VH-CZA, a Boeing 737, was inbound to Brisbane and was being sequenced to follow VH-EWM, a BAe 146, for a visual approach to runway 19 via left base. Both aircraft were under the control of the Brisbane Approach radar controller and were being processed via the Coolangatta One Standard Terminal Arrival (STAR). VH-TNU, a Dash 8, was behind CZA and was being processed via the Jacob's Well Four STAR.

The Brisbane Flow Controller had originally assigned TNU to land on runway 14. This required the controller to position the aircraft east of the Jacob's Well - Brisbane track, and for the crew to track direct to the Brisbane Control Tower before joining right base for runway 14. There were two light aircraft operating to the south and southeast of the aerodrome at 1,500 ft on traffic survey work. The controller assessed that these aircraft may have conflicted with TNU overflying the aerodrome for runway 14. As well, there was departing traffic that he considered might also have conflicted with TNU being processed for runway 14. As a result, the controller decided to process TNU for runway 19.

The controller vectored TNU north to a position approximately abeam and above CZA and informed the crew that they were following the traffic which was low and to their right. The crew of TNU advised that they were still in cloud. At this time, TNU was approaching 5,000 ft heading 360 degrees M and CZA was descending through 3,500 ft, heading 310 degrees M. The controller then asked the crew of CZA to fly a close left base, and advised that they were following a blue and white BAe 146 aircraft (EWM) which was over St Helena Island on left base for runway 19. A short time later, the crew of TNU reported visual. The controller acknowledged this transmission and then instructed EWM to contact the tower.

After making a number of transmissions to other traffic, the controller instructed TNU to turn left heading 310 and to descend to 2,500 ft. He asked the crew if they still had the B737 in sight, adding that it was in their 10 o'clock position at 3 NM. The controller instructed CZA to descend to 1,500 ft. The crew of CZA then reported visual and were cleared by the controller for a visual approach via a close left base. After CZA acknowledged this clearance, the crew of TNU reported that they had the traffic sighted. The controller responded that CZA was for a close left base and cleared TNU for a visual approach. He then instructed CZA to contact the tower and asked the crew of TNU to closely follow CZA. The controller then processed a departing aircraft and accepted the transfer of another inbound aircraft onto his frequency. A turn instruction he gave this aircraft was not initially understood and required two further transmissions for clarification. As this exchange concluded, the controller asked the crew of TNU to confirm that they still had visual contact with the B737. The crew responded that they believed that the B737 had landed. The controller immediately instructed TNU to turn right. He then confirmed with the tower controller that TNU was turning away from CZA. CZA subsequently landed normally on runway 19 and TNU was processed for landing on runway 14.

Recorded radar data indicated that the minimum separation between the aircraft was about 200 ft vertically, and 0.8 NM horizontally. This occurred as TNU was instructed by the approach controller to turn away from CZA. The required separation standard in the circumstances was 1,000 ft vertically or 3 NM horizontally. The data also showed that, at the time CZA was given as traffic to TNU, the groundspeed of CZA was 143 kts while that of TNU was 243 kts. When the crew of TNU reported sighting the traffic, the position of CZA relative to TNU was between 7 and 8 o'clock at about 3 NM rather than 10 o'clock as the controller had indicated a short time earlier. At this time, EWM was at about a 10 o'clock position relative to TNU but at a range of 7 NM.

A review of the automatic voice recording of communications between the approach controller, the aircraft, and other agencies confirmed that there was no request from the control tower for TNU to be processed for runway 14. It also confirmed that the crew of TNU was not given their number in the landing sequence.

Neither TNU, nor CZA, was fitted with a traffic collision avoidance system (TCAS).

The controller

The controller involved was highly experienced. He gained his initial ATC ratings in June 1980. He had been rated and endorsed to perform the radar approach controller duties since April 1990.

The incident occurred two hours after he had commenced duty on the Approach South position. The controller stated that he did not consider the traffic levels and/or workload to be unusually high at the time of the incident. He indicated that he changed the traffic sequence with respect to TNU to reduce his workload and improve traffic management for both himself and the tower.

Information from the crew of VH-TNU

The technical crew of TNU said that, in hindsight, they did not sight CZA at any stage during the sequence. Their report to the controller that they believed the aircraft had landed indicated that they had mis-identified EWM as CZA. The vectors they had been given placed the aircraft high and fast on the descent profile. This, coupled with the checklist actions and the request to sight the traffic, created a high cockpit workload situation. Neither pilot could recall being given their number in the landing sequence.

The cabin flight attendant on TNU became aware of CZA when she saw it through a passenger window as she completed the cabin pre-landing checks. She felt some concern regarding the proximity of the aircraft and considered entering the cockpit to confirm that the pilots knew of its presence. In the event, she decided to take no action, as she felt confident that they would have been aware of the situation.

Brisbane arrival procedures

The management of air traffic in the Brisbane Terminal Area (TMA) involved the formulation of a traffic management plan for arriving and departing aircraft. The Terminal Approach Coordinator (TAC) decided upon the plan after liaison with the Traffic Management Coordinator (TMC) in the control tower. Its purpose was to determine the most efficient use of available runways and types of approaches consistent with prevailing weather conditions and traffic density and patterns. Once the plan was agreed, the TAC directed the Flow Controller to sequence the arriving traffic in accordance with the traffic management plan. The task of the radar approach controller was to direct aircraft in accordance with the plan to ensure that aircraft arrived in sequence at appropriate intervals for efficient traffic flow.

In order to make the best use of the available runways and minimise delays, procedures had been developed to assist controllers with arriving aircraft required to overfly the aerodrome from the south for runway 14. These procedures were detailed in Northern District Local Instructions TMA 27, and were designed to assist in minimising workload on both the approach and control tower controllers by providing a standardised technique which allowed departing aircraft to depart and remain clear of the arriving overflying traffic. The instruction stated:

In the case of RWY 19/01 and an arrival from the east/southeast for landing RWY 14, the following shall apply. The arrival shall be positioned east of the JCW-BN VOR track, tracking direct to the CONTROL TOWER assigned A020 and for the ACFT to be on TWR frequency no later than 5NM with no restrictions.

These tracking requirements will ensure that a RWY 19 LAV SID can proceed unrestricted.

Sight and Follow procedures

At the time of the occurrence, visual separation standards were detailed in the Manual of Air Traffic Services (MATS) Chapter 4, Section 8. Air traffic controllers could use the visual separation standard to separate aircraft flying at or below FL125. However, if this standard was used an instruction to maintain visual separation with, or to follow other aircraft, should have been issued in accordance with requirements of MATS Chapter 6 Section 4. Chapter 8 of MATS, "Enroute/Approach Control", did not make reference to sight and follow procedures.

A Fairchild SA227-AC (Metro) aircraft landed on runway 34R, with an instruction to exit onto taxiway Tango 1 and taxi via taxiway Juliet. The pilot contacted the surface movement controller (SMC) while approaching a mandatory holding point on taxiway Juliet, and was instructed to taxi via taxiway Bravo and hold short of runway 25. The pilot read the instruction back correctly. That instruction required the pilot to continue ahead on taxiway Juliet, passing the entry to taxiway Charlie before turning right into taxiway Bravo.

The tower controllers subsequently observed the Metro taxiing along taxiway Juliet, past the entry to taxiway Bravo, towards runway 34L, which lay a short distance beyond taxiway Bravo. At that time, a Boeing 737 (B737) was touching down on runway 34L. The SMC instructed the pilot of the Metro to stop. As there was no reply, the instruction was immediately repeated, and subsequently acknowledged. The Metro was stopped at the edge of runway 34L, beyond the runway holding point. The pilot of the B737 was also instructed to stop immediately, but the aircraft was too close to the taxiway Juliet intersection to comply. The B737 was steered to the left of the runway centreline to increase separation from the Metro, subsequently passing about 25 m in front of that aircraft, at about 80 kts.

The occurrence happened at night in fine conditions; there were no works taking place in the area; and all relevant taxiway lights, runway holding point lights and movement area guidance signs were reported by the air traffic controllers to have been functioning normally. However, when subsequently interviewed, both pilots asserted that the runway 34L holding point lights on taxiway Juliet were not illuminated. Shortly after the incident an airport safety officer had been requested to conduct an inspection of the ground lighting in the occurrence area. All lighting, including the holding point lights on taxiway Juliet protecting runway 34L, were observed to be functioning normally.

An investigation established that the pilot flying the aircraft was undergoing command training under the supervision of a training captain. A practice Instrument Landing System (ILS) approach had been flown, during which the pilot under training experienced difficulties in maintaining the required flight tolerances. After landing, the pilot correctly obtained and read back the taxi clearance. However, he subsequently failed to notice that the aircraft had passed taxiway Charlie and was approaching the entrance to taxiway Bravo. Believing he was now approaching taxiway Charlie, the pilot continued to taxi along taxiway Juliet, expecting to see the entry to taxiway Bravo.

At about that point, the training captain, noticing that the aircraft had just passed the entry to taxiway Bravo, looked towards the pilot to ask where he was going, but immediately noticed the lights of a landing aircraft on runway 34L. Concerned about the proximity of the other aircraft, he immediately instructed the pilot under training to stop, which coincided with similar instructions from the SMC.

During discussions with the crew of the Metro, it became apparent that, after landing, the pilot under training had been concerned with his performance during the practice ILS approach, and had initiated a brief discussion with the training captain at about the time the aircraft had been taxiing along taxiway Juliet.

Both pilots, who were very familiar with the layout of Sydney Airport, also indicated that the movement area guidance signs (MAGS) were somewhat confusing when indicating the entry to taxiways. However, the pilot under training said that normally he did not refer to the MAGS because of his familiarity with the taxiway layout.

The handling pilot was undergoing a scheduled 6-monthly proficiency check in the Learjet, under the supervision of an experienced instructor who occupied the right control seat. The flight was planned to depart from Essendon and proceed to the Ripley locator, for an entry into the holding pattern, before making a locator approach to Avalon. Three touch-and-go circuits were planned at Avalon, including one 700-ft asymmetric low-level circuit following a simulated engine failure at V1. The aircraft was then to return to Essendon via the Plenty locator for an instrument landing system (ILS) approach.

The instructor briefed the exercise in detail and, because the pilot under check had low experience in the Learjet, it was decided that the instructor would handle all radio communications and conduct the necessary check sequences. The instructor also indicated that it was company policy for the aircraft be landed with less than full flap during an asymmetric landing. The aircraft was to be landed with flap 20 at Vref +10 kts.

The flight proceeded normally up to the point of the simulated engine failure at V1. When the instructor simulated a right engine failure by placing the thrust lever in the IDLE position, the aircraft drifted to the right. Intervention by the instructor enabled the aircraft to return to balanced flight. The handling pilot continued to carry out the 700-ft circuit in the after-take-off configuration of gear UP and flap 8. Flap 20 was selected during the base turn. The aircraft was flared normally with both thrust levers in the IDLE position. As the aircraft settled, a slight vibration was noticed, and both pilots became aware that the landing gear was still selected UP. Go-round power was applied and the aircraft climbed away. The landing gear was cycled normally and the aircraft returned for a full stop landing. A subsequent inspection of the aircraft showed that the only evidence of a runway strike was abrasion of the lower fuselage mounted very high frequency (VHF) blade antenna.

A subsequent investigation revealed that the pilot under check was allocated one hour of Learjet flying every 3 months. The instructor had selected the flight sequences to give the pilot the maximum handling exposure in the limited time available. In doing so, the normal two-crew, challenge and response routines were abandoned and the checks had to be accomplished by the instructor alone. The instructor became distracted by the asymmetric handling issues, and the demands of the low-level circuit, subsequently forgetting the relevant downwind and pre-landing checks. The handling pilot, who had been absorbed with controlling the aircraft, had lost situational awareness and did not notice the lack of check procedures by the instructor, or the lack of a positive gear-down indication.

The approach had been carried out with flap 20 extended, rather than full flap (flap 40), because the operator had previously experienced a partial loss of control during training when attempting an asymmetric go-around at flap 40. However, with flap 20 selected, the landing gear warning system was inhibited, contributing to the late realisation that the landing gear was not extended. The flap 20 asymmetric approach configuration was not in accordance with the manufacturer's flight manual recommendations.

A Boeing 767 (B767) was en route from Auckland to Hong Kong on air route B333. The crew had been cleared to conduct a cruise climb from FL310 to FL330. While FL310 was a standard level for the track being flown, FL330 was not. A Boeing 737 was en route from Sydney to Port Vila on air route B580. The crew of this aircraft had been cleared to, and was maintaining FL330, which was a standard level for their track. The routes cross at LEMIB a waypoint located approximately 490 NM east of Brisbane. The aircrafts' estimates for LEMIB were within three minutes of each other.

Five minutes prior to reaching LEMIB the crew of the B767 received a traffic alert and collision avoidance system traffic advisory warning. The traffic alert and collision avoidance system indicated that an aircraft was passing from left to right 900 ft above the level of, and 12 NM ahead of the B767. The required separation standard was 2,000 ft vertically.

Air traffic controller training was being conducted at the control position responsible for the area at the time of the occurrence. Following a query by the B767 crew regarding the crossing aircraft to air traffic control and a check of the flight progress strip data it was established that a separation standard had not been applied between the aircraft. There was a breakdown of separation.

Organisational aspects

The trainee was a group leader in the Brisbane Centre responsible for the management of the Tops/Ocean/West group. The group was divided into a number of sectors. Tops consisted of sectors 11, 11k and 12. West consisted of sectors 5 and 5D while Ocean had sectors 8, 8O and 10. Sectors 5D and 8O were radar sectors while all the other sectors were procedural control sectors except sector 11K which was a composite radar/procedural sector.

The group leader had previously held ratings and endorsements for all Tops sectors but had let two endorsements lapse due to workload and retained only the sector 11K endorsement.

The Ocean sectors were to be the first of the group to transition to The Australian Advanced Air Traffic Control System. The Ocean sectors had limited experienced staff plus there were other concerns for management that they believed could be suitably handled by the group leader becoming part of the Ocean teams. He also believed that if he obtained Ocean sector ratings he would be able act as a relief controller if required. Consequently, it was agreed that the group leader would transition with the Ocean sectors.

The group leader intended to undergo training to obtain the appropriate endorsements for the Ocean sectors. On completion of training he would function as one of the three Ocean team leaders as well as retaining group leader responsibility for those sectors. The En-route manager was to assume group leader responsibilities for the Tops and West sectors during the transition. To enable the group leader to commit himself totally to the training, an interim group leader was to be appointed to manage the group. The interim leader took over the group in the week immediately prior to the occurrence.

The Ocean sector staff had, and were, undergoing considerable changes due to: the reorganisation of the sectors; developments relating to contingency plans for Papua New Guinea airspace; management of Honiara airspace; and the transfer of the provision of directed traffic information from flight service international to the sectors and changes required for The Australian Advanced Air Traffic Control System transition. There were also a number of staff management issues that limited the options available to the group leader to meet operational demands.

Training plan

The Ocean sectors had a comprehensive training plan with workbooks for trainees and a training guide for the controller supervising the training. However, the supervisors training guide was unapproved and was in draft form only. The training guide and workbook were developed with the intention of providing a basis for ab-initio training. Training normally commenced with the trainee undertaking four weeks familiarisation at the sector operating positions followed by four days of classroom work, approximately one month in the simulator and finally six weeks of on the job training. During the on the job training phase the trainee would be assessed for competency by the supervising controller.

As the group leader already held radar and procedural ratings and an endorsement for sector 11K he elected to rationalise the ab-initio training program to reduce the time spent training. This aspect was in accordance with Airservices Australia's Civil ATS Operations Administration Manual which described "refresher training" for controllers as:

"training which focuses on change in a person's required competence and includes training concentrating on what a person once knew; what a person should (or does) know but hasn't been applying; and what a person hasn't had an opportunity to apply."

The Civil ATS Operations Administration Manual also detailed "local training strategies which recognise refresher training responsibilities include (but are not limited to):

1. personal study of material, including videos, collated and provided by at the direction of team leaders or line mangers;
2. use of simulator sessions, either specific to refresher training topics or included in ongoing training; and
3. use of computer based training (CBT)."

The group leader had reviewed the training material for sector 8 but had not undertaken any classroom or simulator training for the sector. Due to other training commitments the simulator was not immediately available. He planned to join the scheduled Ocean simulator-training program at lesson ten (three weeks after the date of the occurrence) and to complete the lessons through to 23. During the interim period the group leader was to undertake familiarisation training under supervision. He had discussed his training plan with one of the Ocean team leaders responsible for training. There was no process to assist rated controllers intending to train for other ratings/endorsements to develop a suitable training plan.

Supervising controller

A supervising controller had been appointed to oversee the training of the group leader. The group leader's training commenced two days prior to the occurrence. During that time he and the supervising controller had spent approximately three hours together on the first day, none on the second and three hours on the morning of the day of the occurrence. The interruptions to the group leader's training were due to a prior commitment on the first day and the lack of a training position as a result of another controller undergoing familiarisation on the second day.

After spending three hours together on the morning of the occurrence the group leader and the supervising controller conducted a handover of sector 8 to enable them to take a rest period. Subsequently, the supervising controller returned to the console and assumed responsibility for sector 8O, the radar sector.

At 1100 Eastern Standard Time the group leader was waiting at the console expecting the next rostered controller to takeover sector 8O to enable the supervising controller to train the group leader on sector 8. A recently arrived controller offered to supervise group leader on the console instead of taking over sector 8O. Subsequently, this controller and the group leader assumed responsibility for sector 8. The controller had recently completed a stint as a supervising controller for another trainee and had considerable training experience. The nominated supervising controller for training the group leader remained on sector 8O.

The controller was not aware of how much training the group leader had completed and was therefore not aware of the latter's level of knowledge in the position. They did not conduct a pre-training brief to discuss learning aspects to be addressed during the session.

Sector 8 was the only procedural sector that used controller pilot datalink facilities to communicate with flight crews. Between 10% and 20% of all the flights operating through the sector used the controller pilot datalink. As the group leader was unable to operate the controller pilot datalink the controller was required to operate the facility. The controller was also directing and explaining the operation of sector 8 to the group leader while reviewing the flight progress strips for conflictions.

The controller and the group leader were at the position for approximately two hours during which the number of aircraft under their control steadily increased. The group leader noted a number of procedures that were different to what he expected to experience and he found it increasingly difficult to maintain an appreciation of the traffic situation. During this period it became apparent to the controller that the group leader was unfamiliar with the operation of the position.

Just prior to the occurrence the group leader received a request for a clearance from the pilot of an aircraft to enter the oceanic control area on climb to FL250. Because of the proximity of the aircraft to other flight information region boundaries the controller had to coordinate with Nadi and Auckland Centres prior to issuing a clearance. At about the same time, a flight information region boundary position report was received on the controller pilot datalink from a flight. The controller should have transferred this flight to Nadi Centre prior to the position but had not instructed the crew to transfer. Consequently, the position report had to be coordinated by voice intercom.

Flight progress board management

Details of the B737 were annotated on a blue flight progress strip and the sector 8O controller had coordinated the flight at FL330. The B767 flight was on a buff flight progress strip. Blue and buff coloured flight progress strips were used to differentiate between eastbound and westbound flights respectively. Coordination on the B767 was received from Auckland Centre and a clearance for the crew to conduct a cruise climb in the block level FL310 to FL330 was concurred. Requests by crews to conduct a cruise climb were regularly received and approved by the sector 8 controllers.

The recognition and resolution of potential conflicts at positions in the sector where tracks crossed was a regular part of a controller's task. The crossing point of the aircrafts' tracks, LEMIB was not annotated on either flight progress strip. After flight progress strips were activated by a departure report or coordination from another air traffic service unit they were passed to the sector 8 controller for placement on the board. The controller would place flight progress strips under an appropriate designator and then review all flight progress strips for conflictions. Conflicting flight progress strips would then be "cocked" to highlight a problem for subsequent action. Controllers "cocked" a flight progress strip by partially moving a strip out of the display bay.

If the controller identified that there was a potential conflict at a crossing point like LEMIB, a separation standard would be calculated and the respective flight progress strips would be annotated with the position. Once the separation procedure had been applied the flight progress strips would be returned to the bay. In the occurrence the flight progress strips for the aircraft were placed in the bay under the same designator without being cocked. The investigation did not establish why the flight progress strips were placed in the bay without being actioned.

Controller pilot datalink facilities

When the controller pilot datalink had been introduced a simulator had been provided to assist in the training of controllers. The controller pilot datalink simulator had subsequently been removed from the centre following training of the initial group of controllers required to operate the system. There were no facilities to assist controllers to learn how to operate the system prior to conducting the on the job training phase of their training. There was no controller pilot datalink reference material in the training guides. The group leader was not familiar with the operation of the controller pilot datalink.

With the introduction of the controller pilot datalink, two additional positions on sector 8 had been established for a controller to specifically operate the facility during the morning and afternoon shifts. At the beginning of 1998 these positions were disestablished to provide an additional team leader and to enable a controller to undertake operational development tasks. Sector staff had discussed the decision to remove the dedicated controller pilot datalink controller and believed that a controller could adequately manage both the controller pilot datalink and the sector 8 position. There was no workload or safety review conducted prior to amending the roster.

Team management

On the day of the occurrence there were no team leaders rostered for duty to oversee the management of the sector. A full performance controller was fulfilling the team leader functions. This controller had not completed team leader training.

Sector 8

Sector 8 was a procedural sector managing the Class A oceanic controlled airspace east of the coast of Australia, extending to the flight information boundary with New Zealand and Nadi, from just south of Tasmania to a line joining Brisbane and Port Vila. The operation of the sector was considered by controllers to be significantly different from other procedural sectors because of the inclusion of the controller pilot datalink and the different separation standards and procedures used for oceanic control compared to sectors over continental Australia.

Operating position

The sector 8 operating position consisted of a console with a flight progress strip display with the controller pilot datalink keyboard and monitor located on a table on the left side of the console. The sector 8 controller operated the controller pilot datalink and managed the separation of aircraft using flight progress strips placed under location designators on the board. The layout of the facilities required the controller to turn 90 degrees to the right each time the controller pilot datalink was operated.

The sector 8O radar position was located to the left of the procedural console and a map display was fixed to the floor in front of, and midway between both operating positions.

The layout of the operating positions and the additional facilities only enabled a single controller to sit and operate sector 8. Controllers supervising training were required to sit or stand behind the trainee.

The De Havilland Dash 8 was tracking inbound to Tamworth via the 040 radial of the Tamworth VOR (a radio navigation aid) at 3,500 ft in accordance with air traffic control instructions. Meanwhile, a TB10 Tobago was about to depart Tamworth via the 018 VOR radial, a track that would require the Tobago to turn across the inbound track of the Dash 8. Air traffic control had decided to limit the initial climb of the Tobago to 2,500 ft in order to ensure vertical separation with the Dash 8.

The aerodrome control task was being performed by a trainee controller under the supervision of a rated controller. The pilot of the Tobago was instructed to "line-up" and given an altitude restriction of 2,500 ft, which was read back by the pilot. However, the readback was not clear and the aircraft subsequently took off with both the trainee controller and the rated controller uncertain as to the actual altitude read back by the pilot, who was from a non-English speaking background and had an accent that was, on occasions, difficult to understand. It was subsequently determined that the pilot of the Tobago had understood the controller to say 3,500 ft, and that was the altitude he had read back to air traffic control. The crew of the Dash 8 were given traffic information on the position of the Tobago by air traffic control and commenced a look-out for that aircraft.

Because both controllers were unsure that the 2,500 ft restriction had been correctly acknowledged, it was agreed that the trainee controller would request the pilot of the Tobago to confirm that he was maintaining that altitude. Before that could take place, the pilot of the Tobago asked the controller to confirm his assigned altitude. Again, the transmission was not easy to understand. The trainee controller confirmed 2,500 ft and passed traffic information on the Dash 8 at the same time. The only reply from the pilot of the Tobago was the word "affirm" and his callsign.

In fact, the Tobago had been maintaining 3,500 ft, and it was that later transmission from air traffic control that made the pilot realise he should have been maintaining 2,500 ft. He commenced an immediate descent from 3,500 ft, but did not make any radio transmission to that effect.

The crew of the Dash 8 heard the exchange between air traffic control and the pilot of the Tobago. At almost the same time, they saw the Tobago about 400 m ahead, at the same level. They commenced an immediate descent then realised the Tobago was moving to their right. The aircraft passed with a horizontal separation of approximately 200 m, and no discernible vertical separation. The required standard was 1,000 ft vertical separation until the aircraft had passed.

While flying a coupled instrument landing system (ILS) approach to runway 35 at Canberra, the crew received a ground proximity warning system (GPWS) "terrain" warning at 470 ft above ground level. As the aircraft was in visual conditions at the time, the crew elected to continue the approach.

The crew reported that, at the time of the warning, the aircraft was established on the ILS glideslope and descending at 650 ft/min. However, the T-VASIS indicated one dot low. Immediately prior to the warning, the radio altimeter indication rapidly increased, and then decreased, apparently triggering the warning.