

Cabin crew incapacitation

In the event of a cabin crew member’s incapacitation, the CCOM calls for first aid to be administered, and the CSM and captain to be notified as soon as practicable. The CSM would then reassign the duties of the incapacitated crew member to an assist crew member, if available. If the incapacitation results in a crew complement less than the minimum required, the CSM, in consultation with the captain, will determine whether any additional off‑duty crew members are available to assume the role of the incapacitated cabin crew for landing.

The timing of the incapacitation placed concurrent procedural demands on the CSM and disrupted the cabin secure procedures. The CCOM procedures were sequential instructions designed to achieve consistent performance and reduce the potential for miscommunication and non‑conformances. 

Injury response tool

The Qantas Group injury response tool specified the actions the crew must follow if an injury has occurred inflight. In this case, the injury response tool directed crew to seek immediate doctor advice via telephone for a turbulence event and for a head blow or head strike which fell under the classification as a ‘specific circumstance’. Consequently, the crew would be required to cease work immediately. 

However, as neither co‑workers nor the injured crew recognised the symptoms of a concussion in the L2P or a facial injury in the CSM, they were not triaged according to the Group injury response tool. The operator stated that this likely occurred due to the injury response tool process relying on the self‑assessment of injuries which may not be immediately apparent.

Integrated operations controls procedures

The integrated operations control (IOC) was the central point of contact during any kind of disruption or incident on a Qantas aircraft. At the time of this occurrence, the IOC incident notification communications protocol contained in the operations control procedure (Figure 5) included contact with the on‑call doctor in cases of severe turbulence regardless of whether there were reported injuries. The IOC notification process for a ‘significant passenger/crew injury or illness’ event did not include contact with the Qantas on‑call doctor. 

As the turbulence in this event was deemed post‑event by flight crew as moderate, contact with the on‑call doctor was not required under the process.

Figure 5: Qantas incident notification communications protocol version 11, March 2024

Source: Qantas

Research into turbulence detection

There are occasions where it can be challenging to identify turbulence. The US National Transportation Safety Board conducted 10 case studies (National Transportation Safety Board, 2021) of accidents between 2019‍–‍2020 that involved turbulence and embedded convection and determined that: 

Embedded convection may not be easily detected by onboard or ground‑based weather radar, and when not visible outside the aircraft windows, this class of convective activity can act as a hidden source of severe turbulence encounters within an otherwise benign‑looking cloud mass. 

Research into turbulence related injuries 

From 2009 through 2018, the US National Transportation Safety Board (NTSB)[10] found that turbulence‑related accidents accounted for more than a third of all Part 121 accidents. The accident data revealed that the most common phase of flight associated with turbulence‑related accidents in Part 121 operations was during the en route descent, which accounted for 36.0% of accidents.

Further analysis indicated that cabin crew accounted for 78.9% of serious injuries, with the majority occurring in the aft section of the aircraft cabin. Passengers accounted for 21.1% of serious injuries, while no flight crew members were seriously injured (National Transportation Safety Board, 2021).

The distribution of cabin crew injuries found most occurring in or near an aft galley (Figure 6), which likely reflects that the service‑related duties of cabin crew often require them to spend more time working unrestrained in the galley area. The most commonly reported cabin crew activities at the time of serious injury were:

• preparing the cabin for landing (39.2%)
• conducting cabin service (13.4%)
• preparing for cabin service (9.3%).

During this occurrence, the R2P and L2P sustained the most severe injuries while unrestrained in the aft galley as they prepared the cabin for landing, which is consistent with research on turbulence‑related injuries.

Figure 6: Location of cabin crew at time of turbulence‑related serious injury, 2009–2018

Source: National Transportation Safety Board 

Safety analysis

Crew communication 

During the descent into Brisbane, the captain commenced cabin preparations earlier than usual, using standard protocols to account for known weather conditions en route. Although the graphical area forecast indicated the possibility of moderate to severe turbulence, the captain did not observe weather radar returns or receive any other pilot reports indicating the presence of moderate to severe turbulence during the descent.

Approximately 5‍–‍6 minutes after initiating cabin preparations, the aircraft entered stratiform cloud and the captain contacted the customer service manager (CSM) to check on the cabin crew’s progress. The purpose of this communication was to provide the captain with information to guide the timing of the seatbelt sign illumination. However, the captain did not provide any weather‑related information to the CSM during this interaction, leaving the cabin crew unaware of any increased likelihood of turbulence. 

Two minutes later, the seatbelt sign was illuminated, accompanied by the ‘seatbelt’ public announcement (PA). The captain then observed an approaching cumulus cloud along the flight path but determined it did not pose an immediate hazard based on a visual assessment and the lack of radar indications. As a result, the captain did not perform the ‘turbulence’ PA, which would have prompted the cabin crew to immediately secure themselves in the nearest seat or wedge themselves in the aisle to prepare for the turbulence encounter.

Although the captain contacted the CSM to confirm the time remaining to prepare the cabin, the absence of indications to the subsequent severity of the turbulence limited the captain's perception of the possible threat. Therefore, additional precautions were not considered. The captain followed normal descent procedures, however, did not discuss any additional weather‑related information in communications with the CSM. 

As a result, the cabin crew, who relied on information from the flight crew, were unprepared for the turbulence encounter. This situation underscores the difficulties posed by unexpected turbulence, as the procedures for managing in‑flight turbulence rely on the flight crew's ability to predict or avoid these situations.

Crew unrestrained during severe turbulence

When the seatbelt sign was illuminated during the descent, cabin crew members were required to perform several duties whilst being unrestrained. In the moments immediately preceding the turbulence encounter, the CSM and R2P recalled checking their assigned lavatories as part of securing the cabin for landing. 

Cabin crew were required to complete their assigned duties within one minute of the seatbelt sign being illuminated, which was also the case for unanticipated light turbulence. The captain performed the ‘seatbelts’ PA when the seatbelt sign was illuminated, which indicated unanticipated light turbulence to the cabin crew members. However, the cabin crew did not recall hearing this PA and remained unaware of the increased risk of turbulence as the aircraft approached a cumulus cloud. 

Because the turbulence event occurred less than one minute after the illumination of the seatbelt sign, which was accompanied by the ‘seatbelt’ PA, the cabin crew did not have sufficient time to ensure they were seated and restrained prior to the aircraft being affected by turbulence. The injuries sustained during the encounter reflect research showing that cabin crew members face a higher risk of turbulence‑related injuries, especially during the descent phase of a flight when they are preparing for landing (National Transportation Safety Board, 2021).

Cabin management 

After the turbulence event, the CSM and left 2 primary (L2P) turned their attention to the right 2 primary (R2P) who was laying on the floor of the aft galley and was unable to move due to their injury. The turbulence event occurred in the latter stages of the descent, which meant there was little time to provide first aid to the R2P and complete the required preparations for landing. The CSM, L2P and the passengers assisting the R2P were reluctant to return to their assigned seats despite the clear instructions from the captain to do so. 

The situation in the aft galley disrupted the procedural flow and meant that the CSM and L2P became focused on providing first aid rather than returning to their seats to complete the callback and silent review prior to landing. Interruptions often lead people to forget to resume their tasks, while multitasking can further complicate the situation by increasing the overall workload within a limited timeframe (Loukopoulos & Barshi, 2009). In this case, the CSM and L2P had to balance providing first aid and securing the cabin. High stress levels are also known to cause errors (Kim & Hyun, 2022), which likely contributed to the CSM prioritising providing first aid over securing themselves and the cabin for landing.

The CSM subsequently lost situational awareness with respect to the phase of flight and the sequence of the standard operating procedures. As a result of being situated in the aft galley, the CSM likely missed audible cues, such as the extension of the landing gear. 

The captain did not recall receiving any requests for more time to prepare the cabin for landing. Additionally, the single aisle cabin configuration of the Boeing 737 offered limited options for accommodating the R2P anywhere other than the aft galley. After repeated instructions for everyone to be seated for landing, the captain was confident that all uninjured occupants had complied.

The decision of the CSM and L2P to remain unrestrained in the aft galley during a critical phase of flight increased the risk of incapacitation to additional cabin crew, which could have further compromised their ability to manage a landing‑based emergency effectively if one was to happen. Additionally, the 5 occupants in the aft galley created a potential obstruction to emergency exits, increasing the likelihood of delays or complications if they needed to enact an emergency evacuation. 

The aircraft landed with the CSM, L2P, R2P and 2 passengers unrestrained in the aft galley. The flight crew was made aware by the CSM that the injured cabin crew member was unsecured and unable to be made secured for landing and instructed the CSM to ensure everyone else was secure for landing. While this instruction was communicated to those people unsecure in the cabin, the instruction was not followed as described above. The CSM attempted to inform the flight crew by instructing the L2P to communicate with them. However, the captain again instructed that everyone who could be secured needed to be, as they were landing. As such, the flight crew assumed all cabin occupants would be secure apart from the injured R2P crew. At this stage, the cabin crew operating procedures requiring the CSM to inform the flight crew if the cabin was not secure broke down as there was no further communication that the cabin was not secure.

The captain stated that if they had known that 4 uninjured occupants were still unrestrained in the aft galley, they would have taken appropriate action to ensure they had returned to their seats prior to the final approach to land. The lack of a positive signal increased the likelihood that flight crew would be unaware of unrestrained occupants during the approach and landing phases of flight.

While the lack of a positive cabin secure signal played a role in this occurrence, the available data does not indicate it as a significant ongoing risk.

Post‑flight medical assessment 

Shortly after arrival at the gate at Brisbane airport, the R2P was attended to by ambulance personnel. However, the CSM and L2P, who were also injured during the event, did not receive any follow‑up medical assessments or treatment. This situation arose due to procedural gaps, which relied on crew members to self‑assess and report a significant injury to receive a medical assessment. 

While the CSM self‑diagnosed a minor injury and reported it the following day, the L2P was unaware of their injury. As a result, the L2P operated on multiple flights while experiencing symptoms of an undiagnosed concussion, until some days later when co‑workers noticed signs of a possible injury.

The Qantas integrated operations control protocols did not mandate contacting the on‑call doctor in cases where a passenger or crew member was significantly injured. Although the protocol required consultation with the on‑call doctor in cases of severe turbulence, this turbulence event was classified as moderate, and no medical consultation was either required or requested. Additionally, the Qantas group injury response tool also relied on crew members self‑assessing their injuries to determine if medical treatment would be required, but an injured crew member may not realise the extent of their injury at the time. 

In the cases of a concussion, symptoms may include impairments in neurocognitive functioning, primarily affecting attention, concentration, memory, and judgment or problem‑solving (Ryan & Warden, 2003). Returning to work with an undiagnosed concussion likely compromised the L2P’s ability to perform safety‑critical tasks. A subtle incapacity due to an undiagnosed injury could negatively impact operational safety, particularly during emergencies.

Findings

From the evidence available, the following findings are made with respect to the turbulence event and cabin crew injury involving Boeing 737, VH‑VYK, 36 km south‑east of Brisbane Airport, Queensland, on 4 May 2024. 

Contributing factors

• The captain did not communicate to the cabin crew about the expected turbulence, likely as a result of the captain not knowing the severity of the turbulence.
• Three cabin crew were unrestrained while performing duties during unanticipated severe turbulence resulting in all 3 receiving injuries.

Other factors that increased risk

• Although the captain had instructed that the uninjured passengers and crew needed to be seated, 3 cabin crew and one passenger were unrestrained for landing due to being preoccupied with administering first aid to the injured cabin crew member. This increased the risk of injury to the unrestrained occupants and had the potential to compromise a safe emergency evacuation if required.
• The Qantas 737 procedures did not require flight crew to receive positive confirmation that the cabin was secure for landing. This increased the risk that occupants and objects were not secure for landing.
• A crew member with undiagnosed concussion from the accident flight operated on subsequent flights without receiving appropriate medical attention.
• Qantas lacked a procedure to ensure cabin crew fitness was assessed after a significant injury. This increased the risk that a crew member could continue to operate while being unfit for duty. (Safety issue)

Safety issues and actions

Undiagnosed injuries 

Safety issue number: AO-2024-032-SI-01

Safety issue description: Qantas lacked a procedure to assess cabin crew fitness after a serious injury. This increased the risk that a crew member could continue to operate while being unfit for duty.

Glossary

Sources and submissions

Sources of information

The sources of information during the investigation included:

• the captain of the accident flight
• the customer service manager on the accident flight
• the R2 primary cabin crew member on the accident flight
• Qantas Airways Limited
• the manager of Safety, Qantas Airways Limited
• Civil Aviation Safety Authority
• Bureau of Meteorology 

References

Endsley, M. R. (1999). Situation awareness in aviation systems. Handbook of aviation human factors. Retreived from https://www.pacdeff.com/pdfs/AviationSA-Endsley%201999.pdf.

Kim, J. Y., & Hyun, S. (2022). Study on Factors That Influence Human Errors: Focused on Cabin Crew. International Journal of Environmental Research and Public Health, 19(9), 5696. 

Loukopoulos, L. D., & Barshi, I. (2009). The multitasking myth : Handling complexity in real-world operations. Taylor & Francis Group. Taylor & Francis Group.

National Transportation Safety Board. (2021). Preventing Turbulence-Related Injuries in Air Carrier Operations Conducted Under Title 14 Code of Federal Regulations Part 121. Retrieved from https://www.ntsb.gov/safety/safety-studies/Documents/SS2101.pdf. 

Ryan, L. M., & Warden, D. L. (2003). Post concussion syndrome. International Review of Psychiatry, 15(4), 310–316. Retreived from https://doi.org/10.1080/09540260310001606692.

Submissions

Under section 26 of the Transport Safety Investigation Act 2003, the ATSB may provide a draft report, on a confidential basis, to any person whom the ATSB considers appropriate. That section allows a person receiving a draft report to make submissions to the ATSB about the draft report. 

A draft of this report was provided to the following directly involved parties:

• the captain of the accident flight
• the customer service manager on the accident flight
• the L2 primary cabin crew member on the accident flight
• the R2 primary cabin crew member on the accident flight
• Qantas Airways Limited
• Civil Aviation Safety Authority
• Bureau of Meteorology.

Submissions were received from:

• the L2 primary cabin crew member on the accident flight
• the R2 primary cabin crew member on the accident flight
• Qantas Airways Limited
• Civil Aviation Safety Authority
• Bureau of Meteorology

The submissions were reviewed and, where considered appropriate, the text of the report was amended accordingly.

Purpose of safety investigations The objective of a safety investigation is to enhance transport safety. This is done through:  identifying safety issues and facilitating safety action to address those issues providing information about occurrences and their associated safety factors to facilitate learning within the transport industry. It is not a function of the ATSB to apportion blame or provide a means for determining liability. At the same time, an investigation report must include factual material of sufficient weight to support the analysis and findings. At all times the ATSB endeavours to balance the use of material that could imply adverse comment with the need to properly explain what happened, and why, in a fair and unbiased manner. The ATSB does not investigate for the purpose of taking administrative, regulatory or criminal action. Terminology An explanation of terminology used in ATSB investigation reports is available here. This includes terms such as occurrence, contributing factor, other factor that increased risk, and safety issue. Publishing information Released in accordance with section 25 of the Transport Safety Investigation Act 2003 Published by: Australian Transport Safety Bureau © Commonwealth of Australia 2025   Ownership of intellectual property rights in this publication Unless otherwise noted, copyright (and any other intellectual property rights, if any) in this report publication is owned by the Commonwealth of Australia. Creative Commons licence With the exception of the Commonwealth Coat of Arms, ATSB logo, and photos and graphics in which a third party holds copyright, this report is licensed under a Creative Commons Attribution 4.0 International licence. The CC BY 4.0 licence enables you to distribute, remix, adapt, and build upon our material in any medium or format, so long as attribution is given to the Australian Transport Safety Bureau.  Copyright in material obtained from other agencies, private individuals or organisations, belongs to those agencies, individuals or organisations. Where you wish to use their material, you will need to contact them directly.

Executive summary

What happened

On the afternoon of 23 May 2024, loaded WATCO grain train 6839 was being operated between Thallon and Brisbane (Fisherman Island), Queensland. At 1325, the train was travelling through Gooray and approaching a passive level crossing with Gooray Road from the west at about 60 km/h, when the train crew sounded the horn about 440 m in advance of the crossing. 

At that time, a heavy vehicle consisting of a prime mover and low‑loader trailer carrying a skid steer was travelling southbound on Gooray Road at about 40‍–‍60 km/h towards the level crossing. This vehicle had departed a farm in Gooray several minutes earlier. Prior to departure, the driver had conducted several tests to check whether the trailer was connected to the prime mover and whether the brakes were functioning on the consist. During these tests, the truck driver observed a potential problem with the brakes, but did not investigate this issue before departing.

The train and truck should have been visible to each other for about 1 km as both approached the level crossing, but they did not see each other before the sightlines between the road corridor and the rail corridor were impeded by vegetation. The train crew first saw the truck about 300 m before the crossing, and at that moment, they assessed it was not going to stop and applied the train’s emergency brakes and braced for impact. 

The truck driver did not see the train until about 50 m before the crossing, having cleared the vegetation. When the truck driver sighted the train, they attempted to brake but assessed they could not stop safely before the crossing. Instead, the truck driver attempted to veer slightly and accelerated through the crossing in front of the train. The prime mover was able to clear the railway line, but the trailer was still across the tracks when the train reached the crossing. 

The lead locomotive impacted the trailer, the second locomotive impacted the lead locomotive and 14 grain hoppers derailed. The prime mover, low-loader trailer, both locomotives and 12 grain hoppers were damaged beyond repair, while 2 other grain hoppers and the skid steer were substantially damaged. The train crew and truck driver sustained serious injuries.

What the ATSB found

The ATSB found that the truck driver did not stop prior to the level crossing as required by the road signage. Due to the infrequency of trains on that corridor, it is likely that the truck driver had not seen a train at the Gooray Road crossing on previous journeys. This created an expectation that there would be no conflicting rail traffic at the crossing, which probably reduced the effectiveness of the truck driver’s scan and led to them approaching the intersection at a speed too high to stop prior to the level crossing.

The ATSB also found that the configuration of the Gooray Road level crossing provided appropriate signage and stopping distance for a road vehicle driver to notice the passive level crossing controls and bring the vehicle to a controlled stop before the level crossing. Once stopped, there was adequate visibility for a driver to sight a train and give way.

What has been done as a result

Following the incident, Queensland Rail initiated an Australian Level Crossing Assessment Model (ALCAM) assessment and looked at road geometry, traffic volumes (road and rail), visibility and sighting distances, and the existing passive controls. The assessment determined that there were no obstacles to sighting distances or the visibility of signage, which, on the northern side of the crossing was deemed to be compliant. However, the Office of the National Rail Safety Regulator identified that the advance warning signs on the northern side of the crossing were in the incorrect order. Goondiwindi Regional Council advised that it would schedule works to alter the order of the signs. 

The ALCAM assessment also noted that sun glare could impact visibility for all drivers, although this was only relevant at dawn or dusk. Additionally, the assessment found that the approach from the southern side of the crossing lacked an additional advance warning sign, although this had no bearing on this accident. It also found that the stop lines on both sides of the crossing had become faded and scuffed, most likely through a combination of spillage and impact damage during the collision sequence, and vehicle movements and dirt transfer during recovery activities. Although the stop lines remained compliant and visible, it was recommended that these be repainted. Under the existing interface agreement, this work is the responsibility of Goondiwindi Regional Council, and it has been notified. 

Safety message

This accident highlights the risk of expectation bias in road users when negotiating passive level crossings. Passive controls are common at level crossings where road and rail traffic volumes are low, and it is unlikely that most road users will encounter a train at a level crossing. As road users become familiar with a level crossing, at which they have not previously encountered trains, they can unconsciously form an expectation that no trains will be present each time they approach the level crossing. Accordingly, road users can habituate to behaviours which influence their ability to scan for approaching trains, or detect and respond to relevant signage and road markings.

Passive controls cannot physically prevent vehicles from entering a crossing, and the onus is on road users to follow these controls. This makes passive level crossings particularly vulnerable to driver error (unintentional) or driver decisions (intentional), which can place road users at imminent risk of collision with rail traffic. Where stop signs are provided, the maximum sighting distance occurs when a vehicle is stopped at the stop line. When conducting a ‘rolling stop’ a road vehicle driver will spend less time scanning for oncoming trains, increasing the likelihood of an incorrect decision to proceed into the crossing when it is not safe to do so.

This incident also highlights, for truck drivers, the importance of completing preparatory checks and rectifying any problems which may be observed prior to moving their vehicle, as there is a significant risk of harm and damage when driving vehicles with mechanical issues. 

The investigation

The occurrence

On 23 May 2024, loaded WATCO grain train 6839 was being operated between Thallon and Fisherman Island (Port of Brisbane), Queensland (Qld). This train was a regular scheduled grain‑haulage service operated by WATCO for GrainCorp, and Queensland Rail (QR) was the rail infrastructure manager (RIM).[1] Train 6839’s crew comprised one driver-in-control (DIC) and one driver-assisting (DA), who were rostered to drive the train from Thallon to Goondiwindi. The crew signed on at Goondiwindi station at 0600 local time, and drove a WATCO truck to Thallon. As the train crew were driving west to Thallon, a truck driver was preparing to take a prime mover and flat-top trailer from the St George area to a farm on Gooray Road, Gooray (Figure 1).

Figure 1: Geographic overview of area of operations for train 6839 and truck 

Source: Google Earth, annotated by the ATSB

At about 0700, the truck driver hitched the trailer to the prime mover and found no brake or other mechanical issues with the heavy vehicle during their standard pre-departure tests. The truck driver then drove the heavy vehicle to the farm on Gooray Road and arrived several hours later. En route, the truck driver crossed the Gooray Road level crossing from the south, which they had done at least once before.

The train crew arrived at Thallon in the WATCO truck at about 0815 and conducted a handover with the outgoing crew as the train was still being loaded. Train 6839 finished loading at about 0945, but was then held for 45 minutes by network control to allow the operation of hi-rail vehicle ZE09. The hi-rail vehicle was a 4WD equipped to drive on railway tracks, and departed Thallon at 0932 on a regular track inspection patrol between Thallon and Goondiwindi. Although train 6839 was timetabled to depart Thallon at midday, the crew was granted a proceed authority[2] to depart at 1030.

Train 6839 arrived at Toobeah at around 1256, where it was held for several minutes while the hi‑rail vehicle entered Goondiwindi Yard. Train 6839 was then granted an extended limit of authority[3] to hold short of Signal GI28 on the outskirts of Goondiwindi. Train 6839 departed Toobeah at 1306, with a run time of around 18 minutes to Gooray and around 50 minutes to Goondiwindi. 

By the time train 6839 arrived at Toobeah, the truck driver had exchanged the flat-top trailer for a low-loader trailer at the farm on Gooray Road. The farm workers placed a bobcat or skid steer onto the low-loader, and the truck driver secured this equipment to the trailer. The truck driver intended to take this equipment back to the St George area via the Barwon Highway (Figure 2).

Figure 2: Direction of travel for the truck and train through Gooray

Source: Google Earth, annotated by the ATSB

Prior to departure, the truck driver reported having conducted their standard checks, including a tug test and checks of the lights, air hoses and brake lines. The truck driver reported that during the tug test, the trailer was moving and the wheels were turning, which should not occur when a trailer is hooked up and before the air hoses are connected.

At about 1320, the truck departed the farm and turned left onto Gooray Road, heading south towards the Barwon Highway. To reach the highway, the truck driver had to negotiate the passive level crossing they had crossed that morning, but this time from the north. As the truck and train converged on the crossing at about the same speed, neither vehicle’s driver sighted the other vehicle.

In the prevailing light and weather conditions, all advance warning signs, level crossing signage, and the crossing itself would have been visible to the truck driver from up to 400 m away. At the posted speed limit of 100 km/h, from 400 m, road vehicle drivers would have up to 15 seconds to react and slow in preparation to stop. However, the truck driver reported driving at around 40‍–‍60 km/h, which would have provided them at least 30 seconds to see and respond to the crossing or the signs once these became visible. 

At about the same moment the crossing and warning signs would have come into the truck driver’s view, the train DIC first sounded the train horn. Train drivers were required to sound the horn at a whistle board located about 316 m from the crossing. However, the train DIC sounded the horn earlier, 442 m prior to the crossing, aware that trees immediately prior to the crossing prevented extended line of sight for drivers of vehicles and trains (Figure 3).

Figure 3: Location of train and truck 30 seconds prior to the collision

Source: Google Earth, annotated by the ATSB

As train 6839 passed the whistle board, the DIC first became aware of the truck and estimated the truck was doing at least 40 km/h and did not look like stopping. The train was 19 seconds from the crossing at that time, and the train crew sounded the horn almost continuously from this point.

The truck was then about 210 m from the level crossing, and the truck driver recalled taking their foot off the accelerator. Aware they were approaching the rail corridor, the truck driver then coasted towards the level crossing where they were required to stop. However, the truck driver was not aware of the train until they passed the vegetation (Figure 4), but the truck driver reported that the train horn was audible from the roadway as the train and truck both approached the crossing.

Figure 4: View of railway line as the truck passed the vegetation

Source: WATCO

As the truck driver passed the trees and rounded a slight bend in the road, they saw the train, which was just over 100 m from the crossing. At that point, the truck was around 50 m from the crossing and travelling at around 40 km/h. The truck driver reported that they started to brake but that it felt like the trailer was pushing the prime mover, and assessed that they did not have sufficient braking power to stop before the crossing. As a result, the truck driver decided to accelerate past the stop sign and across the railway line into the path of the train.

Simultaneously, the train crew assessed that the truck was not slowing down or under heavy braking action, as they did not see smoke, skidding, or locked tyres. The train DIC applied the emergency brakes, and both train crew braced against the control console, realising a collision was imminent. The DIC pressed the horn until their toggle broke, after which the DA sounded the horn until moments before impact. The prime mover of the truck cleared the tracks, but the low‑loader trailer was still across the tracks when train 6839 reached the crossing at around 1325.

The lead locomotive of train 6839 hit the middle of the trailer. The locomotive body was sheared off the leading bogies and forced perpendicular to the track before coming to a rest on its side in muddy ground within the rail corridor. The trailing locomotive was also turned perpendicular to the track and impacted the lead locomotive during the collision sequence (Figure 5), and 14 grain hoppers derailed and bunched up in a concertina effect. The locomotives and a dozen grain hoppers were damaged beyond repair, and another 2 grain hoppers sustained substantial damage. 

Figure 5: The WRA-class locomotives, post-collision

Source: WATCO

The prime mover experienced secondary impact damage most likely from the grain hoppers (Figure 6), and the bobcat on the trailer landed a short distance away. The prime mover and low‑loader trailer were damaged beyond repair and the bobcat was substantially damaged.

After the collision, the train crew radioed network control and requested emergency services. The truck driver escaped from the prime mover and assisted the train DA with removing the emergency exit window, through which the train crew subsequently evacuated. The train crew and truck driver sustained serious injuries.

Figure 6: Grain hoppers, prime mover, and low-loader trailer post-collision

Source: Queensland Police Service, annotated by the ATSB

Context

Train driver information

The DIC and DA were experienced train drivers who had worked for over 4 decades in several operational and front-line roles with multiple operators across Qld. They had both been driving trains on the South Western system since the early 2000s and had recently recommenced driving with WATCO after several months of inactivity due to the seasonal nature of the grain harvest.

Train information

Train 6839 consisted of 2 National Railway Equipment Company E2250CC (also known as WRA‑class) locomotives, followed by 40 DGWY-class loaded grain-hopper wagons. The total length of the train was around 600 m, which weighed around 2,500 tonnes prior to departure. 

The train crew reported that the train was operating normally throughout the journey prior to the collision. There were no warning lights, error messages, or known or observed mechanical problems with either the locomotives or the consist. 

Data from the locomotive event logger showed that the train was travelling at or below the speed limit of 60 km/h between Toobeah and the Gooray Road crossing. The event logger also showed that:

• the train was maintaining a near-constant speed at a consistent throttle setting in the minutes immediately preceding the collision
• activation and acknowledgement of the vigilance system[4] occurred at regular time intervals
• there was no uncommanded decrease in brake pressures or degradation in brake performance
• the status of headlights and ditch lights was high/bright and operational prior to collision
• full service/emergency brakes activated seconds prior to collision
• there was near-continuous sounding of the horn from before the whistle board to the crossing.

Truck information

The truck was an articulated heavy vehicle consisting of an Iveco Power Star prime mover, and a Drake low-loader trailer carrying a small CAT skid-steer bobcat. The truck driver reported that they had never towed that trailer before. The driver also reported that during their pre-departure checks, the trailer wheels were moving when they should have been fully locked. This should not have occurred as the air hoses had not been connected to the trailer to release the brakes and allow the wheels to move. 

Although the truck driver had observed this condition prior to departure, no further tests were conducted, and no mechanical repairs or brake adjustments were made to the trailer. The truck driver reported that the first time they attempted to apply the brakes was as they approached the level crossing. When the truck driver applied the brakes to come to a stop, they felt the trailer was pushing the prime mover to the extent that the cab would have been struck by the train. As a result, the driver assessed that their only option was to veer away from the train and accelerate through the crossing. 

After the collision, Queensland Police conducted a forensic examination of the prime mover, but did not examine the trailer. The investigator noted that they were ‘unable to make an accurate determination as to the overall mechanical condition of this vehicle due to the extent of impact damage sustained to the pneumatic braking and suspension systems’. It was also found that ‘the first drive axle left side brake servo exhibited a noticeable air leak with the brake pedal applied’. The police concluded that this condition was ‘unsatisfactory’ and could ‘decrease the overall braking efficiency of the prime mover if brake application was continuous’.

The forensic examination also assessed the condition of the tyres on the prime mover but were ‘unable to make an accurate determination as to the overall tyre condition due to the extent of incident damage and tyres not being located within the inspection site’. However, it also concluded that ‘all mounted tyres on the subject prime mover were of a satisfactory tread condition’.

Level crossing information

Gooray Road

Gooray Road was the main vehicle thoroughfare through Gooray. In the year to 26 May 2024, there were 152 recorded train movements through the Gooray Road level crossing (averaging 3 per week), but frequencies varied due to the nature of the grain harvest. QR advised that 3 train movements per week was standard for the month of May, and that this would increase to 1‍–‍3 trains per day through Gooray in peak grain season. Prior to the hi-rail vehicle, the last train went through Gooray 3 days before the collision.

The Gooray Road level crossing was a public crossing in a rural location surrounded by farmland which saw limited road and rail traffic throughout the year. The crossing had passive controls protecting both approaches on Gooray Road. Passive level crossing controls use signage and road markings to warn road users about an approaching level crossing, but do not activate or change when a train is approaching.

Advance warnings

Two advance warnings were provided on the northern approach to the Gooray Road crossing:

• Stop Sign Ahead signage, with accompanying ‘RAIL’ block text painted on road surface (W3-1)
• Rail Crossing Ahead signage, with accompanying ‘X’ painted on road surface (W7-7).

The first advance warning, about 270 m from the crossing, consisted of W3-1 Stop Sign Ahead signage, and a ‘RAIL’ marking on the road surface, which was required due to the high-speed approach (Figure 7). 

Figure 7: First advance warning, northern approach

Source: WATCO

The second advance warning, about 180 m from the crossing, consisted of W7-7 Rail Crossing Ahead signage on both sides of the road and an ‘X’ marking on the road surface (Figure 8).

Figure 8: Second advance warning, northern approach

Source: WATCO

According to the Australian Standard,[5] the W3-1 Stop Sign Ahead sign was required to be placed 180–250 m before the stop sign at the crossing, and the W7-7 Rail Crossing Ahead sign was required to be placed 70 m before the W3-1 sign. The ‘RAIL’ and ‘X’ road markings were in the correct order, but the signs were in the reverse order. 

Passive controls at crossing

The passive control installed at the Gooray Road level crossing was an RX-2 Stop Sign Assembly, with an accompanying solid white block line painted on road surface. This sign had been installed in 2004 following an Australian Level Crossing Assessment Model (ALCAM) assessment, and the assembly consisted of 4 components and a QR incident plaque (Figure 9).

A white, unbroken stop line accompanied the Stop Sign Assembly. The stop line was required to be placed in a position that maximised sightlines for road traffic, 3.5 m from the nearest rail. A roadside white post with red reflective marker for night visibility next to the assembly was impacted during the collision sequence.

Figure 9: Stop sign assembly and stop line, northern approach

Source: Queensland Rail

Sightlines

Queensland Police also found that road users approaching the crossing from the north had ‘clear, unobstructed views’ in an arc of 180° from 1,700 m before the crossing, including of the rail corridor. Further, at the truck’s reported speed, the truck driver had an opportunity of 60‍–‍75 seconds to see the train before the vegetation blocked their view.

Similarly, Gooray Road should have been visible to rail traffic from about 1,650 m before the crossing. These sightlines did not change greatly until about 650 m before the crossing, when the vegetation started to obstruct them (Figure 10). However, the train crew reported first seeing the truck about 320 m prior to the crossing and the truck driver first sighted the train after passing the end of the vegetation 20 m before the stop line.

Figure 10: Available sightlines as the train and truck approached the crossing

Source: Google Earth, annotated by the ATSB

QR calculated that the sightlines down the rail corridor were at least 1 km at the Stop Sign Assembly in both directions. However, 120 m from the crossing, vegetation obstructed the view of the western side of the rail line from the northern approach on Gooray Road. The rail corridor was then not visible until about 20 m prior to the crossing where the vegetation cleared, after which the sighting distance increased exponentially to almost 1 km at the stop line. This vegetation was the primary factor for the installation of the Stop Sign Assembly in 2004. 

The railway line had a shallow right curve, which began at about the whistle board on the western side of the crossing, but this straightened out 130 m before the crossing. Outside of this curve, there was a straight of almost 9 km on the western side and a 7.5 km straight on the eastern side.

QR also assessed that approaching the crossing from the north, Gooray Road had a minor right curve about 50 m before the crossing, and a minor left curve about 150 m prior to the crossing. However, the road was straight for 2.3 km preceding these curves, and the road alignment maintained direct line-of-sight of the advanced warning signs, the stop sign assembly, and the level crossing from up to 400 m away in clear conditions. 

Road and rail corridor visibility

The truck driver reported that their in-cab visibility and sighting distance was not compromised by the prime mover cab design or any external structures or monuments around the road corridor. Post-incident photos and videos supported this recollection. The truck driver also reported that although there was a crack in the windscreen and a problem with a headlight switch, neither of these factors affected their ability to see the road, the signage, or the crossing. The truck driver also reported that about 1 minute before reaching the crossing, they looked in their side mirror and observed a loose chain on the trailer. However, the driver also reported not having been distracted by anything prior to the accident. 

The train crew reported that there were no impediments to in-cab visibility caused by locomotive cab design, windscreen deformities, or track grade and curvature. They also reported that there were no lineside structures, monuments, or foliage that could have prevented them from seeing the crossing or the whistle boards. Trackworkers on the hi-rail in front of train 6839 similarly did not record any obstructions or any limitations of visibility within the rail corridor prior to the incident. QR also noted that trackside fencing was in good condition and was not tall enough to obstruct visibility within the rail corridor. 

Track condition 

The South Western line through Gooray consisted of a single narrow gauge bi-directional track and used Direct Traffic Control[6] as the primary means of safeworking. Although the published speed limit was 70 km/h in this section, QR implemented a temporary limit of 60 km/h through a series of Train Notices to protect the track condition and structure. This preventative measure was first imposed in 2022 and remained in place at the time of the incident.

Although the track condition was reported as ‘good’, the train DA reported that drivers would travel below the speed limit in that section as it was ‘a smoother ride’. The grade was almost entirely flat between Toobeah and Goondiwindi, including through the Gooray Road crossing. 

Scheduled inspections

As the RIM, QR conducted regular inspections of all crossings between Thallon and Goondiwindi. This consisted of twice-weekly visual inspections from the rail corridor to monitor track condition and known defects, and an annual inspection from the road corridor to assess overall compliance. 

The most recent visual inspection of the Gooray Road crossing had been completed less than 30 minutes before the collision, by the hi-rail that train 6839 had been following since Thallon. The trackworkers onboard hi-rail ZE09 did not identify any track misalignment, defects, or faults at the Gooray Road level crossing on 23 May 2024. They also did not identify any observable issues with the road surface or the road/rail interface as the hi-rail passed through the crossing.

The annual inspection for compliance with civil engineering standards was conducted in February 2024. This inspection found that the road surface within the 10 m approaching the crossing was in average condition, and that the bitumen around the rails was also in average condition. It was also noted that the flangeway and guard rails were compliant and in good condition, as were the signs that formed the passive controls. The inspection further determined that vegetation did not impact visibility and sightlines of the rail corridor from the crossing to the whistle boards on both sides.

Interface agreement

An interface agreement was signed in August 2018 between the RIM (QR) and the Road Manager (Goondiwindi Regional Council), and this agreement remained current at the time of the collision. The agreement stated that ALCAM would be used to identify, monitor, and manage risk at all 42 level crossings in the area, including the one on Gooray Road. The interface agreement outlined the responsibilities that QR and Goondiwindi Regional Council each had for corridor maintenance as well as the provision and the upkeep of relevant signage and markings.

The interface agreement stipulated that all elements needed to follow the relevant Australian Standard[7] and responsibility for each individual aspect of the Gooray Road level crossing was allocated as shown in Table 1.

Table 1: Agreed division of responsibilities for the Gooray Road level crossing

Previous incidents at the Gooray Road level crossing

The train DA did not recall any near-collisions with vehicles, but the DIC reported several previous near-collisions on the South Western line. QR did not identify any previous incidents or collisions at the Gooray Road crossing, but reported that the hi-rail would slow down significantly at the Gooray Road crossing from their usual speed of between 20–40 km/h. This was a precaution as trackworkers had regularly observed vehicles cross in front of the hi-rail without stopping, and that this was particularly common on the northern approach to the level crossing. The train DIC attributed this behaviour to expectation bias caused by the infrequency of trains.

Environmental conditions

Gooray Road was sealed with rough bitumen, around 6 m in width, with a speed limit of 100 km/h. The truck driver described the road as narrow and rough. Photos and videos taken post-incident confirmed the absence of any environmental contamination or degradation of the road surface that could have impacted traction, braking, or steering.

The train crew did not identify any conditions affecting the rail head, or any issues with the grade and curvature of the track that could have influenced the operation of the train. Similarly, no such issues with track condition or weather conditions were reported by the trackworkers on the hi-rail that preceded train 6839. QR analysis and post-incident photos confirmed that weather conditions were sunny with partly cloudy skies and good ground visibility. The road surface and rail head were both dry with zero rainfall recorded, and the temperature was 24–27 °C, which reduced the risk of heat-related track deformities and operating restrictions.

The ATSB assessed whether sun glare may have affected visibility due to the north-south alignment of Gooray Road and east-west alignment of the railway line. However, the collision occurred at 1325 when the sun was at or near its zenith and directly above the vehicles. The truck driver and train drivers all confirmed sun glare was not present prior to the collision.

ATSB safety study

The ATSB safety study Review of level crossing collisions involving trains and heavy road vehicles in Australia (RS-2021-001) analysed collisions between trains and heavy vehicles at Australian level crossings between 1 July 2014 and 31 August 2022, in which 24 rail or road users were fatally or seriously injured. The aim of the study was to improve understanding of the risks associated with level crossing collisions involving heavy vehicles. Common factors identified that may have contributed to the actions of heavy vehicle drivers included the following:

• In at least 12 collisions the heavy vehicle driver had regularly used the level crossing prior to the collision with the train. The drivers' previous experience at the level crossings may have led to a low expectancy for trains and contributed to them not detecting a requirement to stop and give way.
• In at least 14 collisions, the heavy vehicle driver’s view of the track or level crossing protection equipment was obstructed by vegetation, the design of the heavy vehicle cab, poor crossing lighting, or sun glare.
• Consistent with prior research showing that train horns have limited effectiveness for alerting road vehicle drivers approaching level crossings, in at least 25 accidents the horn was not effective at alerting the heavy vehicle driver to the presence of the train.

The report also described a ‘rolling stop’, which involves slowing a road vehicle until a decision is made to proceed into the crossing, without coming to a complete stop. When conducting a ‘rolling stop’ a road vehicle driver will spend less time at the stop point for a passive level crossing and therefore will probably employ less time scanning for oncoming trains. In turn, that increases the likelihood of an incorrect decision to proceed into the crossing when it is not safe.

Locomotive emergency windows

The WRA-class locomotives entered service in 2020 and had been designed so that the windscreen panels in the cabs could be used for emergency egress, unlike older locomotives, which did not have designated emergency exits. Inside the WRA-class cab, 2 stickers were fixed to the windscreen to assist emergency egress. The first sticker illustrated the use of a fire hammer to crack the glass and consisted of a graphic in the corner of the windscreen. This graphic was accompanied by block text which stated ‘In an emergency break glass’ (Figure 11).

Figure 11: Emergency exit sticker with illustration and block text

Source: WATCO

The second sticker was affixed to the top of the windscreen under 2 black handles and listed a 3‑step process to remove the window. On the exterior of the cab, an identical sticker with these instructions enabled people outside to remove the glass (Figure 12).

Figure 12: Emergency exit sticker with written instructions

Source: WATCO

WATCO did not have a specific training module for emergency egress, but drivers were told how to remove the windows as part of their overall ‘shed tuition’. Information about the windows was incorporated into the overall teaching of locomotive mechanical and electrical functionality and componentry. This was supported by the train DA, who explained that drivers had general awareness of the exit windshield, and that the relevant information would likely be in the locomotive manuals. The DIC reported that they had not been shown how to open the window in training but had read the instructions on previous journeys.

The train DIC reported that they had difficulty seeing and reading the instructions after the collision due to debris and dust inside the cab. They also noted that reading the instructions and escaping from the cab would be more difficult at night or in low-light conditions. However, the train DA and truck driver were able to follow the instructions and remove the glass, although the final position of the locomotive and the sticker location may have made it harder to do so. The train DA reported that the glass came out easily, and that the space and access available was sufficient to escape through the window despite the build-up of mud in the cab.

Although the train DIC suggested that training may have assisted in understanding how to use the emergency exit, the DA found the instructions were clear, and the exit was easy to manipulate despite the cab being compromised and the sustained injuries. 

Safety analysis

Passive level crossing

Although the Gooray Road level crossing and its associated infrastructure largely complied with the relevant standards, and the signs and road markings were visible, approaching from the north, vegetation obstructed the view of the rail corridor from the road from 650 m until 20 m before the stop line. There were minor curves on Gooray Road in advance of the crossing, there was sufficient sighting distance for a driver scanning for hazards along the road ahead to react and stop before the level crossing. At the stop line prior to the crossing, the truck driver would have had an unobstructed view of the train. 

The level crossing used passive controls and advance warnings to alert road users to stop and look for rail traffic. Although the advance warning signs were in the incorrect order according to the standard, this did not alter the overall meaning of the signs and road markings, and was unlikely to have influenced the ability of the driver to view, interpret, and respond to the warnings as they approached the crossing. 

The passive controls at the crossing were the primary risk control against collisions but did not activate or change to indicate the presence of a train or physically prevent road users from crossing in front of trains. Consequently, the onus was on road users to stop, sight, and remain clear of rail traffic.

Truck driver actions

As the truck and train converged on the level crossing at similar speeds, the truck driver did not see the train before vegetation blocked their view of the rail corridor. 

When the train crew sounded the horn approaching the Gooray Road level crossing, the truck was approaching the first of 2 advance warning signs at about 40–60 km/h. The truck driver could not see the train or hear the train horn at that distance, but the crossing and advance warnings would have been visible. Although at that point the truck driver had sufficient reaction time and braking distance to stop, they did not attempt to brake until they passed the vegetation just before the crossing and saw the train approaching.

By that time, it was too late for the truck to stop prior to the crossing. There was no evidence that the driver then braked heavily, and the prime mover and trailer brakes may have had reduced effectiveness. Given that the truck driver was aware that the trailer brakes may have been defective, it was likely that they had not slowed the truck sufficiently to stop safely prior to the level crossing. Assessing that the truck would not stop before entering the crossing in front of the train, the driver elected to accelerate to get the truck cab past the train, resulting in the locomotive impacting the trailer. 

Expectation bias 

Given the infrequency of rail movements at the Gooray Road level crossing and that the truck driver had not seen trains when crossing there on previous occasions, they likely did not expect to encounter a train. Research showed that drivers with a low expectancy of trains may be less likely to look for trains or respond appropriately to signage and road markings. The truck’s approach to the crossing was consistent with the driver conducting a rolling stop, in which they would proceed unless a train was sighted. In this case, that resulted in insufficient distance remaining to stop prior to the crossing.

Findings

From the evidence available, the following findings are made with respect to the Level crossing collision between freight train 6839 and truck at Gooray Road, Gooray, Queensland on 23 May 2024.

Contributing factors

• The truck driver did not stop prior to the level crossing as required by the signage.
• Due to the infrequency of trains on that corridor, it is likely that the truck driver had not seen a train at the Gooray Rd crossing on previous journeys. This created an expectation that there would be no conflicting rail traffic at the crossing, which probably reduced the effectiveness of the truck driver’s scan and led to them approaching the intersection at a speed too high to stop prior to the level crossing.

Other findings

• The Gooray Road level crossing configuration provided appropriate signage and stopping distance for a road vehicle driver to notice the passive level crossing controls and bring the vehicle to a controlled stop before the level crossing. Once stopped, there was adequate visibility for a driver to sight a train and give way.

Safety actions

Safety action by Queensland Rail

On 12 June 2024, Queensland Rail initiated an ALCAM assessment and looked at road geometry, traffic volumes (road and rail), visibility and sighting distances, and the existing passive controls. The assessment determined that there were no obstacles to sighting distances or the visibility of signage. The ALCAM assessment deemed the northern side signage to be compliant, however, ONRSR subsequently identified that the advance warning signs were in the reverse order to the standard. It was noted that sun glare could impact visibility for all drivers, although this was only relevant at dawn and dusk.

The assessment found that the signage on the southern side of the crossing lacked an additional advance warning sign, although this had no bearing on this accident. It also found that the stop lines on both sides of the crossing had become faded and scuffed, most likely through a combination of spillage and impact damage during the collision sequence, and vehicle movements and dirt transfer during recovery activities. Although the stop lines remained compliant and visible, it was recommended that these be repainted. Under the existing interface agreement, this work is the responsibility of Goondiwindi Regional Council, and it has been notified.

Safety action by Goondiwindi Regional Council

On 6 December 2024, Goondiwindi Regional Council was advised that the advance warning signs on the northern side were in the reverse order to the standard. The council reported that it would alter the order of the signs on the northern approach to conform with the national standard, and that this rectification work would be scheduled.

Sources and submissions

Sources of information

The sources of information during the investigation included:

• the train crew
• the truck driver
• Queensland Rail
• Queensland Police
• WATCO.

Submissions

Under section 26 of the Transport Safety Investigation Act 2003, the ATSB may provide a draft report, on a confidential basis, to any person whom the ATSB considers appropriate. That section allows a person receiving a draft report to make submissions to the ATSB about the draft report. 

A draft of this report was provided to the following directly involved parties:

• the train crew
• the truck driver
• Queensland Rail
• WATCO
• ONRSR.

Submissions were received from:

• ONRSR.

The submissions were reviewed and, where considered appropriate, the text of the report was amended accordingly.

Purpose of safety investigations The objective of a safety investigation is to enhance transport safety. This is done through:  identifying safety issues and facilitating safety action to address those issues providing information about occurrences and their associated safety factors to facilitate learning within the transport industry. It is not a function of the ATSB to apportion blame or provide a means for determining liability. At the same time, an investigation report must include factual material of sufficient weight to support the analysis and findings. At all times the ATSB endeavours to balance the use of material that could imply adverse comment with the need to properly explain what happened, and why, in a fair and unbiased manner. The ATSB does not investigate for the purpose of taking administrative, regulatory or criminal action. Terminology An explanation of terminology used in ATSB investigation reports is available here. This includes terms such as occurrence, contributing factor, other factor that increased risk, and safety issue. Publishing information Released in accordance with section 25 of the Transport Safety Investigation Act 2003 Published by: Australian Transport Safety Bureau © Commonwealth of Australia 2024 Ownership of intellectual property rights in this publication Unless otherwise noted, copyright (and any other intellectual property rights, if any) in this report publication is owned by the Commonwealth of Australia. Creative Commons licence With the exception of the Commonwealth Coat of Arms, ATSB logo, and photos and graphics in which a third party holds copyright, this report is licensed under a Creative Commons Attribution 4.0 International licence. The CC BY 4.0 licence enables you to distribute, remix, adapt, and build upon our material in any medium or format, so long as attribution is given to the Australian Transport Safety Bureau.  Copyright in material obtained from other agencies, private individuals or organisations, belongs to those agencies, individuals or organisations. Where you wish to use their material, you will need to contact them directly.

On 16 April 2024, a pilot and passenger were conducting a private flight in an Aeropilot Legend 600 aircraft, south of Perth, Western Australia. During cruise, controllability issues were encountered which led the pilot to deploy the aircraft's emergency parachute. 

The aircraft collided with trees prior to impacting the ground near Preston Beach. The pilot and passenger received minor injuries. An inspection of the aircraft identified that the vertical stabiliser had detached from the tail section of the aircraft.

Recreational Aviation Australia (RAAus) commenced an investigation in response to this accident. RAAus requested technical assistance from the ATSB to examine components from the aircraft. To facilitate that request, an investigation under the Transport Safety Investigation Act (2003) has been commenced.

Any enquiries relating to the accident should be directed to RAAus.

The occurrence

On 9 May 2024 an Embraer ERJ 190‑100 IGW aircraft, registered VH‑UZI, was being operated by Alliance Airlines on flight[1] QF1887 from Cairns, Queensland (Qld) to Brisbane, Qld with 29 passengers and 2 flight crew on board. The captain was the pilot flying, and the first officer was the pilot monitoring.[2] 

The aircraft departed Cairns at 2101 local time. By about 2243, the aircraft was stabilised on the Brisbane runway 19L instrument landing system (ILS) approach, descending below 1,000 ft above aerodrome level (AAL) – 1,015 ft above mean sea level (AMSL) – in darkness and in visual meteorological conditions.[3] At 2243:49, air traffic control provided a landing clearance, and shortly after, the captain disengaged the autopilot to manually fly the approach.

To assist with following the approach glidepath, the captain enabled the flight path reference (FPR) line on their primary flight display (PFD) (see the section titled Flight guidance system and displays). The FPR displayed the aircraft’s flight path angle[4] reference line and digital readout on the PFD’s attitude indicator. The reference line was initially presented at 3.2° down, and the captain asked the first officer to adjust the line to present 3.0° (the ILS glideslope angle for Brisbane runway 19L). To do so, the first officer first needed to press the FPR button on their display controller panel to display the reference line on their attitude indicator, and then turn the flight path angle select (FPA SEL) knob to the requested value of 3.0°. At 2244:12, with the aircraft at about 460 ft AAL (475 ft AMSL), the first officer inadvertently pressed the flight path angle (FPA) button, which selected the FPA mode and changed the lateral and vertical navigation guidance for the flight director from localiser (LOC) and glideslope (GS) to aircraft roll angle (ROLL) and flight path angle (FPA). 

The captain reported that the mode change was unexpected for the flight crew, while the first officer reported experiencing ‘startle’.[5] After the mode change, the flight crew focused on troubleshooting the unexpected change and recapturing the ILS flight director modes. With the captain still manually flying, the aircraft’s pitch angle began to decrease, with an associated increase in the descent rate. A few seconds later, the captain selected the approach (APP) navigation mode which armed the ILS approach mode but did not capture the LOC or GS navigation modes. One second later, at 2244:16, the aircraft’s glideslope deviation reached 0.5 dot below the ILS glidepath, and a second later, the vertical rate of descent exceeded the operator’s stabilised approach criteria limit of 1,000 feet per minute, reaching a maximum of 1,139 feet per minute at about 300 ft AMSL. At 2244:19, 7 seconds after the FPA button was pressed, the glideslope deviation increased to about 1.0 dot below, which was the stabilised approach criteria limit, and the lateral guidance (LOC) for the ILS approach mode was captured. 

Over the next few seconds, as the aircraft descended to about 295 ft AAL (310 ft AMSL), the glideslope deviation increased to 1.5 dots. During this time, the captain felt that the aircraft’s nose was low and observed the low glideslope indications on their attitude indicator, as well as the precision approach path indicator (PAPI) system[6] showing 3 red lights, indicating the aircraft was below the glideslope. In response, the captain began to increase the aircraft pitch, and immediately after, with the aircraft still descending, the enhanced ground proximity warning system (EGPWS) generated a glideslope warning (see the section titled Enhanced ground proximity warning system (EGPWS)). The first officer reported calling out ‘slope’ at some point before the EGPWS activation (see the section titled Flight crew task sharing and standard calls). A maximum glideslope deviation of 1.8 dots was reached while the excessive descent rate was being arrested. 

As the aircraft descended to about 235 ft AAL (250 ft AMSL), the vertical guidance (GS) for the ILS approach mode was captured. A few seconds later, when the aircraft was about 1 dot below the glideslope, the descent rate reduced to less than 100 feet per minute. The aircraft then levelled at 233 ft AAL (248 ft AMSL) and the glideslope warning deactivated. Over the next 5 seconds, the captain re‑established the aircraft on the glidepath, and then continued the approach, with the aircraft subsequently landing within the touchdown zone[7] at an appropriate speed without further incident. The circumstances of this occurrence meant that there was no air traffic control alert issued to the flight crew for the glideslope deviation and excessive descent rate.

Context

Flight crew

The captain and first officer both held an air transport licence (aeroplane) and class 1 aviation medical certificates. The captain had over 12,100 hours of flying experience, of which 2,100 hours were on the E190 aircraft type, with 110 hours accrued in the previous 90 days. The first officer had almost 8,200 hours of flying experience, of which 975 hours were on the E190, with 154 hours accrued in the previous 90 days. 

Fatigue

At interview, both pilots reported that they obtained poor quality sleep the night before the day of the incident flight. While the captain was uncertain about why they slept poorly, the first officer reported that they went to sleep about 2.5 hours past their usual bedtime and generally did not sleep well outside of their usual pattern. The captain reported obtaining 8‍–‍9 hours of sleep in the previous 24 hours and 17.5‍–‍19.5 hours in the previous 48 hours while the first officer reported 5.5‍–‍6 hours and 13.5‍–‍14 hours respectively.

The crew also reported feeling ‘moderately’ tired towards the later stage of the flight and that flight crew fatigue was identified in the approach briefing as a threat to be managed. While the incident occurred during the approach, both crew remarked that it was not a high workload situation at the time.

The ATSB conducted an assessment of the flight crew’s sleep opportunity, actual sleep obtained, and quality of sleep leading up to the flight as well as other fatigue‑related factors, identifying that:

• the flight crew had an adequate rest opportunity of about 14 hours before the incident flight
• the rest opportunity was overnight which coincided with the circadian rhythm cycle and was unlikely to increase the risk of fatigue
• although the flight crew reported poor sleep quality, the conditions at the hotel accommodation where they spent the night were suitable and therefore conducive to obtaining restful sleep
• biomathematical modelling[8] of the flight crew’s roster data for the 2 weeks leading up to the flight indicated a low likelihood of fatigue.

In addition, research indicated that:

• the crew’s reported hours of sleep in the previous 24 and 48 hours were within limits that were unlikely to increase the risk of fatigue
• the time the flight crew had been on duty at the time of the incident was unlikely to have increased the risk of fatigue
• the time the flight crew had been awake at the time of the incident was not associated with significant performance degradation
• the time the incident occurred (2245 local time) would not have increased the risk of fatigue as it was outside the window of circadian low.[9]

The assessment concluded it was unlikely the flight crew were experiencing a level of fatigue known to adversely affect performance. 

Instrument landing system

An instrument landing system (ILS) is an instrument approach procedure that provides lateral (localiser) and vertical (glideslope) position information using angular deviation signals from the localiser antennas (located past the upwind end of the runway) and the glideslope antennas (located approximately 1,000 ft from the runway threshold). Aircraft systems detect these radio signals and provide instrument indications which, when utilised in conjunction with the flight instruments, enable an aircraft to be manoeuvred along a precise final approach path.

The Brisbane runway 19L ILS approach provided the typical 3° glideslope to the runway (Figure 1).

Figure 1: Brisbane runway 19L ILS approach chart

Source: Airservices Australia, annotated by the ATSB

Flight guidance system and displays

The E190 featured an integrated automatic flight control system (AFCS) that processed inputs from several aircraft systems and sensors. The AFCS supplied this processed data to the flight guidance control system (FGCS), which provided visual and aural information to the flight crew. 

The E190 FGCS also provided flight guidance information to the primary flight display (PFD) flight director (Figure 2) and the autopilot. The flight mode annunciation (FMA) display was located at the top of the PFD and displayed autothrottle, autopilot, approach status, and flight director lateral and vertical mode indications.

The attitude director indicator (ADI) was located below the FMA display on the PFD, and presented the following:

• flight director represented by a magenta diamond providing lateral and vertical guidance
• glideslope and localiser deviation pointers with scales (1 dot spacing), independent of flight director guidance
• flight path angle symbol that showed the current flight path angle in reference to the horizon line
• flight path reference line (FPR) and readout which indicated a manually selected flight path angle for reference.

Figure 2: E190 flight mode annunciation and attitude director indicator displays

Source: Alliance Airlines, annotated by the ATSB

The flight director provided guidance based on pilot selections on the guidance panel (Figure 3). When a mode change was selected by the flight crew, the selected mode was armed and, when certain conditions were met, became active. This active mode was displayed on the FMA display which temporarily flashed in reverse video (black text on green background) to highlight the change. 

Pressing the ‘APP’ button armed approach navigation modes, and when on an ILS approach, activated the ILS approach mode, providing vertical (glideslope) and lateral (localiser) flight director guidance. This navigation mode was displayed on the FMA as ‘LOC’ and ‘GS’. 

Figure 3: Guidance panel and display controller panel

Source: Alliance Airlines, annotated by the ATSB

Pressing the FPA button selected the flight path angle vertical mode and the aircraft roll hold (ROLL) lateral mode. When the FPA mode was active, it commanded the flight director to a flight path angle reference and the flight path reference (FPR) line was displayed as a solid line. The FPA SEL control knob was then used to manually select the desired flight path angle, represented by the FPR line. 

The FPR line feature could also be used by pilots to assist with flight path management when manually flying an ILS approach and was activated by pressing the FPR button located on the display controller panel. The FPR line was then presented as a dashed line when activated and adjusted using the same FPA SEL control knob on the guidance panel (Figure 3). When the FPR button was pressed, the line and numerical flight path angle value presented was that of the aircraft’s flight path angle at that time.

The first officer reported that they had previous experience using the FPR function, but it was not a feature commonly used by the operator’s flight crew. 

The aircraft manufacturer advised the ATSB that it had not received any previous flight crew reports of a similar inadvertent selection of FPA instead of FPR as occurred during this incident. 

Enhanced ground proximity warning system (EGPWS)

The aircraft was fitted with a Honeywell EGPWS, which used aircraft position and configuration information, along with a radio altimeter and a terrain database, to provide flight crew with increased awareness of the terrain along the projected flight path via aural and visual alerts and warnings. These included a mode that alerted pilots to excessive glideslope deviation during an ILS approach (Figure 4), and excessive descent rate (Figure 5).

When the aircraft descended more than 1.3 dots below the glideslope while at a radio altitude less than 1,000 ft, an aural ‘GLIDESLOPE’ would be generated and an amber ‘GND PROX’ alert displayed on each PFD (‘soft’ glideslope alert). If the descent continued to less than 300 ft and the glideslope deviation was 2 dots below, the ‘GLIDESLOPE’ aural alert would be louder and faster (‘hard’ glideslope alert). The aural and visual alerts continued until the aircraft exited the alert envelope.

Figure 4: EGPWS descent below glideslope alert

Source: Alliance Airlines

When the aircraft altitude was lower than 2,450 ft above ground level, aural and visual alerts were generated when the EGPWS calculated that the aircraft had an excessive descent rate towards terrain (Figure 5). When the outer limit of the descent rate envelope was breached, an aural ‘SINKRATE’ would be generated with an amber ‘GND PROX’ alert displayed on each PFD. If the inner limit was breached, a ‘PULL UP’ aural and visual alert was generated.

Figure 5: EGPWS excessive descent rate alert

Source: Alliance Airlines

Recorded data

The ATSB was notified of the incident 4 days after it had occurred. By this time, the cockpit voice recorder audio covering the time of the incident had been overwritten and was unavailable to the investigation. 

The flight data from the aircraft’s quick access recorder was analysed by the ATSB and the aircraft manufacturer, Embraer (Figure 6). The data showed that at 2243:57, the autopilot was disconnected, and the aircraft closely followed the glideslope until 2244:12, when the first officer inadvertently selected the FPA mode. At that time, the aircraft’s flight path angle was 3.3° down with a pitch angle of about 2.5° up. Over the next 10 seconds, the control column pitch up input reduced, with a subsequent reduction in aircraft pitch up angle, and the descent rate and deviation from the glideslope both increased.

Between 2244:18 and 2244:23, the descent rate exceeded 1,000 feet per minute, reaching a maximum of 1,139 feet per minute at about 300 ft above ground level. This was outside of the EGPWS excessive descent rate activation envelope and, therefore, no ‘SINKRATE’ alert was generated. At 2244:21, the glideslope deviation reached 1.0 dot, and 2 seconds later, the captain pitched the aircraft up, with the EGPWS ‘GLIDESLOPE’ alert activating immediately after. Three seconds later, a maximum glideslope deviation of 1.8 dot was recorded and the EGPWS alert deactivated after a further 5 seconds. The glideslope deviation reduced below 1.0 dot at 2244:31, and about 5 seconds later, the aircraft recaptured the glideslope at around the approach minima (220 ft AMSL). The maximum deviation below the glideslope was approximately 60‍–‍70 ft during the EGPWS activation.

Figure 6: VH-UZI recorded flight data during the approach

Source: ATSB

Operator procedures 

Approach briefing

The objective of an approach briefing is to ensure all flight crew understand and share a common mental model for the proposed plan of action. The approach briefing was normally performed by the pilot flying with the pilot monitoring reviewing and checking the information. 

The operator’s procedures required that the flight crew cover several topics during the briefing such as the expected manoeuvring to the initial approach fix, nomination of navigation aids required for the approach (for example, ILS), terrain, weather, obstacles, and any threats. 

The captain stated that the approach briefing for the occurrence flight was ‘normal’ other than fatigue being identified as a threat (see the section titled Fatigue). The captain stated that they did not brief the use of the FPR line feature during the approach briefing as they only decided to use it after the approach had already commenced and following the autopilot disconnection. 

Stabilised approach criteria

An approach is stable when all of the stabilisation criteria specified by the operator are met and an unstable approach is any approach which does not meet these criteria. According to an International Air Transport Association (IATA) report published in 2017, historical commercial aviation accident data indicated that many accidents occur during the approach and landing phase of flight, with frequent contributing factors being an unstable approach together with a subsequent failure to initiate a go‑around. Failure to maintain a stable approach could result in a landing that is too fast or too far down the runway, leading to a hard landing, runway excursion, loss of control, or collision with terrain. 

The operator’s standard operating procedures required all flights conducting instrument approaches to be stabilised by 1,000 ft above aerodrome level, and an immediate go‑around was required for any approach that did not meet the following stabilised approach criteria:

a) the correct flight path;

b) only small changes in heading/pitch are required to maintain the correct flight path;

c) the aircraft speed is not more than VAPP + 10 knots indicated airspeed and not less than VREF;

d) the aircraft is in the correct landing configuration;

e) sink rate is no greater than 1,000 feet per minute

f) thrust or power setting is appropriate for the aircraft configuration;

g) all briefings and checklists have been completed;

h) specific types of approached are stabilized if they also fulfil the following

i. instrument landing system (ILS) approaches must be flown within one dot of the glideslope and localizer

ii. a Category II or Category III ILS approach must be flown within the expanded localizer band

i) unique approach procedures or abnormal conditions requiring a deviation from the above elements of a stabilized approach require a special briefing to have been completed prior to beginning the approach.

• Note 1: A momentary excursion is permitted for points (c) & (e). A momentary excursion is defined as a deviation lasting only a few seconds and where every indication is that it will return to the stabilised criteria as listed in points (c) & (e).

• Note 2: Where the nominal descent path for a particular approach requires a descent rate greater than 1000 fpm. This is only permitted when expected rates of descent have been briefed prior to the approach being commenced.

Flight crew task sharing and standard calls

During a manually flown approach, the pilot flying was responsible for controlling the aircraft flight path. The pilot monitoring was responsible for performing actions requested by the captain and monitoring the aircraft status (for example, configuration, altitude, speed, and flight path). For precision approaches, such as an instrument approach, the pilot monitoring was required to call out flight path deviations, such as glideslope deviations, below the stabilisation height:

Any time the [pilot monitoring] calls deviations from 'on slope' the PF should make corrections to avoid flight path excursions towards full scale.

The [pilot monitoring] should continue slope deviation calls until the glideslope indicator stops moving toward full scale and whenever the indicator is at full scale.

Defined phraseology was used to standardise communication of critical items in high workload situations. Deviation calls were to be made if a deviation limit was exceeded, and no corrective action had been observed (Table 1).

Table 1: Relevant standard calls

EGPWS

The operator’s E190 EGPWS policies and procedures required the crew to take the following action in response to an EGPWS alert:

If an EGPWS alert is associated with a PFD AMBER visual message of ‘GND PROX’, the EGPWS WARNING CORRECTIVE MANEUVER must be performed unless on daylight operations with clear visual conditions (not IMC), and a positive visual verification ensures that no obstacle or terrain hazards exist.

…During daylight in VMC, with terrain and obstacles clearly in sight, the alert may be considered cautionary. Take positive corrective action until the alert ceases or a safe trajectory is ensured. 

Perform the appropriate GPWS warning or alert procedure at all other times and climb the aircraft to the [lowest safe altitude] when enroute or to the [minimum safe altitude] when in the terminal area.

The ‘EGPWS WARNING CORRECTIVE MANEUVER’ required the pilot flying and pilot monitoring to perform various actions and callouts (Figure 7). 

Figure 7: EGPWS warning corrective manoeuvre

The operator advised that flight crew received training on the EGPWS during initial type training and through recurrent cyclic simulator training sessions. The training involved different scenarios involving an EGPWS alert that required the corrective manoeuvre to be performed, with the training focus being on ensuring that the procedure was executed correctly. The operator advised that while EGPWS glideslope alerts were probably not simulated as frequently as other EGPWS alerts, the response to almost all EGPWS alerts, as specified in the operator’s procedures, was to perform the corrective manoeuvre. 

The captain reported being surprised when the EGPWS glideslope alert activated as they could see they were ‘very near the place they needed to be’ and assessed that the safest course of action was to continue the approach.

Safety analysis

Incorrect mode selection

While manually flying the Brisbane runway 19L instrument landing system (ILS) approach, the captain enabled the flight path reference (FPR) line function and requested the first officer adjust the FPR value to the runway ILS glideslope angle of 3.0°. This required the first officer to press the FPR button and then turn the FPA SEL knob to the requested value. Instead, the first officer inadvertently pressed the FPA button.

The first officer’s action constituted a ‘slip’ type error that is a failure of an execution of an action (Reason, 1990). Specifically, slips occur when an individual’s understanding of the situation is correct, but the wrong action is performed (Wickens et al, 2022). Characteristics of this error also occur when people accept a match for the proper object, something that looks like it, is in the expected location or does a similar job. Specifically, it can occur when some characteristics of either the stimulus environment or the action sequence itself are closely related to the wrong action. It occurs during well practiced tasks where the operator may not be carefully monitoring their own action selections (Salvendy and Karwowski, 2021).

The first officer’s prior intention was to press the FPR button, which was a routine action, but this did not go as planned. In addition, the FPA and FPR buttons were both used in conjunction with FPA SEL control knob, which the first officer would have needed to turn after pressing either button. The aircraft manufacturer reported they were unaware of any similar occurrences that would indicate that this was a significant ergonomic issue. 

The captain had not briefed the use of the FPR line function during the approach briefing and therefore the flight crew did not have a shared mental model regarding the use of this function during the approach. However, it was unlikely that this influenced the first officer’s ‘slip’ type error as the first officer knew, and intended to press, the correct button.

Pressing the FPA button disengaged the ILS approach mode and changed the active flight director modes from glideslope and localiser to flight path angle and roll mode on the flight mode annunciator display. As a result, the flight director moved to the aircraft’s flight path angle at the time of the button press (3.3° nose down), and the flight path reference line turned solid with the readout indicating 3.3°. Although the ILS approach can be flown without the flight director guidance, the change to the flight director mode was unexpected and resulted in the crew diverting their attention to correct the mode change.

Diversion of attention

Although the first officer reported being ‘startled’ when the flight director mode change occurred, they were more likely experiencing ‘surprise’, which is when a mismatch is detected between what is observed and what is expected (Rankin et al, 2013). Surprise can be described as a ‘… combination of physiological, cognitive, and behavioural responses, including increased heart rate, increased blood pressure, an inability to comprehend/analyse, not remembering appropriate operating standards, “freezing,” and loss of situation awareness’ (Rivera et al, 2014).

At the time, the flight crew did not expect a mode change and were surprised when it occurred. After the mode change, the flight crew diverted their attention from monitoring the flight path and became preoccupied with resolving the mode change and recapturing the ILS flight director modes. As a result, the aircraft descended below the glideslope and the approach became unstable with respect to glideslope deviation, which was not recognised by either flight crewmember. The descent rate also exceeded the stabilised approach criteria limit of 1,000 feet per minute for about 6 seconds, although the criteria allowed for ‘momentary excursions’ of this parameter.

The first officer reported calling out ‘slope’, however they did not make the ‘slope limit’ or ‘unstable’ call to indicate that a go‑around was required. Furthermore, the time of the ‘slope’ call in relation to the glideslope deviation could not be determined. However, if these calls were made, they were unlikely to have affected the outcome given that the captain began to recover the aircraft’s descent shortly after the 1.0 dot glideslope criteria was exceeded. 

Research has highlighted the difficulties in understanding the aircraft’s present state following a surprising event, which includes an inadvertent mode change (Rankin et al, 2016). As a result, there are also challenges in identifying which response is appropriate. In such circumstances, immediate reference to the flight mode annunciation display offers the best opportunity to promptly identify and resolve the situation. When focus is diverted to a primary task such as manual flying or emergency actions, attention narrows to that task, and so monitoring of other sources degrades (CAA, 2023). This degradation of monitoring often occurs without the flight crew realising it.

Response to ground proximity warning system alert

Shortly after the aircraft exceeded 1.0 dot glideslope deviation, the captain recognised that the aircraft was too low and initiated a pitch up manoeuvre to correct the deviation. Immediately after, the enhanced ground proximity warning system (EGPWS) glideslope alert activated, which was heard by the flight crew. As the alert occurred at night, procedures required that the EGPWS corrective manoeuvre be performed. The EGPWS glideslope alert also indirectly indicated to the flight crew that the aircraft had exceeded the stabilised approach criteria for glideslope deviation. However, the flight crew did not perform the required EGPWS corrective manoeuvre and continued the unstable approach until the aircraft landed. The decision not to perform the corrective manoeuvre and to continue the approach increased the risk of landing too fast or too far down the runway, which in turn increased the risk of a hard landing, runway excursion, loss of control, or collision with terrain.

Findings

From the evidence available, the following findings are made with respect to the unstable approach involving Embraer 190, VH‑UZI, about 4 km north‑east of Brisbane Airport, Queensland on 9 May 2024. 

Contributing factors

• In response to a request from the pilot flying to adjust the flight path reference line on their primary flight display, the pilot monitoring inadvertently disengaged the aircraft’s instrument landing system approach mode by mis‑selecting the flight path angle mode.
• Following the unexpected change to the aircraft’s flight modes, the flight crew diverted their attention to recapturing the instrument landing system approach mode and did not effectively monitor the aircraft's flight path. Consequently, the aircraft exceeded the glideslope limit requirement of the stabilised approach criteria undetected by the flight crew.
• The aircraft continued to descend below the glideslope, resulting in the Enhanced Ground Proximity Warning System ‘GLIDESLOPE’ alert. Subsequently, the flight crew did not perform the required terrain avoidance manoeuvre, and instead continued the approach.

Safety actions

Safety action by Alliance Airlines

Following the occurrence, the operator conducted an internal review of the incident, including interviews with the flight crew and analysis of flight data to assess procedural adherence and identify contributing factors. 

In response to the occurrence, Alliance Airlines implemented the following to enhance safety and learning:

• a discussion was added in the pre‑brief of the cyclic training program to include the EGPWS ‘glideslope’ activations (hard and soft) and required procedures
• issued an Operational Notice to remind crew of the stabilised approach criteria and go‑around requirements
• conducted a thematic review of unstable approaches and analysed data for further review.

Sources and submissions

Sources of information

The sources of information during the investigation included:

• the flight crew
• quick access recorder data
• Alliance Airlines
• the aircraft manufacturer (Embraer)
• Airservices Australia.

References

Civil Aviation Authority (CAA) (2023) Flight-crew human factors handbook, Civil Aviation Authority, United Kingdom Government.

IATA (2017). Unstable Approaches – Risk, Mitigation Policies, Procedures and Best Practices (3^rd ed.). Retrieved from https://www.iata.org/contentassets/7a5cd514de9c4c63ba0a7ac21547477a/iat….

Rankin A, Woltjer R, Field J and Woods D (25–27 June 2013) ‘Staying ahead of the aircraft’ and managing surprise in modern airliners [conference presentation], 5th Resilience Engineering Symposium, The Netherlands, accessed 16 October 2024.

Rankin A, Woltjer R and Field J (2016) ‘Sensemaking following surprise in the cockpit—a re-framing problem’, Cognition, Technology & Work, 18:623–642, doi: 10.1007/s10111-016-0390-2.

Reason J (1990) Human error, Cambridge University Press, Cambridge, United Kingdom. 

Rivera, J., Talone, A. B., Boesser, C. T., Jentsch, F., & Yeh, M. (2014). Startle and Surprise on the Flight Deck: Similarities, Differences, and Prevalence. Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 58(1), 1047-1051. https://doi.org/10.1177/1541931214581219

Salvendy G and Karwowski W (2021) Handbook of human factors and ergonomics, 5th edn, John Wiley & Sons, New Jersey.

Wickens CD, Helton WS, Hollands JG and Banbury S (2022) Engineering psychology and human performance, 5th edn, Routledge, New York.

Submissions

Under section 26 of the Transport Safety Investigation Act 2003, the ATSB may provide a draft report, on a confidential basis, to any person whom the ATSB considers appropriate. That section allows a person receiving a draft report to make submissions to the ATSB about the draft report. 

A draft of this report was provided to the following directly involved parties:

• the flight crew
• Alliance Airlines
• Qantas Airways
• Embraer
• United States National Transportation Safety Board
• Brazilian Aeronautical Accidents Investigation and Prevention Center
• Civil Aviation Safety Authority
• Airservices Australia.

Submissions were received from:

• Alliance Airlines
• Civil Aviation Safety Authority.

The submissions were reviewed and, where considered appropriate, the text of the report was amended accordingly.

Purpose of safety investigations The objective of a safety investigation is to enhance transport safety. This is done through:  identifying safety issues and facilitating safety action to address those issues providing information about occurrences and their associated safety factors to facilitate learning within the transport industry. It is not a function of the ATSB to apportion blame or provide a means for determining liability. At the same time, an investigation report must include factual material of sufficient weight to support the analysis and findings. At all times the ATSB endeavours to balance the use of material that could imply adverse comment with the need to properly explain what happened, and why, in a fair and unbiased manner. The ATSB does not investigate for the purpose of taking administrative, regulatory or criminal action. Terminology An explanation of terminology used in ATSB investigation reports is available here. This includes terms such as occurrence, contributing factor, other factor that increased risk, and safety issue. Publishing information Released in accordance with section 25 of the Transport Safety Investigation Act 2003 Published by: Australian Transport Safety Bureau © Commonwealth of Australia 2025   Ownership of intellectual property rights in this publication Unless otherwise noted, copyright (and any other intellectual property rights, if any) in this report publication is owned by the Commonwealth of Australia. Creative Commons licence With the exception of the Commonwealth Coat of Arms, ATSB logo, and photos and graphics in which a third party holds copyright, this report is licensed under a Creative Commons Attribution 4.0 International licence. The CC BY 4.0 licence enables you to distribute, remix, adapt, and build upon our material in any medium or format, so long as attribution is given to the Australian Transport Safety Bureau.  Copyright in material obtained from other agencies, private individuals or organisations, belongs to those agencies, individuals or organisations. Where you wish to use their material, you will need to contact them directly.

The investigation

The occurrence

On 11 April 2024, a Cessna 404 Titan was being operated by Aero Logistics on an aerial survey flight north of Emerald, Queensland. The flight was being operated under visual flight rules,[1] with one pilot and 2 task specialists operating the survey equipment. At 1317 local time, the aircraft departed Emerald Airport to the north-east before turning towards Moranbah and climbing to an altitude of 5,000 ft (Figure 1).

Figure 1: Survey area location and recorded flight path between Emerald and Moranbah

Source: Google Earth and operator ADS-B data, annotated by the ATSB.

At around 1349, the aircraft approached the Moranbah area and was descended to 4,000 ft to begin the first survey run. This required the aircraft to maintain a constant altitude and speed while travelling along specific parallel ‘lines’ overhead the target area as instructed by the task specialists. Approximately 20 minutes into this run, all 3 crew members started to notice a sporadic smell in the cabin, although their recollection of the smell differed.

The pilot advised that, as the flight progressed, they noticed that they had increasing difficulty setting the aircraft up and aligning it correctly. One of the task specialists also began to feel affected by the fumes, and at 1449 the pilot cancelled the survey and commenced a return to Emerald Airport. The first task specialist moved to the rear of the aircraft due to the extent of their respiratory symptoms. The second task specialist was not experiencing any symptoms at this stage, and repositioned to the cockpit to assist the pilot if needed.

Return to Emerald

During the return to Emerald, the crew opened the windows, vents, and cabin door to ventilate the cabin. They also disconnected the survey equipment and checked several aircraft systems, including ensuring that the autopilot was selected off, in an attempt to control or reduce the fumes. None of these measures resulted in any improvement, and the pilot and task specialists reported experiencing worsening symptoms as the flight progressed, but the nature and extent of these symptoms varied between each person. Although several diversionary airports were available en route, the pilot chose to continue to the base at Emerald.

Approximately 15 minutes from Emerald, the pilot considered conducting a precautionary landing in a field, due to worsening symptoms. They initiated a descent, and made a broadcast to Brisbane Centre, advising they were ‘landing somewhere in a field hopefully’.

Recorded flight data showed that, during this descent, the aircraft reached a maximum descent rate of 2,664 ft/minute. The pilot levelled the aircraft off at around 2,200 ft, approximately 14.5 km north‑east of Capella Airport. However, the pilot did not consider a diversion to Capella at that time. The pilot advised that they had managed to get some fresh air and decided to continue.

Several minutes later the pilot observed that the trim wheel started what they assessed as an uncommanded nose down input, and that ‘the autopilot was actually now trying to push us into the ground.’ In response, the pilot asked the task specialist seated beside them to hold the trim wheel, which rectified the issue. The task specialist confirmed that the trim wheel was moving but that they did not require significant force to stop it. They also advised that it was possible the pilot accidently activated the electric trim switch on the control column.

Landing 

The pilot advised air traffic control that they were experiencing fumes in the cabin, but they did not mention any control issues with the aircraft. They declined an offer for emergency services on arrival, and no MAYDAY[2] or PAN PAN[3] calls were heard or recorded by air traffic controllers at any point during the flight. The pilot continued to Emerald where, after confirming the wind direction on their electronic flight bag, and informing other traffic of their intentions, tracked for a right base leg of the circuit for runway 24,[4] and landed at 1538.

Although the pilot had not requested emergency services, the Brisbane Centre controller called the aerodrome reporting officer (ARO) at Emerald Airport to advise them of the aircraft’s approach. The ARO then called emergency services at around 1530, and several Queensland Fire and Emergency Service (QFES) appliances were waiting at the parking bay when the aircraft landed. The crew was able to disembark unassisted, but one task specialist exited the aircraft and lay down on the grass next to the aircraft due to nausea. The second task specialist reported having a headache towards the end of the flight. 

The ARO did not detect any fumes or smoke when they opened the rear door of the aircraft. The operator’s chief engineer entered the aircraft and similarly could not identify any smells, fumes, or smoke inside the aircraft. QFES crews then attended the aircraft, conducting a thermal scan and gas sampling of the interior with nil results. All of the internal panels and the flooring were removed for inspection and the QFES crews did not detect any fluid leaks internally or externally, but the aircraft was isolated overnight as a precaution.

The crew was attended to by paramedics at the scene before being transported to Emerald Hospital. They were given several hours of high-flow oxygen as a standard treatment and cleared to leave hospital later that evening. Blood samples were not taken from the crew as Emerald Hospital did not have the equipment required for blood gas testing. The aircraft was temporarily withdrawn from service for additional examination and testing.

Context

Crew information

Pilot

The pilot was experienced with piston and turboprop aircraft, and possessed the relevant qualifications and competencies for the work being conducted. The pilot had several decades of experience conducting freight, passenger, and recreational flights with a number of operators across Australia and overseas. The pilot had recently joined the operator to operate their piston aircraft. The pilot had passed their most recent aviation medical examination in July 2023.

Task Specialist 1

Task specialist 1 (TS1) was experienced with aerial survey work, having worked for a number of years with the owner/operator of the aerial survey equipment installed on the aircraft. Most of their experience had been in Cessna 404 and 406 series aircraft, and they had not previously experienced fumes or smells inside the cabins of these aircraft.

Task Specialist 2

Task specialist 2 (TS2) had limited experience on the Cessna 404 and had spent most of their time on Cessna 406 and 441 series aircraft. At the time of the occurrence, they were in the process of obtaining an aeroplane pilot licence, with several hours of flight time already logged. TS2 had experienced fumes in another aircraft several years earlier caused by a fault in the air conditioning system, but had not encountered anything similar in the Cessna 404.

Response to fumes and symptoms

When the crew decided to cancel the survey run, the aircraft was around 37 km north‑west of Moranbah Airport and Emerald was around 200 km south. The pilot advised that they would have had to declare an emergency to land at Moranbah Airport, as it was a private airport, and at that stage they did not consider there was an emergency. They had also experienced a flap issue the day before with resultant fumes in the cockpit and they considered this to be a similar event.

The pilot had completed emergency response and hypoxia awareness training in November 2023 and had been issued a pulse oximeter as part of the operator’s standard induction process. The aircraft had oxygen equipment onboard and the pilot and both task specialists were trained in how to use this equipment. The task specialists confirmed that the pilot had given an emergency briefing prior to departure. They could not recall whether the onboard oxygen equipment had been mentioned specifically, however there had been no plan to fly above 10,000 ft. 

The crew confirmed that during the return to Emerald, they used a supplied pulse oximeter to assess the oxygen saturation levels in their blood several times – however, none of the crew considered using the onboard oxygen equipment. 

Aircraft information

The aircraft was a Cessna Aircraft Company 404 Titan, manufactured in 1978 and equipped with 2 Teledyne Continental GTSIO-520-M engines. It was one of 7 such aircraft in the operator’s fleet. The operator told the ATSB that these aircraft were not modified from the original Cessna 404 design apart from the floor cutouts for the survey equipment unit. This unit was placed above the fuselage cutout at the rear of the cabin and secured to the floor with the cameras and sensors facing downwards.

The survey operators would usually place aluminium tape or cardboard around the unit to seal any gaps between the unit and the cutout in the floor. The operator also confirmed that the cable routing and electrical harnesses for the equipment and aircraft systems had been arranged and routed in a standard configuration around the cutouts in the fuselage.

Maintenance information

On 10 April, the day prior to the incident, the same aircraft with the same crew was conducting similar survey runs north of Emerald. The pilot said that during this flight they experienced an uncommanded extension of the flaps to the full-down position after the flaps were extended to Flap 10. The pilot attempted to troubleshoot the problem by moving the flap lever back and forth, but the flaps did not respond to movements of the flap lever. The pilot also began to notice a smell in the cabin when the flap issue commenced, and that their airspeed was lower than expected. The pilot decided to return to Emerald, but did not advise the maintenance personnel of the presence of fumes.

An engineering inspection found that the wires on the micro-switch had detached from the flap select lever. The operator said that this inspection also found that ‘the wire hadn't touched any surrounding areas, hadn't tripped, hadn't burned or anything like that’. The wires were checked and reconnected on the morning of 11 April, and the flap lever was tested before the aircraft was released back into service that day. No flap issues were detected after this repair was made.

Following the incident flight, the operator undertook flight tests, scheduled maintenance, and strip‑down examinations of aircraft wiring and componentry, including inspections and testing of:

• panels and flooring, which were removed to inspect for electrical wiring damage
• the autopilot and electrical trim systems
• around and underneath the cockpit dash and pilot side electrical panel and relay boxes
• the battery and lighting systems
• the heater and ventilation systems
• the engines, including alternators and electrical looms
• the hydraulic system
• the nose and main landing gear bays and wiring
• the tail interior and flight controls
• the left and right inner wings
• the flap coves and flight controls
• the nose locker and wing lockers including interior lighting.

No mechanical or wiring issues that could have been related to fumes were detected. The survey operator found no wiring defects or component faults in their survey equipment unit, which was then reassembled and installed on a different aircraft. The operator also confirmed that there was a card-based carbon monoxide and a digital aural CO detector on the aircraft, and that neither of these detectors had activated during the incident flight or the test flights conducted after the occurrence. Additionally, neither the pilot nor task specialists recalled either CO monitor activating during the occurrence flight.

Environmental

Weather

Weather data was requested from the Bureau of Meteorology (BOM) for winds aloft and automated METARs[5] in the area between Emerald and Mackay. This data confirmed that visual meteorological weather conditions[6] existed, with temperatures around 25 ˚C and cloud cover[7] ranging between few and scattered up to altitudes of around 5,000 ft. These conditions were consistent for the duration of the flight. 

Bushfires and mining activity

The crew did not observe any blast fumes or mining activity in the Moranbah area during the incident flight. BOM data also confirmed no known plumes and smoke in the area. Queensland government data on mining activity and blast fumes was requested but this was not provided. The pilot did not indicate the presence of any bushfires or smoke that could have generated fumes, and TS2 stated that there were some minor grass fires in the area, but these were not in the immediate vicinity of the flight path.

The operator confirmed through post-flight testing that ground-based fumes and smells could ‘be picked up within the cabin with the environmental settings open in the venting configuration’. The operator also described a smoke smell in the cabin during the post-incident test flight when the aircraft flew directly overhead a local grass fire. However, this smell was only temporary and limited to the specific fire area and was markedly different to the smell described by the crew.

Safety analysis

During the return flight to Emerald following the onset of impairment symptoms, the crew could not identify the source of the fumes inside the cabin. Based on the available evidence, no definitive cause or source for the fumes could be established. There were no visible signs of smoke or leaks, and there were no environmental conditions that could have consistently generated fumes inside the cabin. The onboard CO detectors did not activate, and post-incident testing by QFES did not identify noxious gases or thermal hotspots. The operator was also unable to identify any faults or defects in the wiring or componentry of the onboard systems and could not replicate the fumes in test flights.

Although the decision to cancel the survey run was prudent, flying back to Emerald, rather than diverting to closer suitable airports exposed the crew to the fumes for longer than necessary and may have worsened the impact. However, this decision may have been influenced by the flap‑related incident on the previous day. In addition, although the crew was trained to use the supplemental oxygen equipment onboard, and repeatedly used the pulse oximeters, the crew did not consider using the available oxygen. 

It is probable that all the crew members were affected by the fumes, although these symptoms presented differently in each crew member. TS1, sitting at the rear of the aircraft, reported nausea and respiratory symptoms. TS2 reported experiencing a headache. The pilot did not experience any physical symptoms but may have been experiencing cognitive impairment in terms of their decision-making and aircraft handling, even though the pilot navigated and controlled the aircraft as per the established procedures and communicated effectively on the radio.

The pilot’s decision not to declare a MAYDAY or PAN PAN, or to have emergency services in attendance for the landing, may have been a result of their impaired decision‑making. Fortunately, the controller proactively initiated the emergency response by alerting the ARO.

Findings

From the evidence available, the following findings are made with respect to the fumes event involving Cessna 404, VH-LAD, near Moranbah, Queensland on 11 April 2024.

Contributing factors

• The operating crew was affected by fumes in the cabin and the pilot returned the aircraft to Emerald rather than landing at a closer suitable alternate airport.

Other findings

• Despite extensive post‑occurrence ground and in‑flight examination, the source of the fumes could not be established.

Safety actions

Safety action by Aero Logistics

Prior to the commencement of the ATSB investigation, the operator proactively took several steps in response to the incident.

Placement of oxygen equipment

On 30 April 2024, the operator issued a mandatory requirement for all pilots on all flights – including those below flight levels – to secure the onboard oxygen equipment within seated reach of the pilot-in-command. As an additional risk control, the operator also required photographic evidence of this to be forwarded to the Head of Flight Operations (HOFO) prior to departure. 

Use of oxygen equipment

On 2 May 2024, the operator issued guidance to all pilots via Notices to Aircrew (NOTACs) regarding the circumstances in which oxygen should be used in‑flight. The operator recommended that supplementary oxygen be used in the following circumstances:

• discharge of a fire extinguisher in aircraft cabin
• smoke or fumes in cabin
• suspected CO in cabin
• any other occasion where oxygen may assist the health or wellbeing of a crew member. 

Pilots were also advised to follow existing standard operating procedures and conduct a precautionary landing as soon as possible in the event of smoke, fumes, or gases in the cabin. This guidance was also incorporated into the operator’s emergency training modules.

Sources and submissions

Sources of information

The sources of information during the investigation included:

• the operating crew
• Aero Logistics
• Airservices Australia
• Bureau of Meteorology
• recorded data from the ADS-B unit on the aircraft. 

Submissions

Under section 26 of the Transport Safety Investigation Act 2003, the ATSB may provide a draft report, on a confidential basis, to any person whom the ATSB considers appropriate. That section allows a person receiving a draft report to make submissions to the ATSB about the draft report. 

A draft of this report was provided to the following directly involved parties:

• the operating crew
• Aero Logistics
• the Civil Aviation Safety Authority.

No submissions were received.

Purpose of safety investigations The objective of a safety investigation is to enhance transport safety. This is done through:  identifying safety issues and facilitating safety action to address those issues providing information about occurrences and their associated safety factors to facilitate learning within the transport industry. It is not a function of the ATSB to apportion blame or provide a means for determining liability. At the same time, an investigation report must include factual material of sufficient weight to support the analysis and findings. At all times the ATSB endeavours to balance the use of material that could imply adverse comment with the need to properly explain what happened, and why, in a fair and unbiased manner. The ATSB does not investigate for the purpose of taking administrative, regulatory or criminal action. Terminology An explanation of terminology used in ATSB investigation reports is available here. This includes terms such as occurrence, contributing factor, other factor that increased risk, and safety issue. Publishing information Released in accordance with section 25 of the Transport Safety Investigation Act 2003 Published by: Australian Transport Safety Bureau © Commonwealth of Australia 2024 Ownership of intellectual property rights in this publication Unless otherwise noted, copyright (and any other intellectual property rights, if any) in this report publication is owned by the Commonwealth of Australia. Creative Commons licence With the exception of the Coat of Arms, ATSB logo, and photos and graphics in which a third party holds copyright, this publication is licensed under a Creative Commons Attribution 3.0 Australia licence. Creative Commons Attribution 3.0 Australia Licence is a standard form licence agreement that allows you to copy, distribute, transmit and adapt this publication provided that you attribute the work. The ATSB’s preference is that you attribute this publication (and any material sourced from it) using the following wording: Source: Australian Transport Safety Bureau Copyright in material obtained from other agencies, private individuals or organisations, belongs to those agencies, individuals or organisations. Where you wish to use their material, you will need to contact them directly.