The Boeing 737 (B737) had departed Sydney for Perth and the crew was maintaining the aircraft at flight level (FL) 280. The planned route was to track overhead Portland, Vic., then via air route T134 across the Great Australian Bight. The reporting point SUBUM was located on air route T134, approximately 220 NM south-south-west of Adelaide.
As the aircraft proceeded towards Portland, it was being radar monitored by the Melbourne Sector 4 air traffic controller. Flight crews were not required to report their positions while under radar observation. Accordingly, the time at which the aircraft was overhead Portland was recorded on the air traffic control flight progress strip by the sector controller, who then calculated an estimated time of passing SUBUM. This estimate was based on the flight-planned time interval from Portland to SUBUM and the actual time at which the controller saw the aircraft pass over Portland. The report was then coordinated to Adelaide Sector 4, which had jurisdiction for the route segment to SUBUM.
The Airbus A320 (A320) departed Melbourne for Perth and was tracking to intercept air route T134 via Portland. The crew was maintaining the aircraft at FL310.
The flight progress strips displayed information obtained from the respective company flight plans and showed that the B737 was estimating Portland at 1355 EST and SUBUM at 1444. They also showed that the A320 was estimating Portland at 1415 and SUBUM at 1500.
The Melbourne Sector 4 controller observed the B737 pass Portland at 1357 and, based on that observation, estimated that the aircraft would be approximately on time at SUBUM and elected not to change the estimated time of 1444. The position report was coordinated with Adelaide Sector 4. Air traffic control procedures allowed for a difference of up to 2 minutes between pilot and controller estimates without requiring a cross-check.
A short time later, the Melbourne Sector 4 controller assessed that the A320 passed over Portland at 1402 (13 minutes ahead of the estimate) but due to other duties, did not immediately notate the flight progress strip or coordinate this position with Adelaide Control.
At 1404, the Melbourne controller realised that he had not informed the Adelaide controller of the A320's Portland position report and commenced the coordination process. He reported the time at Portland as 1402 and the level as FL310. At this moment he realised that he had not made a calculation for the SUBUM estimate and used the estimated time of arrival as written on the flight progress strip as his revised estimate. This action resulted in his flight progress strip indicating that the A320 was going to be "on time" at SUBUM despite being 13 minutes early at Portland.
Communications for both aircraft crews were then transferred to Perth Flight Service on high frequency (HF) radio.
The Adelaide Sector 4 controller accepted the coordination from Melbourne Sector 4 on face value because the Melbourne controller was required to check the accuracy of data he was coordinating. The Adelaide Sector 4 controller checked his flight progress strips and noticed that the A320 was early at Portland but estimated to be "on time" at SUBUM. He considered that this discrepancy was probably due to a flight planning error that had been corrected by the Melbourne controller. His decision was influenced by the fact that both aircraft were estimated to be "on time" at SUBUM. He also considered that as they were vertically separated any error in the estimated times would not be significant, and chose not to pursue the matter any further.
The Adelaide Sector 4 controller then coordinated the Portland position details, including the estimated times of arrival at SUBUM, with Melbourne Sector 1, the control position for air route T134 from SUBUM. As a consequence, Melbourne Sector 1 had flight progress strips that indicated both aircraft being "on time" at SUBUM.
At 1444, the Perth Flight Service officer contacted Melbourne Sector 1 with the position report at SUBUM from the crew of the B737. This report included a request for a climb to FL310. As such a climb would negate vertical separation, the controller was required to establish a 10-minute longitudinal separation standard in order to approve the request. As the flight progress strips indicated estimates for SUBUM at 1444 and 1500, this standard appeared to have been achieved. However, because the B737 had only just entered his area of responsibility, he was required to check with the previous sector (Adelaide Sector 4) before authorising such a change. There had been a change of personnel at Adelaide Sector 4 and, as the oncoming controller also had the same time indications as the Melbourne controller, he agreed to the change and the crew of the B737 was instructed to climb to FL310.
After approving the climb, the Adelaide controller decided to check his radar display on the maximum range and saw that the A320 was only 30 seconds east of SUBUM. He immediately contacted the Melbourne Sector 1 controller to inform him of the confliction.
The Melbourne Sector 1 controller contacted Perth Flight Service and issued an instruction for the crew of the B737 to descend to FL290. Because of the amount of coordination required, it took almost 3 minutes to translate the Adelaide controller's observation into an acknowledged instruction for the B737 to descend. As the B737 had reached FL300 before the crew received the instruction to descend, and as the vertical separation standard was 2,000 ft, an infringement of the separation standards had occurred.
The Adelaide radar did not provide reliable coverage on air route T134. It normally gave coverage for aircraft over Portland and for a few miles east of SUBUM, but only provided intermittent coverage for the remainder of the route segment.
As air route T134 was predominantly outside VHF coverage, the route was designated for HF operation. The air traffic service unit responsible for HF communication was Perth International Flight Service. This limitation resulted in a delay in passing information between the controller and the crew of the B737 because the flight service officer was the intermediary for all radio transmissions.
The company flight plan for the A320 included a manoeuvring time of 10 minutes at the departure aerodrome. This time was estimated by the flight planner, using a computerised flight planning system, as that required for manoeuvring prior to setting course after the aircraft had become airborne. However, the flight plan indicated the time as an elapsed time from airborne to abeam/overhead Melbourne airport. The air traffic control strip printing system did not provide for a specific manoeuvring segment. Consequently, it added the 10-minute manoeuvring time to the 12-minute flight-planned time to the first enroute navigation aid.
Having added this 10-minute interval to the departure time, all the flight progress strips for the A320 therefore indicated position estimates which were approximately 10 minutes later than the corresponding flight-planned estimates.
The manoeuvring time was introduced by the company to allow for the time between becoming airborne and setting course, which varied depending on considerations such as runway direction, standard departure instructions and wind velocity. The flight crew had the same initial flight-plan information as air traffic control. However, the crew updated their plan to reflect the actual time at which the aircraft set course.
The B737 was operated by a different company and the flight plan did not include provision for a manoeuvring time. The flight progress strip estimates were therefore within the normal tolerance of plus or minus 2 minutes.
While under radar observation, pilots were not required to report their position as the controller could see that the aircraft was at the reporting point. If there was a difference between the estimated and actual times of arrival, then it was noticed by both parties and independently corrected. Under these conditions separation standards relied on radar.
When under procedural control, separation standards were based on estimated times of arrival at the reporting points and pilots were required to advise air traffic control of any variation of more than 2 minutes. Separation standards were devised to allow for these variations. Controllers crosschecked known times, as reported by pilots, with the flight plan time intervals to provide a check on the estimated time of arrival for the next position. If this crosscheck was within 2 minutes of the pilots estimate, no further action was required.
The air traffic control strip printing system's interpretations of the A320's flight plan led to a latent error in the flight progress strips for the A320 that was not present in the B737 strips. The system defence to negate this error was removed when the Melbourne Sector 4 controller did not update the flight progress strip estimated time for the SUBUM position.
Air traffic control procedures
Once the Melbourne Sector 4 controller had read the flight plan estimated time for the A320 at SUBUM, the coordination procedures were such that this incorrect estimate was passed to Adelaide Sector 4 and then from Adelaide Sector 4 to Melbourne Sector 1 without further check. Consequently, when coordination was required to allow the climb of the B737, the two controllers concerned had incorrect information. They did not question the information because it had been based on a radar observation.
Air traffic controllers
Although the flight progress strips showed that a longitudinal separation standard existed, the Melbourne Sector 4 controller did not review the situation before transferring the aircraft to Perth International HF. However, a scan of the strips should have revealed the error because the flight-planned time for the route segment was 45 minutes which, when added to the known time over Portland, would have given an estimate for SUBUM of 1447 rather than 1500.
The decision by the oncoming Adelaide Sector 4 controller to check his longer range radar display, detected the actual position of the A320 on air route T134 and enabled him to implement remedial action.
The aircraft passed over the last radar-observed position and the crews would have made estimates for SUBUM, but were not required to report those estimates to air traffic control. The controller also made estimates for SUBUM and based a procedural separation standard on those estimates. He had no need to check those estimates with the pilots or any other controller. Consequently, when the error was made, there was no crosscheck with which to provide a safety net.
Both flight crews were operating within 2 minutes of their corrected estimates, based on their actual times at Portland, and therefore were not required to report any minor changes.
Consequently, the crews and the controllers were working a procedural standard from two different time bases, neither of which had been crosschecked with the other. In a procedural environment, pilots and controllers must have a single datum on which to base their reporting and separation.
- The flight plan for the A320 contained a manoeuvring time for the aircraft prior to setting course.
- The air traffic control strip printing system was unable to allow for a discrete manoeuvring time in the strip preparation.
- The Melbourne Sector 4 controller did not conduct a crosscheck calculation on the flight progress strip notation for the A320's estimated time of arrival at SUBUM.
- The Adelaide Sector 4 and Melbourne Sector 1 controllers did not initiate any check action for the observed significant difference in the time interval for the A320 between Portland and SUBUM.
- Air traffic control procedures were such that there was no assurrance that the flight crew's estimated times were the same as those being used by ATC to provide procedural separation.
As a result of the investigation, the company operating the A320 has amended its flight-planning process by making a modification to the flight-planning system. In the 6 months following the amendment's implementation, manoeuvring times remained within the working tolerances of position reporting.
As a result of this occurrence, the Bureau of Air Safety Investigation is currently investigating a perceived safety deficiency. The deficiency identified relates to the validation of air traffic services' flight progress strip data for aircraft operating outside radar coverage.
Any safety output issued as a result of this analysis will be published in the Bureau's Quarterly Safety Deficiency Report.
|Date:||17 August 1997||Investigation status:||Completed|
|Time:||1447 hours EST|
|Location:||6 km S Subum, (IFR)|
|State:||South Australia||Occurrence type:||Loss of separation|
|Release date:||01 December 1998||Occurrence category:||Incident|
|Report status:||Final||Highest injury level:||None|
Aircraft 1 details
|Aircraft manufacturer||The Boeing Company|
|Type of operation||Air Transport High Capacity|
|Damage to aircraft||Nil|
|Departure point||Sydney, NSW|
|Departure time||1226 hours EST|
Aircraft 2 details
|Aircraft manufacturer||Airbus Industrie|
|Type of operation||Air Transport High Capacity|
|Damage to aircraft||Nil|
|Departure point||Melbourne, VIC|
|Departure time||1335 hours EST|