Circumstances

On the afternoon of the accident, the chief pilot was to undergo a 'check pilot' approval flight for company aircraft with a Civil Aviation Safety Authority Flight Operations Inspector (FOI). The chief pilot was initially required to demonstrate his competence to fly from the right seat of the company's De Havilland Beaver, including both into-wind and crosswind circuits. Following the completion of this demonstration, the chief pilot was to hand over control of the aircraft to the FOI. The FOI was to fly the aircraft under the supervision of the chief pilot.

The FOI was endorsed on the Beaver with a total time on type of three hours including one hour as pilot-in-command.

Weather information for the flights had been obtained by Affix. There was a light northerly breeze at approximately 10 kts. In the area where crosswind circuits were to be conducted, the crosswind strength was approximately 8 kts with gusts of up to an additional 5 kts. Surface conditions on the lake at the time of the accident were described as a 'moderate ripple'.

The chief pilot completed a number of circuits. The FOI was satisfied with this aspect of the approval flight. Control of the aircraft was then handed over to the FOI.

The FOI reported that the into-wind circuits he had flown had been uneventful, including a practice forced landing. The chief pilot added that, although the landing technique had been satisfactory, the landing nose attitude had been a little lower than would be considered ideal. He had discussed this with the FOI during these circuits.

The FOI then requested to fly a crosswind landing and take-off. A short discussion followed in which the FOI mentioned that most of his training on 'floats' had been carried out in 'light' wind conditions.

The FOI stated that, as he manoeuvred the aircraft for the crosswind approach and landing, the crosswind was still from the right, but it had moderated a little from earlier observations.

As directed by the chief pilot, the FOI adopted a level nose attitude when the aircraft was close to the water and placed the right wing down to stop the 'drift'. At this time the power was reduced. The right float touched the water first, as expected for a crosswind landing, but the aircraft then 'skipped' slightly.  The second time it touched down the float dug into the water and the aircraft yawed to the right.  At this point, the FOI stated that he applied full left rudder to counteract the yaw. The chief pilot also stated that he took hold of the control column and applied full backpressure to assist with the recovery actions of the FOI.

The aircraft did not respond to these corrective measures.  It proceeded to lurch violently left in what was described as a cart-wheeling motion.

The aircraft came to rest, right side up, facing the opposite direction to that of the intended landing, and with the right wing down approximately 20-30 degrees.

The aircraft rapidly filled with water from the right side. The FOI attempted to open his door on the left side, but it appeared to be jammed by a float, as were all four doors. The chief pilot tried to assist the FOI but without success. While the FOI climbed into the rear seat to attempt to open the rear left door, the chief pilot attempted to exit from the window of the left front door. The FOI used a paddle to assist in opening his door but without results. As the aircraft started to sink the float moved away from the cabin area sufficiently to allow the FOI to evacuate through the left rear door. The chief pilot pulled himself through the left front window as the fuselage was starting to submerge.

The two pilots clung to the aircraft floats to await rescue. Both pilots sustained minor cuts and bruises.

The aircraft sustained major structural damage. Both wings were torn from the wing roots. It was later recovered from the lake.

Analysis

The chief pilot was experienced on the aircraft type but, considering he was in the process of an approval flight at the time of the accident, his experience as a 'check pilot' was limited.

The chief pilot stated that he felt the FOI was using a satisfactory technique to land the aircraft, albeit with a slightly 'flat' nose attitude. He was sufficiently confident in the FOI's ability to safely fly the aircraft that he did not feel it necessary to have his hands placed close to the control column. He pointed out, however, on this landing, the aircraft seemed to 'settle' at a greater rate than he had expected. He considered that, when the Beaver first touched down and 'skipped', the FOI was capable of recovering without assistance. It was not until the aircraft 'dug in' on the second touchdown that he felt it necessary to assist. In retrospect, the chief pilot thought that this action was a little too late to affect the recovery.

The FOI had limited experience in float plane operations. It had been some time since his 'type' endorsement on the Beaver had been completed. He also stated that his flights in the Beaver had all been conducted in 'light' wind conditions. The crosswind on the day of the accident was considered to be greater than he had previously experienced.

Findings

1. The FOI had limited experience in float plane operations.
2. The FOI had limited experience on the aircraft type in crosswind conditions.
3. The chief pilot had limited experience as a check pilot.
4. The FOI lost directional control of the aircraft during a crosswind landing.
5. Both pilots experienced difficulty evacuating the aircraft.

The pilot reported that the aircraft was landed off a normal full flap approach. The strip slopped down from the touchdown point. After lowering the nosewheel the pilot applied the brakes; however, they did not work, and the aircraft's speed did not reduce. He shouted to the passengers to brace themselves and, as the end of the strip was approaching rapidly, he attempted to steer the aircraft into a fence to bring it to a stop. There was insufficient room to complete a go-around. Shortly after turning the aircraft towards the fence the brakes operated, locking the main wheels. The pilot was unable to establish any further directional control, and the aircraft collided with a tree.

A post-accident inspection of the braking system did not disclose any reason for the brake problem.

The pilot was ferrying the aircraft to Killara Station for the owner. He had checked the condition of the strip prior to departure and was advised to land on a road near the homestead as the main strip was unserviceable.

The pilot reported that on arrival he made two low approaches to confirm that the road was suitable and then made a third approach for the landing. There was no windsock visible; however, a windmill indicated that the wind was a crosswind from the right. At about 25 ft on final approach the aircraft started to drift away from the centreline of the road. The pilot was unhappy with the approach and decided to go-around. He opened the throttle; however, before the engine could respond the aircraft descended rapidly and landed heavily on the edge of the road. The aircraft ran off the road into trees, the nosewheel dug in and it overturned.

It was reported that there were no unusual atmospheric conditions present at the time of the accident. It is probable that the sudden descent occurred because the pilot inadvertently allowed the airspeed to reduce to the stall speed during the attempted go-around.

Circumstances

VH-OJJ was cleared to track inbound via the 229 VOR radial, while maintaining 6,000 ft. The Approach South radar control sector controller (APP S) was processing the aircraft for landing on runway 16R. Due to operational requirements, the APP S controller delayed turning the aircraft onto a right downwind for that runway.

VH-TCH had departed runway 16R, cleared initially to climb to 5,000 ft. The Departures South radar control sector controller (DEP S) cleared the aircraft for an early right turn and unrestricted climb to its planned cruising level. The aircraft was then to be vectored left to intercept the 207 VOR radial.

A short time later, as the APP S controller was about to instruct the crew of VH-OJJ to turn downwind and VH-TCH was turning left to intercept the 207 VOR radial, the crew of VH-OJJ advised that they had received a TCAS Resolution Advisory (RA) alert.

Recorded radar information indicated that each aircraft had penetrated its respective airspace boundary buffer, and closed to within 2.5 NM while both were at 6,000 ft.

The runway 16 airspace agreement has a narrow corridor in the APP S airspace to the west of the airport. This airspace design limits the flexibility available to APP S controllers when vectoring aircraft to downwind for runway 16. The latitude that the DEP S controller may apply when turning departing aircraft onto south-westerly tracks, and providing climb, is also restricted.

Separation was not ensured before VH-TCH was cleared for an unrestricted climb. The situation was possibly worsened by the southerly wind which would have drifted VH-TCH closer to the APP S airspace boundary.

Findings

1. The unrestricted climb instruction provided to VH-TCH by the DEP S controller was premature.
2. The vectoring of VH-OJJ onto downwind for runway 16 by the APP S controller was initiated late.
3. The airspace agreement boundary for runway 16 operations did not afford adequate separation assurance.

Factors

1. The radar surveillance techniques used by both controllers were inadequate to prevent the loss of the prescribed separation standards.

Safety action

As a result of the investigation, the Bureau of Air Safety Investigation made the following interim recommendation.

IR950212

The Bureau of Air Safety Investigation recommends that Airservices Australia amend the runway 16 airspace agreement boundaries to ensure that arriving and departing aircraft cannot be on, or near, reciprocal tracks. The amendment should address excursions into either buffer area.

Airservices Australia responded to this recommendation as follows:

"A Standard Instrument Departure (SID) has been developed for non-jet southern departures which will provide greater separation assurance and reduce controller workload. At this stage we expect the new SID to be included in DAP EAST Amendment List 52.

I should, however, point out that there remains a requirement for air traffic control to apply vertical separation between departing aircraft and arriving aircraft until the departing aircraft is clear of approach airspace. Vertical separation is applied at the clearance delivery phase of flight. Jet aircraft are required to maintain 5,000' and propeller driven aircraft are required to maintain 3,000'. The need for crossovers between arriving and departing aircraft within the Sydney TMA is unavoidable."

VH-EAA was processed into the Sydney Terminal Area by the Approach south radar control sector, and was tracking via the 094 VOR radial, on descent to 6,000 ft in visual conditions. At 9 DME, the flight crew, in response to a TCAS RA alert, climbed the aircraft approximately 300 ft. They did not advise ATC of the alert.

VH-NEK had departed runway 16L, on initial climb to 5,000 ft, and was being processed to depart the Sydney Terminal Area by the Departures south radar control sector. The aircraft was vectored onto a north-easterly track from a left turn after take-off.

Recorded radar information indicated that VH-EAA had continued descent to a pressure altitude of 5,700 ft and VH-NEK had climbed to a pressure altitude of 4,900 ft. When their flight paths crossed, VH-EAA was 800 ft above, and 1.86 NM to the right of VH-NEK.

The traffic management plan, techniques and procedures, for processing both aircraft were adequate to maintain prescribed separation standards and tolerances in accordance with the runway 16 airspace agreement. However, neither controller had recognised the incident, probably due to the short duration of the encounter, and the lack of advice from either flight crew.

Findings

1. The traffic management procedures adopted by the controllers for processing the aircraft were adequate.

2. VH-EAA descended below the assigned altitude.

3. In response to the TCAS RA alert, the flight crew returned the aircraft to the assigned altitude.

Significant factors

1. The required separation standard was infringed when the flight crew permitted VH-EAA to descend below their assigned altitude.

As VH-NJR was approaching Perth from the east APP (E) was responsible for initial radar vectoring. To improve aircraft handling APP (W) agreed that APP (E) would maintain vectoring after the aircraft crossed into APP (W)'s airspace. APP (W) also indicated that he would keep all his other traffic clear of VH-NJR.

The pilot of VH-MPZ was cleared to depart Jandakot’s runway 06 in a left turn with a climb to 2,000 ft. He read back his allocated transponder code as 4503 instead of 4502 and the error was not detected despite a request to repeat his read-back. VH-MPZ's clearance for a departure from Jandakot meant that separation between VH-NJR (initially cleared to descend to 2,500 ft) and VH-MPZ (cleared to climb to 2,000 ft) was not assured. APP (W) did not realise this error, nor did he coordinate VH-MPZ's departure with APP (E) and as a result APP (E) was not warned there might be a traffic conflict. APP (W) had placed VH-MPZ's flight strip in the bay between the two controllers and had intended drawing APP (E)'s attention to it. APP (E) was busy, and APP (W) decided to leave it until later. He was then distracted by communications with another aircraft and forgot about it.

When VH-MPZ departed Jandakot its transponder code was set at 4503. A code of 4502 had been programmed into the radar's computer; consequently, when the return appeared on both APP (W) and APP (E)'s screens it did not have a label indicating the aircraft's callsign. It appears both controllers initially assumed that the un-labled return was circuit traffic at Jandakot and paid little attention to it. VH-MPZ exited the Jandakot control zone at the same time as APP (E) was giving further directions to VH-NJR to turn right and continue on a visual descent. Because APP (E) had the quick-look facility selected on her radar she observed that the traffic, which was now heading towards VH-NJR, was at 2,000 ft and would be in conflict. She immediately turned VH-NJR further right and assigned the aircraft 2,500 ft.  This was to provide emergency separation of 500 ft between the aircraft. VH-NJR descended to 2,300 ft before returning to 2,500 ft.

Analysis

The factors leading to the occurrence started when APP (W) cleared the pilot of VH-MPZ to depart Jandakot at 2,000 ft on a track that would take the aircraft towards the flight path of VH-NJR which had been cleared to descend to 2,500 ft. The separation requirement were not assured by this instruction. APP (W) should have cleared VH-MPZ via a different route or restricted its climb to 1,500 ft. APP (W) could also have coordinated a descent limit of 3,000 ft for VH-NJR with APP (E), to provide the required assurance. A lack of commitment to the principal of separation assurance was identified as a possible factor in APP (W)'s under-performance.

Having started the process APP (W) did not ensure that APP (E) was aware of the possible traffic by drawing her attention to the flight progress strip located in the bay between the controllers. Distraction probably played a part in this failure. APP (E) was operating on the assumption that APP (W) was going to keep all other traffic clear and was not expecting any conflicting traffic and therefore, was not looking for it.

All the visual cues that may have been provided to both APP (W) and APP (E) by their radar screens were not there because VH-MPZ's pilot did not have the correct transponder code set. In addition, APP (W) did not have the quick-look facility selected and was therefore, not aware, at least, of the un-labled returns altitude. It was APP (E)'s use of the quick-look facility which alerted her to the approaching confliction. Use of the quick-look facility is not mandated because of the radar return clutter that results from the circuit traffic at Jandakot, when it is selected.

Safety Action

Airservices Australia's Western Australian office is reviewing the airspace layout around Perth and Jandakot, Jandakot departure procedures and coordination requirements, the use of the radar quick-look facility and methods of reducing radar clutter in the Jandakot area. They will also continue to emphasise, to their staff, the mandatory use of defensive control techniques that are designed to provide separation assurance in all appropriate circumstances.

FACTUAL INFORMATION

History of the flight

While the aircraft was in cruise at FL350, approaching waypoint MILUV, the flight crew was cleared by Brisbane sector control to track inbound to Sydney via the CHEZA 2 Standard Arrival Route (STAR). The controller advised the crew that the aircraft was radar identified and to expect runway 34R (right) for landing. Both pilots reviewed the approach information, which was then programmed, including reference to runway 34R, into the aircraft's flight management computer.

The aircraft was subsequently cleared to descend to 8,000 ft and at about 35 NM east of Sydney the crew was instructed to contact Sydney Approach (north) on 124.4 MHz. On initial contact with Approach the crew was cleared to descend to 5,000 ft and was advised, “runway 34 left, localiser frequency 110.1, independent approach". In accordance with current procedures the crew acknowledged "5,000". Soon after, the controller cancelled the requirement to approach via the Cheza STAR and issued a vector for the approach.

When establishing communications with the Director on 125.3, the crew advised that they were visual with the runway in sight. They were issued further descent instructions and a vector to intercept the localiser final approach. At 12 NM from touchdown, they were instructed to join the centreline for runway left, make a visual approach and contact Sydney Tower on 120.5. The tower frequency was then confirmed at the request of the crew. Shortly after, the Director advised the tower that the aircraft had intercepted the final approach for runway 34R. The controllers resolved that, as there was no potential conflict with other aircraft, VH-TAI could continue its approach to runway right.

On initial contact with the tower the crew reported that they were on a six-mile final approach for runway 34R and were subsequently cleared to land on that runway.

Flight crew information

The pilot in command was experienced in domestic airline operations into Sydney. This was the return leg of his first international flight. Both pilots had reviewed the required pre-flight briefing material, including audio-visual and printed route and terminal area information. The co-pilot was experienced, both in operations into Sydney and on international operations.

When informed that they had conducted an approach for the wrong runway, neither crew could recall being advised that they were to land on runway 34L. Both pilots were aware that the advice to "expect" runway 34R could be subject to change.

Communications

There were no identified technical deficiencies in any communications with the aircraft, nor was there evidence of over-transmitting or other interference with the broadcast or reception of critical information.

Parallel runway operations at Sydney

With the introduction of parallel runway operations, an agreement was established between Sydney and Brisbane Air Traffic Services to provide early advice of the landing runway at Sydney to all international arrivals from the north and the east. This agreement required that Brisbane Control advise aircraft of B767/A300 size and smaller to expect to land on runway 16L/34R and for larger aircraft to expect to land on runway 16R/34L.

The advice to flight crews of a landing runway, different from that for which an expectation had been provided, was not seen by the management of the Sydney district office of Airservices Australia to constitute a change of runway.

The processing of VH-TAI for runway 34L reflected the requirement for the aircraft to taxi to the international terminal and the lack of potential traffic conflictions.

When they were advised of runway 34L at both 35 NM and 12 NM, there was no requirement for the flight crew to read back the runway identifier. Current parallel runway operational procedures required only that pilots confirm "visual" and/or runway "left/right in sight" on first contact with the Director, or as soon as each is the case. When the crew of VH-TAI first contacted the Director at about 28 NM from touchdown, they reported visual, with the runway sighted. They did not identify the runway on which they were intending to land, nor did the Director request runway identification.

The investigation of this occurrence identified safety deficiencies in procedures relating to the operation of parallel runways at Sydney. Current operational documentation contain areas of ambiguity between standard operating procedures and instructions published in the AIP, AIP supplement H36/95, the Manual of Air Traffic Services and the Sydney Terminal Area Temporary Local Instructions. These ambiguities include the use in the various documents of the terms "expected", "nominated" and "specified" in reference to landing runways.

ANALYSIS

The published standard operational procedures and instructions did not address the implications of providing a runway expectation, which was subject to change, without the provision of either change alert or read back requirements. Consequently, the significance of the change to the runway expectation was apparently not recognised by those involved. To contend that the advice of 34L as the landing runway did not constitute a runway change is to discount the effect of the provided expectation. It would then be reasonable to question the value of providing an early expectation of the landing runway beyond the orientation of either runway 16 or 34.

The Director and the tower controller were unaware until late in the approach that the crew was not intending to land on runway 34L. This was due, in part, to the lack of a requirement for the crew to read back the runway identifier when first advised by approach control. The omission by the crew and the Director to confirm the runway identifier when the crew reported having the runway in sight also delayed recognition of the error.

Neither pilot recognised the advice of the nomination of runway 34L for their arrival, although it had been given twice and included confirmation of the localiser and tower frequencies. This is probably a consequence of the lack of a specific alert of the change to the runway expectation without a requirement for the flight crew to read back the runway designator. The inclusion of the runway designator when the crew entered the arrival information in the flight management computer preceded the confirmation of the landing runway. This action and the crew's experience that B737 aircraft generally use runway 16L/34R would have served to reinforce to them that they were being processed to land on runway 34R.

The ambiguities in the Sydney parallel runway procedures terminology have the potential to lead to misunderstandings between controllers and pilots. This is due to the apparent interchangeability of terms used in the various documents to describe runway application.

CONCLUSIONS

Findings

1. The flight crew was familiar with parallel runway operations at Sydney.
2. Brisbane Control, in accordance with normal procedures, provided the flight crew with an expectation to land on runway 34R.
3. Current procedures did not require that controllers alert flight crews to any change to a landing runway expectation.
4. Sydney approach, and later the Director, advised the flight crew of runway 34L for their landing.
5. Current procedures did not require that flight crews read back runway information to confirm recognition of the correct runway.
6. On initial contact with the Director, the flight crew did not advise, and the Director did not confirm, their landing runway intention.
7. The flight crew intercepted the final approach for runway 34R and was subsequently cleared to land.
8. There was no breakdown in separation standards.
9. Instructions, concerning some aspects of parallel runway operations at Sydney, are inconsistent and ambiguous.

Significant factors

1. The flight crew did not recognise the change of runway advice.
2. The controllers were unaware, until late in the approach, that the flight crew intended to land on runway 34R.

Safety Action

As a result of the investigation into this occurrence, the Bureau of Air Safety Investigation issued the following interim recommendations:

IR950213 issued 3 January 1996

That the Civil Aviation Safety Authority and Airservices Australia jointly:

1. remove from all published standard operating procedures, publications and instructions any ambiguities concerning the meanings of the terms "expected", "nominated" and "specified" in relation to landing runways.
2. amend the published standard operating procedures, publications and instructions to include advice to flight crew that the approach will involve a cross-over manoeuvre if their aircraft is required to cross a parallel approach path;
3. rectify the ambiguity in the present requirement for flight crew to report "visual" for independent visual approaches to parallel runways;
4. include a requirement for flight crews to read back the runway in aeromobile communications phraseology; and
5. amend standard operating procedures to ensure that flight progress strip notations include verification of complete runway information transfer.

CASA response dated 12 March 1996.

I refer to your Interim Recommendation number IR950213 concerning the accident involving Boeing 737, VH-TAI at Sydney Airport on 0 October 1995. The Authority wishes to forward the following response.

Summary

1. Agreed. CASA and Airservices will review all documentation.
2. Agreed. CASA will pursue with Airservices to provide necessary phraseology.
3. CASA does not concur with BASI' s concern.
4. CASA and Airservices are working to implement new phraseologies.
5. This is an issue for Airservices to address.

Background to response

1. CASA agrees that terminology used by ATC to identify the landing runway must be unambiguous. CASA will, with Airservices, review all documentation to satisfy this requirement.
2. In the case of a crossover to final, CASA agrees that ATC phraseology should positively alert a pilot to the situation. CASA will pursue this issue with Airservices to provide the necessary phraseology.
3. CASA is unable to understand BASI's concern with regard to the use of the phrase 'VISUAL'. It is only used by a pilot to indicate to a controller that the aircraft can proceed to destination in accordance with the visual approach requirements. When localizer tracking is required and the aircraft is established, pilots report

'ESTABLISHED' and when tracking visually to the runway pilots must report 'RUNWAY....LEFT/RIGHT IN SIGHT' when the appropriate runway is sighted. As each situation has a separate and district phrase, there is no ambiguity.

4. CASA and Airservices are working to implement ICAO phraseology in Australia.  As runway readback is included in ICAO procedures, this recommendation will be satisfied when the new phraseologies are introduced.
5. Amending standard operating procedures to ensure that flight progress strip notations include verification of complete runway information transfer is an issue for Airservices to address.

Response status: Closed-Accepted.

Airservices Australia response dated 22 July 1996

I refer to your letter dated 3 January 1996, concerning interim recommendation IR950213, and relating to an incident involving parallel runway operations at Sydney airport. The particular circumstances relating to this incident were discussed at a special meeting of the Sydney Parallel Runway Operations Group, held on Tuesday 28th November 1995. This group comprises representatives from Airservices operational and policy areas, CASA and industry, and is charged with the development of standards and procedures for parallel runway operations.

Several recommendations were made at that meeting, proposing changes to AIP SUP H36/95, AIP, and Sydney Local Instructions, to preclude the possibility of a recurrence of this incident. It was felt, however, that major changes in reaction to a single incident should be considered carefully and take into consideration the amount of pilot and controller training material in circulation, and the potential risks of changing procedures so soon after implementation of independent visual approaches.

In relation to your recommendation (i), references to runway expectation are being removed from approach and director phraseology. Instead, the runway will be specified without the word "expect", and any subsequent change will only take place with the concurrence of the pilot in command.

In relation to recommendation (ii), a controller awareness program will be initiated, with guidance material being published in an information circular, highlighting the need for caution and advice to pilots when crossing centrelines. It was felt, however, that pilots should normally be aware of centreline crossover through the controller advice of runway and circuit direction.

In relation to recommendation (iii), it is difficult to see that there is any ambiguity in the pilot report requirements. During conduct of any approaches, pilots established on a LLZ should report established. The LLZ established call confirms to the controller the track guidance being used, and the report of "visual" is required to enable the application of the standard. When a pilot is not tracking on the LLZ, the track guidance is confirmed by pilot report of "runway in sight".

In relation to recommendation (iv), consultations are continuing separately on a proposed completely revised AIP and MATS phraseology package. This review is intended to bring Australian phraseologies into line with ICAO, as far as is practicable. One of the items being considered as part of this process, is the ICAO recommended practice of runway readback. It was considered prudent to wait for the results of this project, instead of a unilateral implementation at Sydney.

In relation to recommendation (v), any outcome from the revision of phraseologies that requires a pilot readback will be recorded on the flight progress strip as currently specified in MATS page 10-2-2 paragraph 32g.

It is proposed to review and re issue the AIP SUP relating to parallel runway operations in the near future, and account will be taken of the recommendations contained in BASI interim recommendation IR950213.

Response status: Closed Accepted

At 1139 the Canberra approach controller cleared Hawk 02 to track to Canberra from the east-south-east at an altitude of 3,500 feet. Just after this VH-GJZ departed from runway 12, tracking via Bindook. The pilot had been told to turn left onto a heading of 070 degrees and climb to 4,500 feet. A short time later the approach controller passed the responsibility for Hawk 02 to Canberra tower, saying this aircraft was to track for a left base on runway 17. The pilot of Hawk 02 was given a visual approach and commenced descent.

At 1142 the pilot of VH-GJZ advised approach control he was climbing to 4500 feet and turning left onto a heading of 070 degrees. The approach controller realised that separation would be lost between VH-GJZ and Hawk 02. The controller told the pilot of VH-GJZ to stop the turn on a heading of 110 degrees. Horizontal separation reduced to about one mile before vertical separation was established. The pilot of VH-GJZ reported sighting Hawk 02 and was told to pass behind this aircraft, then turn left onto a heading of 040 degrees.

A third aircraft, callsign Squirrel 005, had been cleared direct to Canberra from the north-east at an altitude of 4500 feet. As a result of the controller's instructions to VH-GJZ this aircraft and Squirrel 005 were now on conflicting tracks. The pilot of Squirrel 005 was told to turn 30 degrees to the right and the pilot of VH-GJZ was told to climb to 7,000 feet. Horizontal separation between VH-GJZ and Squirrel 005 reduced below the minimum standard before vertical separation was established.

Significant Factors

The following factors were considered relevant to the development of the incident:

1 The approach controller did not satisfactorily plan the handling of the sequence.

2 The approach controller did not take timely steps to avoid potential traffic conflicts.

The pilot stated that during the take-off run with both engines at full power, the aircraft veered to the left. At 40 kts the aircraft did not respond to full right rudder, so the pilot aborted the take-off. The aircraft came to rest in an adjoining cultivated paddock with the nosewheel broken off and both propellers damaged.

The pilot had recently passed his General Flying Progress Test (GFPT), and on the day of the accident had completed a dual check on VH-PXB. He had subsequently made three local passenger carrying flights. During the fourth flight the pilot stated that, as the aircraft approached Archerfield circuit area, and shortly after he had changed the fuel selector from the right tank to the left tank, the engine lost power. The pilot decided to conduct an emergency landing on a road. During the landing roll, the left wing struck a signpost.

Both fuel tanks were drained after the accident and the left fuel tank yielded 18.5 litres and the right fuel tank 9 litres. The published unusable quantity for each tank is 1 litre. This implies that adequate fuel was available for continued engine operation. However, the pilot reported that the flying conditions were turbulent at about the time he changed tanks, and it is possible that air may have entered the fuel system while the right tank was selected. The engine was ground run after the aircraft had been recovered and functioned normally.

The reason for the loss of engine power was not determined.