Loss of separation

The Cessna Citation (C550) was on climb to flight level 310 (FL310) but had been restricted to FL280 due to opposite direction traffic on descent to FL290. The Sector 16 controller had coordinated the planned flight level with the next sector and had notated the flight strip accordingly. A following Boeing 737 (B737) was on the same track and had planned to climb to FL350 but had also been restricted to FL280 due to the same conflicting traffic.

The controller had also coordinated this flight with the following sector and had notated the flight strip accordingly. The B737 reached FL280 first and began to increase speed as the aircraft maintained level flight. At this time the slower C550 was approximately 25 NM ahead of the B737, but still climbing. An opportunity to establish positive vertical separation was available at this time but the controller did not invoke separation assurance techniques. The controller had intended to climb both aircraft to their respective flight planned levels after they had passed the opposite direction traffic and expected to maintain the 5 NM radar separation standard prior to that event.

Changes to airspace procedures had taken place 3 days prior to the occurrence and it was the controller's first shift involving the new procedures since the introduction of Class "E" airspace. He had been on duty just 30 minutes when he received his first request for a clearance involving "E" airspace. This request resulted in him having to perform coordination tasks with both Flight Service and Sector 19. The consequent workload due to his unfamiliarity with the new procedures resulted in his attention being diverted from the radar display, as the B737 closed on the C550. The C550 subsequently levelled at FL280. A controller not performing tasks related to Sector 16 noticed the rate of closure between the C550 and the B737 and brought this to the attention of the Sector 16 controller who immediately issued a radar vector to the crew of the B737.

As the aircraft passed, the horizontal separation reduced to approximately 1/2 NM. The crew of the B737 did not sight the C550. Radar analysis indicated that the closing speed varied between 70 and 220 kts. It had been planned that all Sector 16 controllers were to have undergone 5 simulator exercises prior to the implementation of "E" airspace. However, due to time constraints and technical difficulties, only 4 exercises had been developed. An agreement had been reached for all Sector 16 controllers to receive those 4 exercises. However, the controller on duty had only experienced 2 complete and 1 "over the shoulder" simulator exercises at the time of implementation. Moreover, the changes issued by the Civil Aviation Safety Authority, 7 days prior to implementation, resulted in a new version of Local Instructions for Sector 16 only becoming available to the controller 4 days before the occurrence. This limited the time available for adequate training of air traffic controllers.

SAFETY ACTION

As a result of this occurrence, the Bureau of Air Safety Investigation is currently investigating a perceived safety deficiency. The deficiency identified relates to air traffic controller training. Any safety output issued as a result of this analysis will be published in the Bureau's Quarterly Safety Deficiency Report.

The military Lockheed C-130 aircraft was tracking to Pearce from Edinburgh via reporting points BADJA and PEPPA, whilst an Avro RJ-70 was tracking from Perth to Telfa via CLIFY and PEPPA.

The pilot of the C-130 had been cleared to descend to 8,000 ft by Perth ATC and the RJ-70 was on climb to FL290. However, the C-130 was observed on radar to drift south of track and then turn left prior to PEPPA. As a result of the early turn, the C-130 passed approximately 3 NM south of PEPPA. The C-130 crew could not recall why the aircraft had drifted south or turned early. However, it was reported that it was possible that the C-130's inertial navigation system (INS) may have experienced some drift. It was also reported that because the INS calculates an intercept to the next leg based on the aircraft's expected turn radius, it may have commanded a turn earlier than ATC expected.

Perth ATC directed the pilot of the RJ-70 to turn the aircraft right to pass behind the crossing C-130. Just after the RJ-70 pilot gave his read-back to the instruction, he advised Perth ATC that he had received a TCAS alert. Perth ATC passed traffic and the two aircraft passed abeam each other at 4 NM with no vertical separation. There was a breakdown in separation as the required standard was 5 NM.

Perth ATC reported that just prior to the incident, Pearce Control had experienced a radar failure. As a result, the Perth "Inner" air traffic controller, who was under training and being supervised at the time, experienced a substantial workload increase requiring him to scan a larger area than normal, thus changing his radar scan parameters. He also had 12 aircraft on frequency. The trainee air traffic controller's workload, whilst heavy, was not considered excessive because he had completed 4 weeks of the 6 week training course and was already an experienced procedural controller.

Perth ATC reported that at 15 NM inbound to PEPPA, the C-130 had a groundspeed of 330 kts and the trainee controller assessed that the separation between the aircraft would be adequate. However, abeam PEPPA, the C-130's groundspeed unexpectedly reduced to 220 kts before increasing again to 255 kts. The speed changes may have been associated with the aircraft accelerating in the descent and then slowing following the level off. The C-130's unexpected speed changes and early turn near PEPPA adversely affected the trainee controller's planned separation between the two aircraft.

The trainee had issued frequency transfer instructions to the pilot of the C-130 just before the impending breakdown in separation was noticed. Although the training officer assumed control of the position as the separation breakdown developed, he could only issue instructions to the pilot of the RJ-70. This limited the options available to the training officer when he was attempting to correct and avoid the separation breakdown.

It is probable that the workload increase caused by the non-operative Pearce radar associated with the C-130's unexpected deviation from the air route and groundspeed changes, were factors in allowing the breakdown in separation to develop. When the training officer noticed the developing separation breakdown, his options in ensuring the required separation were limited by the C-130 being on another frequency.

VH-AJJ, a Westwind, departed Sydney runway 16L on a COOKS ONE standard instrument departure (SID) for a maintenance test flight. The SID required the aircraft to turn left on reaching 500 ft. Radar observation, however, indicated that the aircraft remained on runway heading and did not comply with the SID. The departures controller instructed the pilot to turn left immediately as there was an international B747 departing from runway 16R on a GUEST TWO SID. The tower had both aircraft in sight at all times but there was a breakdown in separation.

The operator of VH-AJJ reported that the flight test was to observe the operation of the aircraft's pressurisation system. The pilot was aware of the departing B747 from the parallel runway but was intent on cabin and engine instruments during the initial climb. He realised he had failed to comply with the SID when he was instructed to turn left.

A Dash 8 aircraft, VH-TQP departed Sydney for Coffs Harbour on climb to FL190. Shortly after this aircraft departed, a second Dash 8, VH-TQG departed Sydney for Port Macquarie on climb to FL170. The second aircraft in the departure sequence was a later model Dash 8, capable of maintaining a high-speed cruise climb. To maintain separation between the two aircraft, both were restricted to 160 KIAS on climb.

As TQG approached top of climb (FL170), the controller observed TQP, the slower Dash 8 ahead, still well below FL170. To expedite the departure of the faster aircraft and achieve coordination with a southbound aircraft, the controller decided to restrict the climb of TQP to FL160. The crew understood their instruction was to FL170, selected this level with their assigned altitude indicator and read this altitude back to the controller. The controller did not detect the discrepancy between the assigned altitude and the crew's read back of the instruction.

The controller was then distracted for a short period of time with other sector controllers involved with an aircraft experiencing communication difficulties. During this period the mode C readout from TQP was noticed at FL168, in close proximity to TQG maintaining FL170.

TQP was instructed to immediately descend to FL160 and TQG initiated a turn to increase the separation between the two aircraft. The aircraft proceeded safely to their destinations following this occurrence.

A Boeing 747 (B747), en route from Sydney to Hong Kong, was maintaining flight level 350 (FL350) and had been radar identified by Manila Air Traffic Control. The flight crew observed another aircraft maintaining the same level approximately 20 NM to the east displayed on their aircraft's traffic alert and collision avoidance system (TCAS). The B747 crew sighted the other aircraft as it crossed their track from east to west. The crossing aircraft was en route from Manila to Kuala Lumpur and had initially been assigned FL280.

On first contact with radar control that crew had been instructed to climb to FL350. The controller noted the proximity of the two aircraft and instructed the crew of the B747 to climb immediately to FL370. At the same time the B747 TCAS announced a "Climb" resolution advisory (RA) instruction. The B747 crew complied with the climb instruction. The crew of the crossing aircraft advised the B747 crew that they had received a TCAS RA instruction to descend which they had followed. There was a breakdown of separation. Following an investigation of the occurrence the Republic of the Philippines, Department of Transport and Communications air traffic services investigator reported that all area control centre controllers would undergo radar simulator training to enhance their proficiency. The training was to commence in January 1998.

A domestic Boeing 767 was inbound to the holding pattern at Cullerin and the crew had been issued a requirement to reach their assigned level of FL 290 before entering the pattern. The crew of a Boeing 747 already established in the Cullerin pattern observed the inbound aircraft descending through its assigned level and alerted ATC. The descending aircraft was visually sighted and was displayed as traffic on the TCAS.

The pilot of the descending B 767 indicated that the autopilot had not captured the assigned level as programmed. The pilot indicated that this was possibly complicated by late amendments to the holding instruction. The crew intervened when it became apparent that the autopilot had not captured the assigned level, but not before the aircraft had descended approximately 400 ft through the assigned level.

The controller indicated that 1,000 ft vertical separation between the two aircraft was quickly re-established, and the aircraft were horizontally separated by 3 - 4 NM.

FACTUAL INFORMATION

The Airbus A340 (A340) was on air route A587 and the Boeing 737 (B737) was on air route G462. The routes intersected at waypoint KETUT, which is 200 NM bearing 103 degrees magnetic from Bali. Both aircraft had been cleared to enter the Bali flight information region (FIR) at flight level (FL) 350. The waypoint KETUT is a reporting point for aircraft on A587 but not for aircraft on G462. All communications prior to this point were made on high frequency (HF) radio SELCAL, which does not alert pilots that there may be conflicting traffic.

The conflicting traffic was under the jurisdiction of the Bali East controller. There was radar coverage to approximately 150 NM, and a combination of procedural and radar control was being used by the controller. Flight plans for both aircraft had been received at the Bali Control Centre. However, the flight plans did not indicate an estimate for the KETUT position.

At the time of the incident, the controller had three aircraft on his frequency - the B737, the A340, and a Boeing 747 (B747) en route Singapore to Sydney. The B747 was tracking via Bali at FL330 and the B737 would have had to descend through that level for arrival at Bali. The B747 was identified on radar and was estimating overhead at Bali at 1436 UTC. The other two aircraft were not identified on radar.

At 1410, the A340 pilot reported to Bali East at ONOXA, a reporting point on the FIR southern boundary. ONOXA is 197 NM from KETUT on the designated route A587. The Bali East controller then requested an estimated time of arrival (ETA) for IKAPI, a reporting point on the northern FIR boundary where the aircraft would exit the FIR. The pilot advised an estimate of 1530 and the controller coordinated this information to Ujung Pandang, which was the next control agency.

KETUT was a compulsory reporting point for aircraft on A587. The reason the A340 crew did not advise an estimate for KETUT when reporting at ONOXA was not established. The reason the controller did not request A340's ONOXA estimate was also not established.

At 1428, the crew of the B737 reported at SOTRA at FL350. The Bali East controller requested an ETA for GABIT, a reporting point 150 NM east of Bali, and the crew advised that their ETA was 1447. At 1439 the crew of the A340 requested advice of any traffic at FL350. The controller asked the crew to repeat the request. The A340 crew then stated that they were descending to FL310 to avoid conflicting traffic. The crew of the B737 asked the crew of the A340 to report their position. There was no reply to this request. The A340 crew then reported to the Bali East controller that there was traffic 2 NM to their north, passing from right to left at FL350. The controller questioned this report. The A340 crew repeated the information and reported that they were maintaining FL310. The controller then advised the A340 to "maintain FL310 for a while". At 1443, the A340 was cleared to climb to FL350 by the Bali East controller.

At the time of the occurrence, the control centre was staffed by four controllers and a supervisor. The Bali East controller had left the control room on a break and another controller had taken over control duties of the Bali East sector. The controller taking over had not been aware of the potential conflict between the aircraft on air routes A587 and G462.

Although the occurrence location was within normal radar coverage, the radar data was not being recorded during the period of the incident due to a technical problem.

Each control position had a single flight progress strip bay. The bay was used to store pending flight progress strips. Active flight progress strips were placed on the flat console surface in front of the radar screen. There was a checklist for the compilation of flight progress strips, but this had not been compiled with.

ANALYSIS

The controller had taken over the Bali East sector duties only a short time before the incident and he had not been properly briefed on the traffic that was within his jurisdiction. His attention appeared to have been focussed on the potential conflict between the southbound B747 at FL330, and the B737 inbound to Bali at FL350. He was aware that the B737 would require descent through the B747's level but had not been briefed that there was a requirement to establish vertical separation between the A340 and the B737. The situation was compounded because the flight progress strips were not prepared in accordance with the checklist and the flight progress bays were not being utilised in the most appropriate manner to provide situational awareness of the traffic in the sector.

Although, in this instance, an estimate for the KETUT position was omitted by the crew of one aircraft, the provision of an estimate by the crew of the other aircraft may have assisted the controller to better appreciate the traffic disposition, and alerted him to the confliction. In this context, the promulgation of KETUT as a compulsory reporting point for the two intersecting routes (A587 and G462) would have provided a higher level of safety.

Controllers normally combined flight progress strips for aircraft being controlled on the radar display console. However, this action ensured that aircraft being controlled by different procedures (radar and procedural control) were not easily differentiated. It also provide limited cues to controllers to assist them in maintaining an awareness of traffic disposition in their area of responsibility. Use of the available flight progress board with a number of key position designators would assist controllers in the management of aircraft under procedural control. Positioning the flight progress strips for radar-identified aircraft at the radar display console, adjacent to the procedural flight progress board, would have provided a better overall representation of traffic disposition.

SIGNIFICANT FACTORS

1. KETUT was not a compulsory reporting point for both routes.

2. The controller who compiled the flight progress strip for the A340 did not notate KETUT on that aircraft's strip.

3. The controller handing over responsibility for Bali East control did not conduct an adequate handover.

4. The Bali East controller did not adequately review the active flight progress strips after taking over responsibility for the position.

5. A procedural flight progress board was not used to manage aircraft operating in the Bali East area of responsibility.

6. The crew of the A340 did not advise an estimate for KETUT, a compulsory reporting point.

7. The Bali East controller did not appreciate the conflict between the A340 and the B737.

SAFETY ACTION

As a result of this occurrence the B737 operator:

1. issued a safety notice to all aircrew advising them of the details of the occurrence; and

2. recommended that the priority be given for the fitment of a traffic alert and collision avoidance system (TCAS) to company aircraft operating international flights.

In addition, the Bureau of Air Safety Investigation suggests that the Indonesian Directorate of Aviation Safety, in conjunction with Angkasa Pura I:

1. review flight progress strip display practices and procedures, including strategies to ensure their correct compilation;

2. review published handover/takeover procedures, including strategies to improve controller compliance; and

3. 3. Include KETUT as a compulsory reporting point for aircraft operating on route G462.

A SAAB 340 aircraft was conducting a right circuit for a landing on runway 34 right at Sydney, NSW. The aircraft subsequently positioned itself behind a Boeing 767 on final approach to runway 34 left, infringing wake turbulence separation standards.

The crew report that the controller had inquired while the aircraft was downwind if they would like a radar vector to slot them between to heavy aircraft inbound for runway 34 left. The subsequent vector for base leg was assumed by the crew to be a right base, for runway 34 left. The crew did not recall hearing the controller refer to the assigned runway at any stage of their approach. On being cleared for a visual approach at approximately 10 miles the crew dialled up the runway 34 left ILS.

The controller observed the SAAB flying through the centreline of runway 34 right at approximately 7 miles. The crew was cautioned about wake turbulence, taken out of the landing sequence and processed for a right-hand circuit for runway 34 right. The aircraft landed safely without further incident.

Two hot air balloons had departed from Archerfield in company at approximately 0615 EST and the pilots had been cleared to operate not above 2,500 ft by the terminal area controller. One of the balloons was operating a transponder but the transponder was not equipped with a mode C (height readout) function. The other balloon was not equipped with a transponder. The ATC radar display of the transponder return from the balloon was intermittent.

After initial identification of the balloon, radar contact was lost for a period of some minutes. At the same time, a Boeing 767 (B767) was inbound to Brisbane from the north and was being radar vectored for a left circuit for a runway 01 ILS approach. At about the same time, a light aircraft departed Archerfield for Brisbane, and the pilot requested airways clearance. The pilot was instructed to proceed outside controlled airspace towards the TV towers, and to expect clearance at that position. The pilot was advised of the presence of the balloons. The radar return of the balloons was not visible to the controller at this time. The B767 had by this time reached a position 5 NM south-west of Archerfield, was cleared to 3,000 ft and vectored onto a heading of 090 for the base leg. At about 1 NM west of Archerfield the B767 pilot requested a heading of 040 "due to balloon" and the aircraft was observed to climb by the controller. The heading requested was approved with a further clearance to descend to 2,500 ft and to commence the ILS approach. As the B767 intercepted the localiser, the radar return for one of the balloons reappeared on the radar, just south of the return for the B767. The controller then realised that there had been a breakdown of separation standards between the balloon(s) and the B767.

The balloon pilots had seen the approaching B767 and said that it had appeared from the west out of cloud. The aircraft was heading towards them but climbed and turned left. One of the balloon pilots estimated that the B767 passed 500 metres to the north and about 600 ft above.

During the time that the B767 was being radar vectored, the balloon transponder return had not been visible on the controller's radar screen, and the controller had temporarily forgotten their presence.

The balloon operator said it was not uncommon for radar contact with balloons to be lost, because the transponder antenna was sometimes masked by the gas installation in the basket. The operator has undertaken to improve the antenna installation to overcome this problem. The balloon operator was operating under a Civil Aviation Safety Authority (CASA) Permit. One of the conditions of the permit required the balloon to be fitted with a transponder capable of mode C operation, to be operated as directed, unless exempted by ATC. At the time of the occurrence, the balloon was operating under an exemption. As a result of the occurrence, Airservices Australia removed the exemption was removed and stated that, in future, no such exemptions for operations of this nature would be made.

A recommendation was made to the Archerfield District Flying Operations Manager that the words "unless exempted by ATC" be removed from the permit.

At the time of the incident, Brisbane ATC Sector 11K Radar and Sector 11 Procedural were being operated as a combined control position. Tindal radar was out of service. A Sector 11K Radar trainee, who was already rated on Sector 11 Procedural, was operating the combined position under the supervision of a training officer.

A Westwind aircraft was tracking Darwin - Tindal - Alice Springs and climbing to Flight Level 330 (FL330), with an estimate for overhead Tindal of 1911 CST. A Boeing 747 (B747) was tracking in the opposite direction to the Westwind and estimating overhead Tindal at 1912. The Westwind was still displayed on radar, but the B747 was outside radar coverage.

Two other aircraft were also under the control of the trainee. The trainee initiated co-ordination with Melbourne ATC to amend the clearance for one of these aircraft. Before Melbourne ATC answered, the pilot of the Westwind requested an amended level of FL370. The trainee immediately cleared the Westwind to this level. Shortly afterwards, the crew of the B747 reported that there was traffic 1,000 feet below their aircraft. (The crew later advised that they had received a Traffic alerting and Collision Avoidance System (TCAS) traffic alert on the other aircraft.) The Westwind was observed on radar to be climbing and approaching FL340. The controller instructed the Westwind pilot to maintain FL340 and, immediately afterwards, to descend to FL330. He informed the Westwind pilot of the relative position of the B747.

The investigation determined that the two aircraft had passed with approximately 1.5 NM lateral separation, as the Westwind descended through FL335 and the B747 maintained FL350. In the procedural control environment existing at the time, the required vertical separation standard was 2,000 ft. The required lateral separation standard was 10 minutes combined flight time.

Neither the trainee controller, nor the training officer, checked the flight progress strips for the two aircraft prior to the Westwind being cleared to climb to FL370. The training officer was checking lateral separation calculations for other traffic on a map at the time and was not closely monitoring the trainee.