A Jetstream 31 aircraft was inbound to Sydney runway 34, on descent from 6,000 ft to 4,000 ft, when severe turbulence was encountered. Whilst being subjected to sustained buffeting, the aircraft experienced uncommanded pitch and roll events. Both pilots responded on the controls, in an attempt to minimise the excursions. At the crew's request, the air traffic controller advised that their separation from a preceding Airbus A300, also tracking for runway 34, was six miles.

Examination of recorded radar information confirmed that the required wake turbulence avoidance separation standard had been maintained. At the time at which the Jetstream 31 had experienced the upset, the A300 was 6.3 NM ahead. However, when closest to the point at which the upset occurred, the A300 had been 600 ft higher than the Jetstream, and the flight path of the A300 had been about 1 NM upwind of that of the Jetstream.

The effect of the prevailing wind, and the descent of the Jetstream to an altitude of less than 1,000 ft below that of the A300, placed the significantly smaller aircraft in the wake turbulence generated by the wide-bodied aircraft.

1. FACTUAL INFORMATION

1.1 History of the flight

The aircraft was the second to take off in a stream of six on a night flight from the Tangalooma Resort strip to Coolangatta aerodrome. Shortly after take-off it struck the ground, nosed over, and was consumed by a fuel-fed fire.

The take-off run appeared normal, but the initial climb was shallow according to the witnesses, some of whom were pilots waiting their turn to take off. At about 150 ft above ground level the aircraft entered a descent which continued until ground impact, 164 m beyond the departure end of the strip. The nose gear collapsed at impact, but the aircraft remained upright and skidded along the ground on its main gear and front fuselage. It traversed a low sand dune, fell 10 ft to the beach and overturned. The aircraft came to rest 112 m beyond the first ground contact. All four passengers were able to evacuate the aircraft which had started to burn. The pilot was rescued by her passengers.

1.2 Weather and visibility

The weather was fine with some cumulus cloud over Moreton Bay to the west of the strip. It was a dark night with only stars visible. The wind was almost calm but slightly favoured a take-off towards the north. The northern take-off path extended over north-eastern Moreton Bay. Except for a possible light from a fishing trawler or house at Cowan township, there was no surface illumination and no discernible horizon. A take-off to the south on the other hand was available with a distinct horizon reference due to an illuminated Tangalooma Resort some 5 km south and the very bright lights of the Brisbane Port facilities at Fisherman Islands some 40 degrees to the right of runway heading.

1.3 Personnel information

1.3.1 Pilot in command

The pilot was a Grade 2 instructor, employed by the flying school where she learnt to fly. She had a current command multi-engine instrument rating and was endorsed to fly the aircraft type. The pilot was required to wear spectacles when flying and she did on this occasion.

On 14 September 1995, the pilot was checked on the route by the chief pilot. The route check included a night take-off at Tangalooma strip. Her most recent flights in the aircraft were on 15 October and 2 November 1996. Both flights were return flights between Coolangatta and Tangalooma, with a night take-off from the Tangalooma strip. The night take-off on 2 November was conducted towards the south.

1.3.2 Previous 72-hours history

The weather on the two days before the accident precluded any flying duties. During this period, the pilot worked at the flying school office and brought her book-keeping duties up to date. On the two nights prior to the accident, she had retired to bed at 2030 hours and 2130 hours local time, respectively. Prior to then she had been off duty for 9 days. On the day of the accident, the pilot commenced duties at 1400 hours, planning the afternoon flight to Tangalooma and the night return. Between her arrival at Tangalooma and the night take-off, she rested and had dinner at the resort.

1.3.3 Information from the pilot

The pilot was hospitalised with serious leg injuries. She had also suffered a blow to the head. She was interviewed in hospital and said that she could not remember the take-off and subsequent flight into terrain. The pilot said that the aircraft carried a printed checklist on a single sheet of paper which her employer required company pilots to use in normal aircraft operations. However, for reasons not established, the pilot did not use the checklist after landing at the island strip or during the subsequent night departure.

1.4 Passenger action

The four passengers were foreign nationals and were seated in rows 2 and 3. They were not injured in the crash and were able to crawl out through the open cabin doorway. A fuel-fed fire had started almost immediately, and the aircraft was well alight by the time they cleared the aircraft. They soon realised that the pilot was still in the cockpit. Some of the passengers crawled back inside and pulled the semi-conscious pilot clear of the aircraft. Other people arrived on the scene soon afterwards and moved the passengers and pilot further away from the burning aircraft.

1.5 Wreckage examination

The aircraft had come to rest upside down on a beach. The front of the cockpit was crushed during the nose-over. The cabin was completely destroyed by a post-impact fire which also damaged the right engine.

Examination of the wreckage found that the flaps were retracted and that the elevator trim was set about halfway between the take-off setting and the fully forward position. Both engines were removed for specialist bulk strip examination. This examination did not discover any defect which could have prevented the engines from developing rated power. This evidence supports witness information that the engines' note did not change during the entire flight sequence. Due to the high wing configuration, neither propeller contacted the ground until the aircraft nosed over. Examination showed significant torque twisting and bending to both propellers indicating that the engines were still developing significant power when the aircraft nosed over.

1.6 Aircraft flight characteristics and flight test

1.6.1 Flap retraction

Discussions with pilots experienced on the aircraft type indicated that during initial climb, the take-off flap must be retracted in stages, trimming the elevator at the same time. Their opinion was that if the flap was retracted in one movement instead of stages, the nose-down trim change would be significant.  This would result in a lower nose attitude unless the pilot held back pressure on the control column.

1.6.2 Flight test

At the Bureau's request a flight in a Partenavia was undertaken by a Grade 1 instructor with the view to establishing elevator control forces under specific conditions. The aircraft was flown on a final approach with full flap selected. The elevator was trimmed so that there was no residual force on the elevator control. This resulted in a trim indication halfway between neutral and a fully nose-down position. A take-off was conducted with the elevator trim and flap in the previous position. The pilot reported that a medium-to-heavy rearward elevator force was required to rotate the aircraft and establish a positive rate of climb. At a safe height, the flaps were retracted. This procedure was accomplished twice with the following result:

1. If the indicated airspeed and rate of climb were maintained the elevator force increased to the point of being very heavy.
2. If the elevator force was not increased, the aircraft settled into a 400 ft per minute rate of descent.

The pilot conducting this test commented that unless a pilot was prepared, the high elevator force required to maintain a positive rate of climb was excessive and well outside the normal experience range of an average pilot.

1.7 Dark night take-off research

Bureau research has shown that dark night take-offs can present a number of special problems for pilots, particularly when departing from aerodromes in areas with little or no ground lighting. Specifically, the somatogravic illusion has been suggested as being a contributing factor in the majority of these accidents. A somatogravic or false climb illusion occurs when a pilot who is deprived of outside visual cues attempts to maintain a desired pitch attitude without reference to instruments whilst the aircraft is accelerating. Under such conditions, pilots can experience a sensation of excessive pitch-up. The sensation is thought to exist primarily at take-off, go-around and during visual flight into cloud. At take-off, pilots who attempt to correct for this sensation by relaxing the backward pressure on the elevator control, may fly the aircraft into the ground.

2. ANALYSIS

2.1 The checklist

The pilot said that she did not use the checklist following the daylight landing at the Tangalooma strip, nor did she use it prior to the night take-off. The pilot could not give an explanation for this omission. A checklist written on loose paper can be difficult to handle during busy periods of operation such as circuit work or in low-light conditions, specifically at night. This could explain the lack of use in this case. If a checklist is not used, the pilot must commit all checks to memory.

2.2 Elevator trim

The elevator trim was found in a position consistent with the trimmed position for a full-flap landing. It is highly likely that the pilot missed the trim item off her memorised checklist both following the daylight landing and during pre-take-off checks at night.

2.3 Evidence from the test flight

The test flight pilot reported that the elevator force induced by an incorrectly set elevator trim would overpower a pilot who was not prepared for this predicament. The shallow take-off witnessed by others was probably the result of the pilot encountering such an unexpected down-elevator force. The pilot probably retracted the take-off flap in one movement rather than in a staged retraction since she would have needed both hands on the control column. Such an action could only have exacerbated the downward force on the elevator.

2.4 Dark night take-off

In addition, the pilot was faced with a dark night take-off. Unless the pilot's attention was focussed almost solely on the correct climb attitude, it was likely that the nose attitude would be lowered inadvertently. Because of the unexpected elevator force, she was probably distracted from her proper instrument scan. Under these circumstances, an illusion of a false climb would have resulted in her not maintaining the high load needed on the control column. This would have resulted in the aircraft descending back onto the ground. That this came as a totally unexpected event is borne out by the lack of any action in correcting the aircraft attitude, reducing descent, or reducing power.

The collective decision by the pilots to conduct take-off operations towards the north was unwise, considering that a useful visual horizon was available for take-offs towards the south.

3. CONCLUSIONS

3.1 Findings

1. The pilot was fully qualified to undertake the flight.
2. The night was dark and the area beyond the departure end of the runway was devoid of any useful lighting.
3. A southerly take-off direction was available, with a visible horizon.
4. The pilot did not make use of the written checklist carried in the aircraft.
5. The elevator trim was set halfway between neutral and fully forward.
6. The flaps were fully retracted at impact.
7. The aircraft impacted wings-level and in a shallow descent.
8. Full power was maintained throughout the take-off and accident sequence.

3.2 Significant factors

1. The take-off direction was dark and had no visible horizon.
2. The elevator trim was not set for take-off.
3. The elevator load on take-off was high.
4. The pilot did not monitor the aircraft attitude after lift-off.
5. The flap was retracted in one movement, increasing the elevator load.
6. The pilot may have been affected by somatogravic illusion to the extent that she thought the climb attitude was adequate.

Factual Information

VH-MWG had departed Armidale for Wollongong and was maintaining 10,000 ft outside controlled airspace.  The crew reported at Armidale east at 1244 ESuT, with an estimate for Mount Sandon at 1303.  The Flight Service Officer (FSO) passed conflicting traffic information on VH-DMI to the crew of MWG.

DMI had departed Sydney for Armidale and was maintaining flight level (FL) 190 in controlled airspace.  The crew reported their position to sector control as passing West Maitland at 1228, with an estimate for Armidale of 1258.

At 1233, the sector controller passed the DMI position report to the FSO and coordinated the frequency transfer as vacating FL 190.  The FSO considered that, as DMI would need to descend 9,000 ft prior to leaving controlled airspace, there was sufficient time to pass traffic information on MWG after the crew of DMI had transferred to her frequency.

At 1247, the crew of DMI requested descent into Armidale and were cleared to leave control area on descent by the sector controller.  The crew commenced descent a few seconds later but did not report leaving FL 190.

Although the altitude readout for DMI was displaying correctly on the radar screen, the controller did not notice that DMI had left FL 190 and did not instruct the crew to change to the flight service frequency.

At 1251, the FSO contacted sector control to find out the actual position of DMI. She had realised that the time by which she had expected the crew to call had expired and the traffic information on MWG had still not been passed.

The controller then looked at the radar screen specifically to fins DMI and noticed that the aircraft was descending through FL120.  He also noticed a conflicting aircraft 4 NM ahead and on a reciprocal track and informed the FSO.

The FSO confirmed that the conflicting aircraft was the one she wanted to pass as traffic to the crew of DMI. 

During this discussion, the sector controller instructed the crew of DMI to transfer to flight service on 122.1 Mhz. He did not pass the traffic information at that time, but did pass traffic a few seconds later.  However, the crew of DMI had already changed frequency prior to those instructions.

As they approached 11,000 ft, the crew of DMI realised that they needed to contact flight service prior to leaving controlled airspace and initiated a call to sector control.  This call was made at the same time as the sector controller had initiated his call to DMI and consequently neither party heard the other transmit.

As they passed 10,800 ft, the crew of DMI sighted MWG and adjusted their flight path to pass behind that aircraft.

Radar analysis indicates that the aircraft passed by 500m horizontally, with no vertical separation.

Analysis

The sector controller normally waited for a pilot to report leaving a level before instructing the crew to change frequency.  In this case, the crew did not make this required report and the reminder to the controller was lost.  At the same time, the controller became occupied with other tasks and did not notice that DMI was on descent.

The crew of DMI were required to leave FL 190 within one minute of the instruction from sector control and to report leaving that level.  The controller is required to check that the aircraft has commenced its descent if no such report is received.  This check may be carried out by reference to radar information.

The FSO had correctly calculated the confliction between DMI and MWG and needed to pass traffic information to each crew.  However, she was cognisant of the historical adversity that air traffic controllers had to passing such information on control frequencies.  She considered that 9,000 ft of descent gave sufficient time to allow the crew of DMI to contact her, receive the traffic advice and make any operational decision.  Therefore, the FSO decided not to ask the sector controller to pass the traffic information on MWG to the crew of DMI.

The FSO had only recently obtained her rating on the FIS 3 circuit and, although instructions on the passing of traffic information to aircrew via air traffic control existed, she was not familiar with them and had not had them made available to her during training.  These instructions were contained in the Temporary Local Instructions book which was not readily available for officer reference.

Findings

1. The Temporary Local Instructions book was not readily available for FSO reference.
2. The training of FSOs at Sydney did not adequately cover temporary local instructions in relation to the passing of traffic information.
3. The FSO correctly calculated the need for traffic information to be passed to the crews of MWG and DMI.
4. The FSO did not pass traffic information to the crew of DMI at the earliest possible opportunity.
5. The crew of DMI did not report leaving FL 190.
6. The sector controller did not check whether the crew of DMI had commenced descent after no report had been received within one minute of issuing the clearance.
7. The sector controller did not notice that DMI was on descent.
8. The sector controller did not instruct the crew of DMI to transfer frequency to flight service at the coordinated time.
9. The crew of DMI sighted MWG and adjusted their flight path to ensure separation from that aircraft.

Significant Factors

1. Traffic information on MWG was not passed to the crew of DMI at the earliest possible time.
2. The crew of DMI were not transferred to the flight service frequency at the appropriate time.

Safety Action

As a result of the investigation, Sydney flight service management have:

1. Re-issued instructions on the passing of traffic information through air traffic control;
2. Updated the Temporary Local Instructions book and relocated it in a readily available position;
3. Instructed training officers to ensure that all trainees are instructed in all matters included in temporary local instructions; and
4. Introduced new sign-on rules that require all FSOs to read temporary local instructions prior to commencing duty each shift.

The B737 aircraft was being radar vectored for an approach to Sydney runway 34. The Approach South controller issued instructions for the crew to expect runway 34 Right and proceeded to transfer the aircraft to the Director for final approach.

The Director vectored the B737 onto base leg and then onto final approach with an instruction to intercept final for runway 34 Left. This runway expectancy was immediately questioned by the aircraft captain as he had briefed his crew for an approach to runway 34 Right.

The Director confirmed that the runway would be 34 Left and commented that he thought the crew had already been given that information. The captain replied that he had not, but that the change to 34 Left could be safely carried out. The aircraft continued its approach and landed on runway 34 Left without further incident.

It was determined that the coordination for the change of runways was carried out between the controllers concerned, but the exact manner in which this was performed was not established. The controllers did discuss the possibility of changing aircraft between the two runways in general terms, but there was no specific reference to this aircraft on the voice recordings. The flight progress strips, which may have provided further insight into this aspect of the occurrence, were not located during the investigation and were unable to be scrutinised.

No definite reason for the failure to advise the crew of the change of runway could be determined.

FACTUAL INFORMATION

The Piper Navajo Chieftan (PA-31) was inbound to Perth from the south and was descending to 4,000 ft. The approach west controller (AppW) initially intended to put the aircraft onto a right downwind for runway 21; however, due to traffic near the city, he co-ordinated with the approach east controller (AppE) for the aircraft to track via a left downwind. This was accepted by AppE and the pilot was transferred to his frequency at 1213 WST.

The Boeing 767 (B767) was taxiing for runway 21 for a departure to the west. At 1220, AppW advised the aerodrome controller that the departure clearance for the B767 was to maintain runway heading and to maintain 6,000 ft. This clearance was issued to the crew at 1221, when the B767 was cleared for take-off.

At the same time, and with the aim of expediting its landing, AppE arranged with AppW for the PA31 to change to a right circuit for runway 21. The aircraft was still south of the airport when AppE advised the pilot of this change and instructed him to turn left onto a heading of 340 degrees and to descend to 2,500 ft.

At 1223, the aerodrome controller contacted AppE and asked what the intention was with the PA31. Shortly after, AppE noticed an aircraft (the B767) departing and asked AppW what instructions the crew had received. Almost simultaneously, the aerodrome controller contacted AppE and told him to turn the PA-31. AppE immediately instructed the PA-31 pilot to turn right onto a heading of 070 degrees and asked if the pilot had the B767 in sight. The pilot replied that he did.

AppE then asked the tower to maintain the B767 at 2,000 ft. The aerodrome controller instructed the crew of B767 crew to turn right onto a heading of 240 degrees and then informed AppE that he could provide visual separation. The two aircraft passed with approximately 1 NM lateral separation and 300 ft vertical separation.

ANALYSIS

Approach/Departure controllers record information on flight progress strips to assist in the processing of aircraft. AppE and AppW sit side by side and use a common central strip holding bay for the display of active strips. A feature of this display is a "next" designator which houses the current arrival/departure sequence. When busy, this bay can become overcrowded with strips, and they are often placed elsewhere on the console for convenience. AppW could not recall whether the strip for the B767 had been placed under the "next" designator in the central bay. AppW omitted to take the B767 into consideration when he gave approval to AppE for the PA-31 to track for a right downwind for runway 21.

AppE was unsure of the position of the strip for the B767 or whether he had looked at the "next" designator bay prior to co-ordinating a change of circuit direction for the PA-31. The alteration of the track of the PA-31 resulted in the tracking of the aircraft across the departure track of aircraft taking off from runway 21.

SIGNIFICANT FACTORS

1. The procedures adopted by the approach east controller resulted in the PA-31 being directed to track across the departure path of aircraft departing from runway 21.
2. The approach west controller overlooked the impending departure of the B767 from runway 21 when he approved the request from the approach east controller to track the PA-31 for a right downwind for runway 21.
3. The central strip holding bay on the approach console was overcrowded and did not readily indicate all aircraft pertinent to the traffic situation.

SAFETY ACTION

As a result of the investigation, several safety actions were initiated by both Airservices Australia and the Bureau.

Airservices Australia

Instrument departure procedures for Jandakot aerodrome have been published which have minimised the variation of random departure paths.

Bureau of Air Safety Investigation

The Bureau issued interim recommendation IR960044 on 14 May 1996.

"IR960044

"The Bureau of Air Safety Investigation recommends that Airservices Australia conduct a review of Perth terminal airspace and procedures with a view to reducing the number of conflicting flight paths and the amount of intra-unit co-ordination. Such a review should include:

-jurisdiction over the runway heading departure track from each runway;

-airspace division between the approach/departure control positions;

-improved SID/STAR interface; and

-the introduction of standardised holding patterns."

Airservices Australia replied on 16 June 1996 as follows:

"I refer to your Air Safety Recommendation No. IR960044 involving VH-NMP and V8-RBH at Perth and concur with the recommendations.

"Your Interim Recommendations to review aspects of Perth Terminal Airspace and Procedures have generally been addressed. Ongoing planning of the revised airspace and procedures under ROTAP (Restructure Of Terminal Area Perth) will reflect other recommendations provided they are predicted to be compatible with forthcoming and future developments of the ROTAP structure requirements.

"The introduction of additional outer holding patterns is under consideration. Prior to any selection of holding points, a full assessment of any implications as to their placement will need to be made to ensure compatibility with the future airspace revisions.

"SID and STAR changes have been actioned under ROTAP, an Environment Impact Statement has been submitted, and implementation is planned for December 1996. This will provide an improved interaction of Arrival and Departure traffic at Perth.

"Temporary Local Instruction (TLI) WADO/95/023, issued 20 Dec 1995, redefines the areas of responsibility.  A more comprehensive documentation of procedures and airspace divisions will be incorporated progressively with ROTAP amendments.  The TLI addresses:

• enhancements to jurisdiction over the runway heading departure tracks from each runway;
• enhancements to airspace division between Approach East and Approach West, and Departures Control positions;
• Procedure, Co-ordination, and flight Progress Strip annotation requirements."This response was classified as CLOSED - ACCEPTED.

Other aspects of the investigation involving the provision of radar advisory services by approach/departure control are still being researched by the Bureau and any safety outputs from this research will be published in the Quarterly Safety Deficiency Report.

On take-off the left engine exhaust gas temperature exceeded the maximum limit. The power was subsequently reduced to idle, and the aircraft was manoeuvred for a return to Melbourne. During the manoeuvring the turbine vibration increased and when upper limits were exceeded the engine was shut down. The aircraft landed with the airport emergency services on standby.

The high engine exhaust gas temperature reading was found to be a false indication which was traced to a high resistance within the exhaust gas harness assembly probably caused by moisture ingress in connectors. The high vibration reading had been investigated without determining the cause. The engine's vibration was being monitored by the operator who advised that since this incident the engine vibration readings have returned to normal.

Sequence of Events

As part of the take-off checklist, the pilot checked both magnetos and the carburettor heat control for function. Following take-off and climb to about 1,000ft, the engine commenced to run rough with large drops in rpm. The pilot decided to return to the aerodrome but during the return the engine suffered a total power loss. He was unable to restore power by completing the trouble checks which included the application of carburettor heat and ensuring a fuel supply with the fuel selector on both tanks.

During the forced landing, the pilot landed with a 10-knot tail wind more than halfway along the strip. He was unable to stop the aircraft from overrunning the end of the strip. The aircraft was damaged by foliage.

Afterwards the engine's fuel and air supply were checked by a maintenance organisation and found to be serviceable. Both magnetos were benchchecked and functioned normally. The Bureau of Meteorology indicated that at the time of the accident the relative humidity was 100% with a temperature of 25 degrees.

Analysis

The aircraft was being operated in an environment where serious icing at descent power was probable. Icing at higher power settings was also possible.

During the return to the aerodrome, the pilot misjudged the forced landing approach, and the aircraft overran the end of the strip

The aircraft operator said that the aircraft struck the ground during a turn when the pilot was involved in a crop spraying operation.

The pilot has declined to forward any information.

The pilot with three passengers intended operating the Warrior in the training area for about one and half hours. During the take-off roll, the pilot became concerned with the apparent lack of aircraft acceleration. Immediately after liftoff, the aircraft settled and, suspecting a reduction of power, the pilot closed the throttle. During the landing the aircraft overran the end of the runway sustaining minor damage. A subsequent engineering inspection found that the engine was capable of normal operation.

The pilot reported that the flight was required to conduct an urgent medical evacuation. He was advised that the strip had been recently graded and was acceptable. The strip owner was not contacted to ascertain its current status. The strip appeared satisfactory during a pre-landing inspection.

During the latter stage of the landing roll, on runway 20, the nosewheel entered an area of soft dry soil. The additional drag from the soil caused the nosewheel landing gear extension strut to fail in overload and the landing gear to collapse.

Although the whole strip had been graded only the first 750 m was suitable for aircraft operations. The remaining 550 m was covered with soft soil that had been washed onto the strip during recent rains. The soil had been levelled but not removed during the grading. There were no markers on the strip to indicate the unserviceable area.