Loss of separation

Report not released due to no IP comment from the Indonesian agency on this joint investigation report.

ASOR hard copy held on occurrence file.

Report not to be released without the approval of the Director or a DD.

Signed

Alan L Stray

Deputy Director Investigations

6 July 1998

The crew reported for duty for a scheduled flight from Sydney to Moruya. The co-pilot was to be the handling pilot for the sector. He had recently been endorsed on the Fairchild Metro 23 and was undergoing line training under the supervision of the pilot in command.

After take-off, the aircraft was cleared to climb to 5,000 ft. This instruction was acknowledged by the crew. As the Metro was climbing to the south, opposite direction traffic was a Boeing 737 maintaining 6,000 ft, inbound to Sydney. A short time later the Metro was observed on radar to be approaching 5,600 ft, in close proximity to the B737. The departures controller asked the crew of the Metro to report their altitude, whereupon the aircraft was observed to commence a descent. The Metro passed below the B737 while descending to 5,000 ft.

Recorded radar data later showed that separation between the aircraft was reduced to 1.3 NM laterally, and some 600 ft vertically, at their closest point of approach. The Manual of Air Traffic Services specified that 1,000 ft vertical or 3 NM lateral separation was required at the time.

Both the pilot in command and the co-pilot of the Metro did not recall noticing an altitude alert, which should have provided the crew with both aural and visual warnings 1,000 ft before the aircraft reached the selected altitude. The pitch trim control was being used during the climb, which also provided an aural annunciation through the cockpit loudspeaker. Company procedures required the non-handling pilot to alert the handling pilot some 500 ft prior to the aircraft reaching its cleared altitude. In this instance that procedure was not carried out.

The pilot in command later reported that he had been completing paperwork during the initial climb but felt that the co-pilot appeared to be handling the departure sequence adequately and did not require close supervision. However, the co-pilot said he had working hard to cope with the workload and may have channelised his attention on the speed control of the aircraft.

The breakdown of separation in this occurrence resulted from a combination of factors. These included: the attention of the handling pilot being unduly focussed on the speed control of the aircraft, at the expense of monitoring altitude; and the attention of the pilot in command being diverted from monitoring and supporting the performance of the handling pilot. The aural annunciation of the pitch trim control may have masked the sound of the altitude alerting system.

As a result of this occurrence, the standard operating procedures of the company will be amended to provide additional defences designed to prevent a recurrence.

VH-LRT was on a visual flight rules (VFR) flight from Riddell via Albury to Bankstown and cruising at 7500 feet. The pilot was having difficulty with visual navigation due to cloud cover and as a result the aircraft entered the Wagga control area steps without an airways clearance.

VH-CMH was on climb out of Wagga for Sydney when the crew saw LRT on a converging heading and estimated it was 500 ft to 700 ft below. The crew of CMH reported that there was scattered to broken cloud from 1500 ft to flight levels. They reported the intruding aircraft to Wagga tower. The tower controller in turn checked with the Melbourne sector eight radar controller who advised there was an unidentified aircraft painting at 7500 ft 25 miles to the northeast of Wagga. That aircraft was duly identified as LRT, the offending aircraft.

LRT had previously transgressed twice on the trip from Riddell, on one occasion entering controlled airspace, climbing through 10,000 ft without a clearance and on the second occasion leaving a cleared altitude without a clearance while transiting Albury control area steps. For this reason, the Melbourne sector eight radar controller had been monitoring the aircraft's progress but had lost radar contact when the aircraft was to the south of Holbrook.

The radar tape was monitored and revealed that a code 2000 return appeared on radar about 16 minutes later at a position 15 miles south of Wagga at 8400 ft on a northerly track entering Wagga CTA. It was later determined that this return was LRT. The tape showed that LRT proceeded in a north-north-easterly direction through Wagga CTA and gradually descended to 7000 feet when, at 15 miles northeast of Wagga, the two aircraft were one mile apart at the same altitude.

Factors

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

• The pilot of LRT was not competent to navigate his aircraft in the cloud conditions existing at the time. This resulted in the aircraft entering the Wagga CTA without a clearance and because the Wagga air traffic controllers were unaware of LRT, they had not arranged separation with CMH.
• The Melbourne sector eight controller had been watching LRT, because of earlier incursions, until its return disappeared from radar. Although Wagga airspace is not the responsibility of the Melbourne sector eight controller, it is considered that in this instance when the code 2000 return appeared entering Wagga airspace, the sector eight controller should have suspected that it may be LRT and alerted the Wagga controllers accordingly. Such an alert should have resulted in this incident being avoided.

The pilot of the C182 had been cleared to operate over the metropolitan area at 5,000 ft, provided he remained 2 NM west of the extended centreline of Perth runways 03/21. This clearance had been issued several hours before, by a different controller to the one involved in this incident.

A C172 had also been operating over the metropolitan area at 1,500 ft and had been cleared to return to Jandakot.

The B767 was departing Perth from runway 21 and the crew intended to turn right and depart towards the north-west.

During the day, the pilot of the C182 had been following a regular, orbiting flight path which was displaced from the B767's planned departure track. After the B767 took off the C182 made two irregular orbits which took it further south than previously, and closer to the B767's path.

The approach-west controller was aware of a possible conflict between the B767 and both the C172 and C182, prior to the B767's departure. He had formulated a plan to ensure that separation was maintained. The plan required the B767 to maintain runway heading until it was clear of both aircraft. The initial clearance to the crew of the B767 was to maintain runway heading after take-off and to climb to, and maintain 6,000 ft. The altitude requirement was to keep the B767 clear of military airspace.

When the crew of the B767 gave their departure report the approach-west controller was engaged in a lengthy coordination discussion with another controller and he was not maintaining a continuous radar watch on the airborne traffic. He glanced at the radar and noted that the B767 was clear of the C172 before he directed the crew of the B767 to turn right onto 280 degrees. This turn was to allow the aircraft to use a preferred noise abatement track, one of a number that have been established in the metropolitan area. The controller had forgotten about the C182 and did not note its location when he glanced at the radar.

As the B767 established a heading of 280 degrees the C182, which was at the end of its southerly track, turned onto an easterly heading. This resulted in a reduction in vertical separation to 400 ft as the aircraft passed 2.8 NM apart. This was below the requirement to maintain at least 1,000 ft if the aircraft are less than 3 NM apart. The crew of the B767 received an electronic warning of the C182's proximity and also acquired it visually. No evasive action was necessary.

The evidence indicates that C182's long-standing clearance and its regular pattern may have caused the controller to establish a mind-set that the B767/C182 separation problem was less important than it actually was. The change in pattern was unexpected and went unnoticed. The controller's attention was distracted from the radar console by his discussion with another controller.

The combination of the unexpected and unobserved change in the C182's flight path and the distraction resulted in the controller only ensuring the B767 was clear of the C172 before he cleared the crew to turn onto a new heading.

The controller also indicated that he was mindful of the need to comply with noise abatement procedures and the priority he assigned to this may also have been a factor in his missing the impending confliction.

A Boeing 747-338 aircraft VH-EBU was inbound to Cairns on the 153 Very High Frequency (VHF) Omni-Directional Range (VOR) radial and had been assigned 3000 feet not below the DME steps. The pilot reported visual at approximately 15 NM south of Cairns and was then cleared to track for a left base for runway 15. VH-XFW an Embraer EMB-120ER was departing Cairns runway 15 for Bamaga and was given a departure instruction to turn left onto a heading of 030 degrees magnetic and to maintain 2000 ft. VH-EBU was observed by radar to descend below 3000 ft. The approach controller instructed VH-XFW to maintain 1500 ft and VH-EBU to maintain 2500 ft. The pilot of VH-EBU said it was too late and that he had the other aircraft in sight to his right. VH-EBU was then cleared for a visual approach and instructed to call the tower. VH-XFW was given climb and tracking instructions to intercept the outbound track.

Examination of the Automatic Voice Recording (AVR) tape and the radar tape showed that at 0119:20 EST the minimum horizontal separation was 1.33 NM, and the vertical separation was approximately 600 ft. The minimum separation standards are 3 NM and 1000 feet.

Interview with the crew of VH-EBU indicated that when VH-EBU was cleared to track for a left base for runway 15 (after reporting visual) it was assumed that they had been cleared for a visual approach. Accordingly, a descent to 1500 ft was initiated, which is circuit altitude. The crew of VH-EBU had not been cleared for a visual approach however, and thus had not been cleared to descend below 3000 ft.

Safety Action

The crew of VH-EBU have been counselled as to visual approach requirements, and a notice to pilots will be issued to prevent a recurrence.

The PA-28 had departed Moorabbin at 1734 ESuT and was on climb to 9,000 ft tracking via Melbourne, Fentons Hill and Mangalore to Tocumwal.  The pilot had initially been in Instrument Meteorological Conditions but broke out of cloud just prior to passing over Fentons Hill.

The B737 had departed Melbourne for Brisbane at 1747 ESuT and was tracking via a Runway 27 DOSEL 3 standard instrument departure (SID).  The crew requested, and were given, a climb at 300 kts in lieu of the compulsory 250 kts below 10,000 ft, and the aircraft passed close to Fentons Hill while tracking on the SID.  The crew reported that they were in cloud initially and broke out just past Fentons Hill.

The air traffic controller had both aircraft on his frequency and attempted to coordinate a shortened track for the PA-28.  This request was denied due to the volume of traffic. Consequently, the controller had to radar monitor the climb of three previously departed aircraft past the PA-28. Track shortening on this particular route is normally available.  Additionally, the controller had two aircraft arriving from Adelaide which required a combination of radar vectoring and vertical separation to ensure their safe and timely arrival. The arrival track from the west crosses over the departure track from runway 27.

The traffic volume and complexity required significant concentration by the controller who did not notice that the B737 and the PA-28 were on converging tracks. Radar analysis indicates that the two aircraft came within 1.3 NM whilst at the same altitude with the B737 passing in front of the PA-28.

Neither crew saw the other aircraft and were not aware of the conflict.

There was a breakdown of separation standards.

Findings

1. The crews of both aircraft were operating in accordance with air traffic control instructions.
2. The traffic volume and complexity were such as to result in a high workload situation.
3. The air traffic controller did not notice that the two aircraft were converging into conflict.
4. The two aircraft came within 1.3 NM with no vertical separation.

Safety Action

Airservices Australia have initiated changes to the Melbourne terminal area airspace that will remove the area of conflict involving the inbound tracks from the west. These changes will come into effect on 20 June 1996 and will significantly reduce the workload on approach/departures control.

Factual Information

GIA829 was maintaining flight level (FL) 350 on air route G326 and was estimating the flight information region boundary at ONOXA at 0330 UTC.  This information was correctly coordinated by Brisbane sector control to Bali control.

GIA899 was maintaining FL 330 on air route A576 and had passed CURTIN at 0245 with an estimate for the boundary at ATMAP of 0347.  This information was also correctly coordinated with Bali control and included the specific indication of a non-standard flight level.

During the coordination process, Bali control advised that they could not accept GIA899 at the non-standard level (FL 330) and required that aircraft at FL 350.  Brisbane sector accepted this requirement and after checking his flight progress strips, the controller calculated that there would be 13 minutes between the two aircraft at Bali (Den Pasar).  As the required standard is 10 minutes, he then instructed the crew of GIA899 to climb to FL 350.

At 0329 the Brisbane controller was checking his flight strips and realised that the flight plan time interval for GIA899 between ATMAP and Bali was approximately 10 minutes in error and the separation between the two aircraft was in fact only 2-3 minutes at Bali.

When the flight times were corrected, it was realised that GIA899 and GIA829 were already within the zone of conflict and that a breakdown in separation standards had occurred.

The Brisbane controller immediately initiated coordination with Bali control, but it was some minutes before contact was made.  When coordination was completed the Bali controller accepted responsibility for re-establishing separation and commenced distance checks from the two aircraft.

It is not known exactly which separation standard was affected by Bali control.

Analysis

When Bali control made the request for both aircraft to be at FL 350 it was based on the previous reported positions of the aircraft and the flight planned times from those positions to Bali.  In the case of GIA899 this time interval was incorrect and contained an approximate 10 minute error.

As both Brisbane and Bali control were using the same flight plan details, it is likely that both controllers considered that a 10 minute separation standard existed.  It was only when the Brisbane controller checked his time intervals against those being commonly used by other aircraft during the period, that the error was noticed.

The computer aided strip printing system only provided a proportional breakdown of the block times notified on the flight plan.  In this case, the first block times were correct and there was no trend to alert control to any time interval error.

There was also a table of estimated time intervals provided on the control console for controller reference.  This table had several erroneous time intervals included in its data and was commonly not used by controllers.

Findings

1. The flight plan for GIA899 contained an erroneous time interval.
2. The table of estimated time intervals held at the Brisbane control console contained several errors.
3. Bali control was unable to accept GIA899 at a non-standard level (FL 330).
4. Both the Brisbane and Bali controllers had information that indicated that a 10 minute separation standard existed.
5. All coordination between Brisbane and Bali was carried out in accordance with the Letter of Agreement.
6. Brisbane control issued a clearance for GIA899 to climb to FL 350 in accordance with Balis requirements.
7. The actual separation was approximately 2- 3 minutes.
8. A breakdown in separation standards occurred.

Significant Factor

The flight plan details for GIA899 were errant, in that they contained an incorrect time interval between ATMAP and Bali.

Safety Action

As a result of the investigation, Brisbane ATS management have removed the erroneous time interval table from the console.  It has been replaced with a table of average times for each type of aircraft using particular air routes.

Factual Information

An F27 Mk50 (F50) departed Sydney for Canberra from runway 34L and was being radar vectored by Departures South (DepS) for a left turn to intercept the 207-degree radial from the Sydney VHF omni-range beacon (VOR). The controller had restricted the climb of the F50 so that it was maintaining 5,000 ft.  This restriction was initiated in order to provide vertical separation from an arriving B747 that would be radar vectored from the west to make a left circuit to runway 34L.

The B747 was arriving at Sydney from Kuala Lumpur, Malaysia, and had been assigned 6,000 ft on descent in order to provide vertical separation from the F50. The crew had correctly read back the assigned level to Approach South (AppS) control who subsequently issued instructions for the B747 crew to turn their aircraft onto a downwind leg.

The radar vectors issued by DepS and AppS resulted in the two aircraft following flight paths that would cross each other approximately 10 NM south west of Sydney Airport.  As the aircraft converged, the controllers noticed that the height readouts from the two aircraft became 'garbled' (unreadable) and it was only when the aircraft had passed, and were approximately 2 NM apart, that they became readable again.  The AppS controller then observed that the height readout for the B747 showed 5,500 ft. At the same time the crew of the B747 queried their assigned altitude and were told that it was 6,000 ft.

By the time this conversation was completed, the aircraft were more than 3 NM apart and radar separation standard had been re-established.  The AppS controller then instructed the crew of the B747 to continue descent to 5,000 ft.

The crew of the B747 had been involved in making a public address announcement to the passengers as the aircraft had been passing 7,000 ft.  As a result, the attention of the crew was not totally on the altitude restriction and, as a result of high cockpit workload, the aircraft was allowed to descend below the assigned altitude.

The aircraft passed approximately 700-800 ft vertically apart, with no discernible horizontal displacement.  As the applicable standard is either 1,000 ft vertically or 3 NM horizontally, a breakdown of separation standards occurred.

Analysis

The operator's Standard Operating Procedures specified that the public address system should not normally be used below the transition altitude (10,000 ft in Australia).  The crew had been informed that the film being shown to the passengers was running late and the captain elected to wait until it had finished before making the public address announcements regarding arrival at Sydney.

Findings

1. The crew of the B747 allowed their aircraft to descend below the assigned level.
2. Both air traffic controllers acted in an appropriate manner.
3. The crew of the B747 made a public address announcement at a time that was contrary to company standard procedures.

Safety Action

As a result of the investigation the operating company reinforced its Standard Operating Procedures in respect to public address announcements below the transition altitude.

The separation standard was compromised between two international flights on crossing air routes. UAL862 was flying north-east along A579 at FL290 when the crew asked for clearance to climb to FL310 to avoid turbulence. This request was granted at 1255 and the aircraft climbed to the new level. Its calculated time at position RUSUM was 1312.

`ANZ39 was flying north-west along A464 at FL310. Its calculated time at RUSUM was 1311. The controller did not realise the confliction until 1309, and she immediately issued a descent to UAL862 which was relayed through the International HF Net manned by Flight Service. The aircraft did not respond to SELCAL. At 1312 an instruction to climb to FL330 was issued to ANZ39 via the HF link. The crew of ANZ39 reported that they had another aircraft displayed on their TCAS 14 to 15 nm south-east and at the same level.

The separation standard of 15 minutes was breached at 1256 when UAL862 climbed to FL310.

The Sector 8 controller had no direct communications with aircraft under her control. All air traffic instructions were relayed through Flight Service International on HF. The controller had approximately 6 to 8 aircraft under control in the sector. For reasons not established, the crew of UAL862 did not answer SELCAL.

The New Zealand Air Traffic Controllers were on strike and international aircraft were directed on nonstandard tracks. This added to the complexity of the operating position where lateral separation points had to be calculated for each situation.

Both aircraft crews were aware of the other aircraft due to the TCAS fitment in each aircraft.

After departing from Bankstown, the pilot of the Turbo Commander was instructed to climb to 5000 ft. At the same time a Boeing 767 was being vectored for a right downwind for runway 16R, on descent to 6000 ft. As the B767 approached that altitude the Departures (S) controller noticed that the Turbo Commander had climbed above 5000 ft and was approaching 6000 ft. The Turbo Commander was immediately turned right but lateral separation between the aircraft reduced to about 2 NM. The required separation standard was 3 NM lateral or 1000 ft vertical.

The pilot of the Turbo Commander later reported that he thought he had been cleared to 6000 ft, but may have confused the altitude clearance with a heading change to 160 degrees given at the same time.