BEECH 400

Full Narrative

On April 13, 2011, at 1222 Central daylight time, an air traffic control operational error and near midair collision occurred at the Midland International Airport (MAF), Midland, Texas, when TONE97 flight, consisting of two T-1 Jayhawk U.S. Air Force training aircraft (civilian-type Beech BE40 turbojets), lost separation with N631ME, a Raytheon-Beech King Air 300, on final approach to runway 28 at MAF. There were no injuries reported, and no damage to any of the involved aircraft.

MAF is a civil airport that is heavily used for training by Air Force aircraft based at Laughlin Air Force Base, located about 170 miles southeast of MAF. During the period surrounding the operational error, there were several different USAF flights operating in the MAF traffic pattern. TONE97 was a flight of two T-1 Jayhawks that entered the airport area on a left downwind approach to runway 28. Initial contact with the radar approach controller was routine, and at 1211:50, the approach controller advised the local controller that TONE97 was two aircraft. At 1217:35, TONE97 was switched to tower frequency and performed a communications check to ensure that the flight wingman was also on tower frequency. After completing the check, TONE97 contacted local control at 1218:02, reporting, “…5 mile initial.” The controller responded, “…left break midfield, descent in the break approved, runway 28 cleared for the option, traffic is a company Beech about three mile final full stop.” TONE97 acknowledged the clearance. At 1219:19, the pilot of N631ME checked in on tower frequency reporting on a visual approach to runway 28. At the time of the report, N631ME was on a wide right downwind about 7 miles north of the runway. The controller acknowledged the call and instructed the pilot to, “…continue for the runway.”

According to recorded radar data, TONE97 commenced the left break maneuver at 1220. As part of the break procedure, the wingman (TONE97-2) dropped back to a position about 1.4 miles behind the lead aircraft. Both aircraft entered close-in left downwind for runway 28. TONE97's transponder was on and the aircraft was displayed as a beacon target. TONE97-2’s transponder was off, so the aircraft produced only a primary radar target with no altitude reporting information available. According to the local controller, N631ME and both aircraft in TONE97 flight were visible from the tower.

At 1221:13, the local controller transmitted, "King Air 1ME you'll be number 3 for the field following a flight of two Beechjets currently about a mile on the left base, report the traffic in sight." The pilot replied, "...uh – I’ve got one of them in sight." The controller continued, "There’s one behind him a little bit higher just south of runway 34R." The pilot did not acknowledge the second transmission, and the controller responded to an unrelated coordination call from the approach controller. Recorded radar data showed that TONE97-2 and N631ME were converging on opposite base legs about 1.5 miles from runway 28. At 1221:53, the local controller instructed N631ME to, "...fly a little bit south and uh do you have the second aircraft in sight?" The pilot responded, "King Air 631ME's waving off, this is entirely ridiculous. We’re going to do a 360 if that’ll work for you." The controller instructed the pilot to make a right 360 and rejoin the final.

At 1223:09, an unidentified voice similar to that of N631ME transmitted, "...that was a perfect recipe for a midair." Following the incident, both aircraft in TONE97 flight and N631ME all landed safely on runway 28.

Radar Data

According to recorded radar data, the closest point of approach between TONE97-2 and N631ME was about 0.3nm. Graphics showing the flight tracks of the aircraft involved in this incident have been entered into the docket.

Personnel Interviews

MAF Local Control Trainee

At the time of the incident, training was in progress at the MAF Local Control (LC) position. The LC trainee had been previously certified at Kansas City Downtown tower in 2008, and transferred to MAF in May 2010. He was certified on the flight data, ground control, and clearance delivery positions. His work schedule included two night shifts and three day shifts. He described traffic at Midland as changeable from day to day, and noted that occasionally the peak traffic can move from the day shift to the night shift. It is very dependent on the military flight schedule.

The LC trainee had completed over 90 hours of training on the LC position. He had not yet received any instruction on the approach radar positions, but stated that he did try to obtain observation time in the radar room to familiarize himself with what the radar controllers do and how they sequence aircraft to the airport. Observation time was not documented on any training forms, as the activity was voluntary and not part of his formal training program.

The LC trainee described his training session as, "going fine, with light traffic." He recalled a pair of T-1 aircraft entering the pattern on an overhead approach. He stated that he instructed the pilots to break midfield, but the flight actually conducted their break further toward the departure end of the runway. As the two aircraft completed their overhead approach, the LC trainee stated that he advised the pilot of N631ME about them. The pilot responded that he had one of the aircraft in sight, but did not see the other. When the first T-1 was arriving at the runway 28 threshold, the second T-1 was on the left base. the LC trainee said that he instructed the pilot of N631ME to fly southbound if he did not have the second T-1 in sight, but it was too late because N631ME had already turned final. The LC trainee stated that his instructor on the position was standing right next to him, but did not intervene or provide any advice.

The LC trainee stated that his objective when sequencing arrivals conducting overhead approaches was to do the sequencing as early as possible. The radar controllers were expected to sequence aircraft to the runways, but the local controller was responsible for sequencing pattern traffic. He stated that he instructed the T-1 pilots to break midfield, and continued that if the pilots had actually executed the break at midfield, "...it would have worked out."

The LC trainee stated that he responded to the initial call from N631ME with an instruction to "continue" instead of advising the pilot that he was number three behind traffic because he was having a hard time breaking the habit he developed at the Kansas City tower, which has only class D airspace, of waiting to sequence aircraft until they are close in to the airport. His trainers at MAF were trying to get him to complete his sequencing earlier. The LC trainee noted that operations in class C airspace such as that surrounding MAF were different from operations in class D airspace, and that controllers had more separation responsibilities.

The LC trainee had used a STARS tower radar display at MKC, where the tower radar display was typically set to a range of approximately 12 to 13 miles. At Midland, the tower radar display range was typically set at 20 or more miles. That range setting was required by the local standard operating procedures manual. He stated that he had been told by his trainers that at MAF the radar range was set to a higher value in order to assist in sequencing traffic. Aircraft at Midland include military and air carrier jets that move faster than the traffic operating at MKC. The LC trainee stated that the approach controller’s planned sequence was normally clear from looking at the tower radar display, but if training was in progress in the radar room it was sometimes harder to tell what the plan was.

The LC trainee was shown a radar replay of the incident in order to refresh his memory of the event. He stated that as the King Air and the T-1 converged, he was looking at the two aircraft out the window and noted that the T-1 was higher than the King Air. He did not expect the King Air to turn inbound toward the airport without having seen the second T-1. When he and his instructor saw both aircraft turning onto the final approach course, their intent was to break one of the aircraft out to resequence it. When they discussed the incident later, the trainee said that the instructor's plan was to turn the T-1 out, while his own plan was to turn the King Air. Before they had a chance to act, the King Air pilot reported that he was going to make a 360 degree turn for spacing.

The LC trainee stated that when applying tower visual separation he was required to have both aircraft in sight and to resolve conflicts by issuing control instructions.

The LC trainee stated that using low approaches as a mechanism for resolving poorly executed sequencing situations was not a technique he had been taught or used. Though he had not yet begun radar training, he did have to complete a radar qualification test to work in the tower. He stated that when adjusting the sequence of inbound aircraft, he was only required to coordinate with the approach controller if his actions would change the sequence. Minor spacing adjustments did not need to be coordinated.

After the incident occurred,he was not involved in any debriefings or other discussions about it except between himself and his trainer. Their conclusion was that "assumption bit us" in that they did not expect the King Air pilot to turn toward the runway on his own.

Local Control Instructor

The LC instructor began working for the FAA in 1988 and transferred to Midland in August 1991. His schedule included Fridays and Saturdays off, but he typically worked every Friday on overtime. He had been working six day weeks for approximately the last two years. He said that it was difficult to get time off except for occasional hour or two at the end of a shift. Taking an entire day off on leave was very difficult.

On the day of the incident, the LC instructor recalled no equipment or operational issues, or other distractions. At the beginning of the shift, he provided two sessions of on-the-job training for a trainee in the radar room, and then came up to the tower cab to begin training on the local control position. When the incident occurred, he and his trainee had been on the LC position for about 50 minutes. The incident occurred in the last 10 minutes of that session. TONE97 called in and was following another T-1. The LC trainee cleared TONE97 for the overhead approach and told the pilot that he was "cleared for the option" on runway 28.

Shortly afterward, N631ME called and was instructed to "continue." TONE97 passed over the airport and commenced the break. As the aircraft proceeded outbound on the downwind leg, the LC trainee advised the pilot of N631ME about the two T-1s also operating in the pattern. The King Air pilot stated that he did not see the second aircraft. The LC instructor said that he was about to intervene and tell the King Air pilot to fly through the final approach course for spacing when they received a coordination call from the approach controller about an aircraft requesting a surveillance approach. During the call, the King Air was turning inbound on final. The LC instructor stated that he was about to instruct the second T-1 in the flight to climb, but saw that the T-1 was already above the King Air. At that point the King Air pilot said something about the aircraft being too close.

The LC instructor said that he could see the King Air and the second T-1 out the window, and he did not believe that the two aircraft were in dangerous proximity to each other. His plan was to break out the second T-1 and instruct the pilot to climb, but he did not do so because the King Air pilot said that he would be executing a 360 degree turn. The LC instructor noted that the second T-1 was further in trail of the lead aircraft than would be normal for a standard formation flight.

Discussing overhead approaches, the LC instructor noted that under the former MAF standard operating procedures manual, the tower would sequence such aircraft with other inbound aircraft by controlling the timing of the pattern break of the overhead approach aircraft. In this situation, he believed it would have been appropriate for local control to have either put the King Air in front of the lead T-1, or to have instructed the T-1s to extend their downwind to fall in behind the King Air. It was LC's responsibility to do that.

Asked when he first realized that the sequence was not working out, the LC instructor said, "...when the King Air turned in on final." It caught him off guard when the pilot turned toward the airport without knowing the location of the traffic he was following. The LC instructor said that he did not believe he would have any trouble breaking out the second T-1, because the aircraft was definitely higher than the King Air. To some extent, the sequencing was left too late because he was trying to let the LC trainee "work it out" for training purposes. He stated that the situation went too far, but that under the circumstances he could have resolved the problem by climbing the second T-1 and turning it for spacing.

Asked about his familiarity with the LC trainee, the LC instructor stated that he had enough awareness of his background to be comfortable training him. It was his understanding that the trainee was at approximately 87 hours of training completed out of the 90 allocated from the position. At that point, the trainee's training would either be extended or terminated.

The LC instructor said that he was unable to recall the specific definition of tower visual separation other than that "...he was required to keep airplanes from running together." He stated that he has not had issues with keeping airplanes apart in the past.

Asked about use of the tower radar display to obtain information not available by looking out the window, such as the speed of arriving aircraft, the LC instructor said that instructors do teach about that. While this incident was occurring, he stated that he was trying to look out the window at the traffic and was not monitoring the radar. The pattern flown by the T-1s was wider than he expected. If he had been handling the position alone, he stated that he would have broken out the T-1s either sooner or later to establish spacing with the King Air.

The LC instructor said that the sequence between the King Air and TONE97 flight was not readily apparent when the King Air called. He noted that local control can change the sequence by adjusting what the aircraft do on overhead approaches. Local control at MAF had a lot of responsibility, almost like a miniature approach control. The local controller was responsible for sequencing traffic to all the runways, even with multiple runways in use.

MAF Front Line Manager (Supervisor)

The supervisor began working for the FAA in October 1998, and came to Midland as a controller in 2001. He became a supervisor in 2007. He was working in the radar room conducting a surveillance approach at the time of the incident involving TONE97 and was not immediately aware of it when it occurred. He stated that supervisors are occasionally required to provide aircraft with surveillance approaches because of staffing, and that it can take approximately 10 minutes to complete such an approach. Since the incident, he had reviewed a radar replay but had not heard the audio of the local control position. He was provided the opportunity to view an audio/visual replay of the incident as part of his interview.

The supervisor said that the handoff of the King Air to LC appeared to be fine as it took place, but the problem appeared to be that the local controllers did not manage the sequencing. The radar approach controller did not always know what LC was going to do with aircraft on overhead approach, so they depended on local control to determine the sequence of the arrivals. In the supervisor's opinion, the King Air should have been number one, with the T-1s following behind. The King Air’s groundspeed was high, and should have been a factor in the sequencing decision. Spacing was normally a cooperative decision between the tower and the radar controller. The tower controller had more flexibility to adjust spacing to the runways.

The supervisor said he was uncertain why the event occurred, but said that the instructor should have forced a sequence when the King Air called rather than allowing the trainee to just instruct the pilot to continue. He noted that the trainee had a control tower operator certificate and was certified at another tower before coming to Midland. Under those circumstances, the expectation would have been that the trainee should have been able to handle the problem. That assumption may have been what got the trainer in trouble. The supervisor noted that the instructor involved had about 20 years experience and was known as a good trainer.

The supervisor stated that he was working about 8 to 16 hours of overtime a month. Until March, he was working 10 hour shifts each day from Monday through Friday, and then eight hours of overtime on Saturdays once or twice a month. Many controllers at Midland were working six day weeks, but recently the supervisors have been attempting to minimize that and release controllers from the sixth-day overtime when there was a way to do so.

MAF had a very high proportion of trainees. The staffing problem started about five years ago when MAF lost numerous certified controllers to other facilities, and they were still trying to catch up.

Asked about use of the tower simulator in San Antonio for training of Midland controllers, the supervisor stated that Midland has been unable to afford to send people there or to put the facility's data into the simulator database. He felt that if the tower simulator in San Antonio was configured for use by MAF, the facility would probably be able to send an instructor and three trainees there to use it.

Supervisors at MAF are responsible for monitoring the operation, as well as managing the training program, including authoring training materials, performing associated administrative tasks, evaluating performance, etc. The facility currently has three full-time supervisors and one temporary supervisor.

There are typically supervisors present in the tower and the radar room from 0900 until 1900, but that was a goal, not a requirement. The operation "tended to be better" when there was a supervisor present. Position audits typically showed improved phraseology and procedures. The facility had also conducted briefings on expectations for the controller-in-charge (CIC) position. He said that the CIC's were often in a difficult position, especially when younger CIC's were expected to take actions affecting more senior controllers. He had heard stories about older controllers ignoring instructions from the younger ones. He counseled the younger controllers that if that occurred, they should advise him so that he could take care of the problem. When the CIC position was in use, the facility tried to have it operate as a stand-alone position. If staffing did not permit the CIC to be split off, the facility tried to combine it with the ground control position, which was typically the least busy position in the cab. Occasionally, because of the qualifications of the individuals in the cab, CIC had to be combined with local control. However, the facility tried to avoid that. When only one supervisor was available, they typically worked in the radar room.

Following the incident involving TONE97, the facility management issued two notices amending procedures for use on the local control and approach control radar positions. The supervisor did not feel that these notices would necessarily be completely effective, but said that they were a temporary measure put in place until the facility could establish a workgroup to review the operation and make additional changes if necessary.

R3 Position Radar Trainee

The R3 trainee began working at Midland tower on February 22, 2009. He was certified on the flight data, clearance delivery, ground control, local control, and tower CIC positions. At the time of the incident, he was training on the R3 radar position. He was assigned Saturday and Sunday off, but worked overtime on Saturday about every other week. Until recently he had been working overtime every Saturday. His typical shift pattern was 1600, 1300, either a 0730 or 1300 shift, 0730, 0730, 0545 (overtime.) He had completed about 180 hours of training on the radar position. The incident occurred during his second or third training session of the day.

When the King Air called, TONE97 flight was already on approach frequency. The R3 trainee said that as he was taking the King Air toward the airport, he wanted to give the tower some "wiggle room" to do what they needed to do for their sequencing. Spacing between aircraft on an overhead approaches and other aircraft was the responsibility of the LC position. The size of the military overhead approach pattern could be 4 to 10 miles depending on the actions of LC. The R3 trainee said he was trying to establish approximately 4 to 6 miles of spacing between the King Air and the flight of T-1s. When he transferred the King Air to LC frequency, he expected that the local controller would turn the King Air if necessary to adjust the spacing. This was a normal operation, and what the R3 trainee would have expected if he had been working the local position.

He had not seen a replay of the incident since it occurred, but did watch it on the radar display in real time. His expectation was that the local controller would turn the King Air toward the south, but that did not occur. He then thought perhaps the local controller would extend the downwind for the two T-1s, but that did not occur either. The R3 trainee was not sure what the local controller's intentions were. Overhead approaches were the only case where the local controller is totally responsible for sequencing.

Asked to provide a definition of tower visual separation, the R3 trainee was unable to state the specific definition beyond saying that the aircraft involved must be in sight of the tower and that the controller should try to keep them apart.

R3 Position Radar Instructor

The R3 instructor began working for the FAA at MAF in 1987. From 1991 until 1992, he was assigned to San Antonio, Texas. In November 1992, he returned to MAF. At the time of the incident he was providing on-the-job training on the R3 radar position.

Asked about the incident from the perspective of the radar position, the R3 instructor noted that he had discussed the incident with his trainee at the conclusion of the training session and noted some alternative ways of handling the situation on the training form for the session. The King Air could have gone first, and the radar position could have provided more space between the King Air and the T-1 flight, which would have prevented the incident. However, the way it was handled was in compliance with the facility's standard operating procedures. The R3 instructor said that he was talking to his trainee about the situation while they monitored the aircraft as it proceeded toward the airport. He commented to the trainee that the local controller was probably going to have to break up the flight of T-1s to accommodate the King Air. At that point he noticed that the second T-1 was turning base leg toward final. That was unexpected. The R3 instructor stated that he had no idea why the tower did not simply extend the downwind leg of the T-1s to allow the King Air to go in first. As noted, the handling of the aircraft by the radar position was in compliance with the local SOP, but the R3 instructor expected that the spacing was going to require some adjustment by the tower controllers. The R3 instructor did not think that the situation required intervention on his part.

The usual practice for the T-1s was that the pilots planned the spacing on their overhead breaks so that the lead aircraft was exiting the runway when the following aircraft was crossing the threshold. In the case of TONE97, the spacing extended beyond what he expected, and the two aircraft ended up over a mile apart, to the point that they had become a non-standard formation.

The R3 instructor said that the usual thought process when setting up aircraft for the tower was, “How would you like to be on the receiving end of this?”

The R3 instructor stated that classroom-type training should occur in classrooms, not on position during OJT. Pre-OJT training preparation was typically done by the supervisors, who try to do a good job. Asked about the possible use of the tower simulator in San Antonio, he said that the former acting facility manager tried to set that up for 3 years but was unable to get it done because of budget issues.

Instructors did know that they should intervene when necessary, but there was always a need to let trainees get in trouble and then work their way out of it. He did not think that there was a general reluctance to take control, but instead instructors were trying to balance jumping in too early with waiting until it was too late to prevent an incident.
On April 13, 2011, at 1222 Central daylight time, an air traffic control operational error and near midair collision occurred at the Midland International Airport (MAF), Midland, Texas, when TONE97 flight, consisting of two T-1 Jayhawk U.S. Air Force training aircraft (civilian-type Beech BE40 turbojets), lost separation with N631ME, a Raytheon-Beech King Air 300, on final approach to runway 28 at MAF. There were no injuries reported, and no damage to any of the involved aircraft.

MAF is a civil airport that is heavily used for training by Air Force aircraft based at Laughlin Air Force Base, located about 170 miles southeast of MAF. During the period surrounding the operational error, there were several different USAF flights operating in the MAF traffic pattern. TONE97 was a flight of two T-1 Jayhawks that entered the airport area on a left downwind approach to runway 28. Initial contact with the radar approach controller was routine, and at 1211:50, the approach controller advised the local controller that TONE97 was two aircraft. At 1217:35, TONE97 was switched to tower frequency and performed a communications check to ensure that the flight wingman was also on tower frequency. After completing the check, TONE97 contacted local control at 1218:02, reporting, “…5 mile initial.” The controller responded, “…left break midfield, descent in the break approved, runway 28 cleared for the option, traffic is a company Beech about three mile final full stop.” TONE97 acknowledged the clearance. At 1219:19, the pilot of N631ME checked in on tower frequency reporting on a visual approach to runway 28. At the time of the report, N631ME was on a wide right downwind about 7 miles north of the runway. The controller acknowledged the call and instructed the pilot to, “…continue for the runway.”

According to recorded radar data, TONE97 commenced the left break maneuver at 1220. As part of the break procedure, the wingman (TONE97-2) dropped back to a position about 1.4 miles behind the lead aircraft. Both aircraft entered close-in left downwind for runway 28. TONE97's transponder was on and the aircraft was displayed as a beacon target. TONE97-2’s transponder was off, so the aircraft produced only a primary radar target with no altitude reporting information available. According to the local controller, N631ME and both aircraft in TONE97 flight were visible from the tower.

At 1221:13, the local controller transmitted, "King Air 1ME you'll be number 3 for the field following a flight of two Beechjets currently about a mile on the left base, report the traffic in sight." The pilot replied, "...uh – I’ve got one of them in sight." The controller continued, "There’s one behind him a little bit higher just south of runway 34R." The pilot did not acknowledge the second transmission, and the controller responded to an unrelated coordination call from the approach controller. Recorded radar data showed that TONE97-2 and N631ME were converging on opposite base legs about 1.5 miles from runway 28. At 1221:53, the local controller instructed N631ME to, "...fly a little bit south and uh do you have the second aircraft in sight?" The pilot responded, "King Air 631ME's waving off, this is entirely ridiculous. We’re going to do a 360 if that’ll work for you." The controller instructed the pilot to make a right 360 and rejoin the final.

At 1223:09, an unidentified voice similar to that of N631ME transmitted, "...that was a perfect recipe for a midair." Following the incident, both aircraft in TONE97 flight and N631ME all landed safely on runway 28.

Radar Data

According to recorded radar data, the closest point of approach between TONE97-2 and N631ME was about 0.3nm. Graphics showing the flight tracks of the aircraft involved in this incident have been entered into the docket.

Personnel Interviews

MAF Local Control Trainee

At the time of the incident, training was in progress at the MAF Local Control (LC) position. The LC trainee had been previously certified at Kansas City Downtown tower in 2008, and transferred to MAF in May 2010. He was certified on the flight data, ground control, and clearance delivery positions. His work schedule included two night shifts and three day shifts. He described traffic at Midland as changeable from day to day, and noted that occasionally the peak traffic can move from the day shift to the night shift. It is very dependent on the military flight schedule.

The LC trainee had completed over 90 hours of training on the LC position. He had not yet received any instruction on the approach radar positions, but stated that he did try to obtain observation time in the radar room to familiarize himself with what the radar controllers do and how they sequence aircraft to the airport. Observation time was not documented on any training forms, as the activity was voluntary and not part of his formal training program.

The LC trainee described his training session as, "going fine, with light traffic." He recalled a pair of T-1 aircraft entering the pattern on an overhead approach. He stated that he instructed the pilots to break midfield, but the flight actually conducted their break further toward the departure end of the runway. As the two aircraft completed their overhead approach, the LC trainee stated that he advised the pilot of N631ME about them. The pilot responded that he had one of the aircraft in sight, but did not see the other. When the first T-1 was arriving at the runway 28 threshold, the second T-1 was on the left base. the LC trainee said that he instructed the pilot of N631ME to fly southbound if he did not have the second T-1 in sight, but it was too late because N631ME had already turned final. The LC trainee stated that his instructor on the position was standing right next to him, but did not intervene or provide any advice.

The LC trainee stated that his objective when sequencing arrivals conducting overhead approaches was to do the sequencing as early as possible. The radar controllers were expected to sequence aircraft to the runways, but the local controller was responsible for sequencing pattern traffic. He stated that he instructed the T-1 pilots to break midfield, and continued that if the pilots had actually executed the break at midfield, "...it would have worked out."

The LC trainee stated that he responded to the initial call from N631ME with an instruction to "continue" instead of advising the pilot that he was number three behind traffic because he was having a hard time breaking the habit he developed at the Kansas City tower, which has only class D airspace, of waiting to sequence aircraft until they are close in to the airport. His trainers at MAF were trying to get him to complete his sequencing earlier. The LC trainee noted that operations in class C airspace such as that surrounding MAF were different from operations in class D airspace, and that controllers had more separation responsibilities.

The LC trainee had used a STARS tower radar display at MKC, where the tower radar display was typically set to a range of approximately 12 to 13 miles. At Midland, the tower radar display range was typically set at 20 or more miles. That range setting was required by the local standard operating procedures manual. He stated that he had been told by his trainers that at MAF the radar range was set to a higher value in order to assist in sequencing traffic. Aircraft at Midland include military and air carrier jets that move faster than the traffic operating at MKC. The LC trainee stated that the approach controller’s planned sequence was normally clear from looking at the tower radar display, but if training was in progress in the radar room it was sometimes harder to tell what the plan was.

The LC trainee was shown a radar replay of the incident in order to refresh his memory of the event. He stated that as the King Air and the T-1 converged, he was looking at the two aircraft out the window and noted that the T-1 was higher than the King Air. He did not expect the King Air to turn inbound toward the airport without having seen the second T-1. When he and his instructor saw both aircraft turning onto the final approach course, their intent was to break one of the aircraft out to resequence it. When they discussed the incident later, the trainee said that the instructor's plan was to turn the T-1 out, while his own plan was to turn the King Air. Before they had a chance to act, the King Air pilot reported that he was going to make a 360 degree turn for spacing.

The LC trainee stated that when applying tower visual separation he was required to have both aircraft in sight and to resolve conflicts by issuing control instructions.

The LC trainee stated that using low approaches as a mechanism for resolving poorly executed sequencing situations was not a technique he had been taught or used. Though he had not yet begun radar training, he did have to complete a radar qualification test to work in the tower. He stated that when adjusting the sequence of inbound aircraft, he was only required to coordinate with the approach controller if his actions would change the sequence. Minor spacing adjustments did not need to be coordinated.

After the incident occurred,he was not involved in any debriefings or other discussions about it except between himself and his trainer. Their conclusion was that "assumption bit us" in that they did not expect the King Air pilot to turn toward the runway on his own.

Local Control Instructor

The LC instructor began working for the FAA in 1988 and transferred to Midland in August 1991. His schedule included Fridays and Saturdays off, but he typically worked every Friday on overtime. He had been working six day weeks for approximately the last two years. He said that it was difficult to get time off except for occasional hour or two at the end of a shift. Taking an entire day off on leave was very difficult.

On the day of the incident, the LC instructor recalled no equipment or operational issues, or other distractions. At the beginning of the shift, he provided two sessions of on-the-job training for a trainee in the radar room, and then came up to the tower cab to begin training on the local control position. When the incident occurred, he and his trainee had been on the LC position for about 50 minutes. The incident occurred in the last 10 minutes of that session. TONE97 called in and was following another T-1. The LC trainee cleared TONE97 for the overhead approach and told the pilot that he was "cleared for the option" on runway 28.

Shortly afterward, N631ME called and was instructed to "continue." TONE97 passed over the airport and commenced the break. As the aircraft proceeded outbound on the downwind leg, the LC trainee advised the pilot of N631ME about the two T-1s also operating in the pattern. The King Air pilot stated that he did not see the second aircraft. The LC instructor said that he was about to intervene and tell the King Air pilot to fly through the final approach course for spacing when they received a coordination call from the approach controller about an aircraft requesting a surveillance approach. During the call, the King Air was turning inbound on final. The LC instructor stated that he was about to instruct the second T-1 in the flight to climb, but saw that the T-1 was already above the King Air. At that point the King Air pilot said something about the aircraft being too close.

The LC instructor said that he could see the King Air and the second T-1 out the window, and he did not believe that the two aircraft were in dangerous proximity to each other. His plan was to break out the second T-1 and instruct the pilot to climb, but he did not do so because the King Air pilot said that he would be executing a 360 degree turn. The LC instructor noted that the second T-1 was further in trail of the lead aircraft than would be normal for a standard formation flight.

Discussing overhead approaches, the LC instructor noted that under the former MAF standard operating procedures manual, the tower would sequence such aircraft with other inbound aircraft by controlling the timing of the pattern break of the overhead approach aircraft. In this situation, he believed it would have been appropriate for local control to have either put the King Air in front of the lead T-1, or to have instructed the T-1s to extend their downwind to fall in behind the King Air. It was LC's responsibility to do that.

Asked when he first realized that the sequence was not working out, the LC instructor said, "...when the King Air turned in on final." It caught him off guard when the pilot turned toward the airport without knowing the location of the traffic he was following. The LC instructor said that he did not believe he would have any trouble breaking out the second T-1, because the aircraft was definitely higher than the King Air. To some extent, the sequencing was left too late because he was trying to let the LC trainee "work it out" for training purposes. He stated that the situation went too far, but that under the circumstances he could have resolved the problem by climbing the second T-1 and turning it for spacing.

Asked about his familiarity with the LC trainee, the LC instructor stated that he had enough awareness of his background to be comfortable training him. It was his understanding that the trainee was at approximately 87 hours of training completed out of the 90 allocated from the position. At that point, the trainee's training would either be extended or terminated.

The LC instructor said that he was unable to recall the specific definition of tower visual separation other than that "...he was required to keep airplanes from running together." He stated that he has not had issues with keeping airplanes apart in the past.

Asked about use of the tower radar display to obtain information not available by looking out the window, such as the speed of arriving aircraft, the LC instructor said that instructors do teach about that. While this incident was occurring, he stated that he was trying to look out the window at the traffic and was not monitoring the radar. The pattern flown by the T-1s was wider than he expected. If he had been handling the position alone, he stated that he would have broken out the T-1s either sooner or later to establish spacing with the King Air.

The LC instructor said that the sequence between the King Air and TONE97 flight was not readily apparent when the King Air called. He noted that local control can change the sequence by adjusting what the aircraft do on overhead approaches. Local control at MAF had a lot of responsibility, almost like a miniature approach control. The local controller was responsible for sequencing traffic to all the runways, even with multiple runways in use.

MAF Front Line Manager (Supervisor)

The supervisor began working for the FAA in October 1998, and came to Midland as a controller in 2001. He became a supervisor in 2007. He was working in the radar room conducting a surveillance approach at the time of the incident involving TONE97 and was not immediately aware of it when it occurred. He stated that supervisors are occasionally required to provide aircraft with surveillance approaches because of staffing, and that it can take approximately 10 minutes to complete such an approach. Since the incident, he had reviewed a radar replay but had not heard the audio of the local control position. He was provided the opportunity to view an audio/visual replay of the incident as part of his interview.

The supervisor said that the handoff of the King Air to LC appeared to be fine as it took place, but the problem appeared to be that the local controllers did not manage the sequencing. The radar approach controller did not always know what LC was going to do with aircraft on overhead approach, so they depended on local control to determine the sequence of the arrivals. In the supervisor's opinion, the King Air should have been number one, with the T-1s following behind. The King Air’s groundspeed was high, and should have been a factor in the sequencing decision. Spacing was normally a cooperative decision between the tower and the radar controller. The tower controller had more flexibility to adjust spacing to the runways.

The supervisor said he was uncertain why the event occurred, but said that the instructor should have forced a sequence when the King Air called rather than allowing the trainee to just instruct the pilot to continue. He noted that the trainee had a control tower operator certificate and was certified at another tower before coming to Midland. Under those circumstances, the expectation would have been that the trainee should have been able to handle the problem. That assumption may have been what got the trainer in trouble. The supervisor noted that the instructor involved had about 20 years experience and was known as a good trainer.

The supervisor stated that he was working about 8 to 16 hours of overtime a month. Until March, he was working 10 hour shifts each day from Monday through Friday, and then eight hours of overtime on Saturdays once or twice a month. Many controllers at Midland were working six day weeks, but recently the supervisors have been attempting to minimize that and release controllers from the sixth-day overtime when there was a way to do so.

MAF had a very high proportion of trainees. The staffing problem started about five years ago when MAF lost numerous certified controllers to other facilities, and they were still trying to catch up.

Asked about use of the tower simulator in San Antonio for training of Midland controllers, the supervisor stated that Midland has been unable to afford to send people there or to put the facility's data into the simulator database. He felt that if the tower simulator in San Antonio was configured for use by MAF, the facility would probably be able to send an instructor and three trainees there to use it.

Supervisors at MAF are responsible for monitoring the operation, as well as managing the training program, including authoring training materials, performing associated administrative tasks, evaluating performance, etc. The facility currently has three full-time supervisors and one temporary supervisor.

There are typically supervisors present in the tower and the radar room from 0900 until 1900, but that was a goal, not a requirement. The operation "tended to be better" when there was a supervisor present. Position audits typically showed improved phraseology and procedures. The facility had also conducted briefings on expectations for the controller-in-charge (CIC) position. He said that the CIC's were often in a difficult position, especially when younger CIC's were expected to take actions affecting more senior controllers. He had heard stories about older controllers ignoring instructions from the younger ones. He counseled the younger controllers that if that occurred, they should advise him so that he could take care of the problem. When the CIC position was in use, the facility tried to have it operate as a stand-alone position. If staffing did not permit the CIC to be split off, the facility tried to combine it with the ground control position, which was typically the least busy position in the cab. Occasionally, because of the qualifications of the individuals in the cab, CIC had to be combined with local control. However, the facility tried to avoid that. When only one supervisor was available, they typically worked in the radar room.

Following the incident involving TONE97, the facility management issued two notices amending procedures for use on the local control and approach control radar positions. The supervisor did not feel that these notices would necessarily be completely effective, but said that they were a temporary measure put in place until the facility could establish a workgroup to review the operation and make additional changes if necessary.

R3 Position Radar Trainee

The R3 trainee began working at Midland tower on February 22, 2009. He was certified on the flight data, clearance delivery, ground control, local control, and tower CIC positions. At the time of the incident, he was training on the R3 radar position. He was assigned Saturday and Sunday off, but worked overtime on Saturday about every other week. Until recently he had been working overtime every Saturday. His typical shift pattern was 1600, 1300, either a 0730 or 1300 shift, 0730, 0730, 0545 (overtime.) He had completed about 180 hours of training on the radar position. The incident occurred during his second or third training session of the day.

When the King Air called, TONE97 flight was already on approach frequency. The R3 trainee said that as he was taking the King Air toward the airport, he wanted to give the tower some "wiggle room" to do what they needed to do for their sequencing. Spacing between aircraft on an overhead approaches and other aircraft was the responsibility of the LC position. The size of the military overhead approach pattern could be 4 to 10 miles depending on the actions of LC. The R3 trainee said he was trying to establish approximately 4 to 6 miles of spacing between the King Air and the flight of T-1s. When he transferred the King Air to LC frequency, he expected that the local controller would turn the King Air if necessary to adjust the spacing. This was a normal operation, and what the R3 trainee would have expected if he had been working the local position.

He had not seen a replay of the incident since it occurred, but did watch it on the radar display in real time. His expectation was that the local controller would turn the King Air toward the south, but that did not occur. He then thought perhaps the local controller would extend the downwind for the two T-1s, but that did not occur either. The R3 trainee was not sure what the local controller's intentions were. Overhead approaches were the only case where the local controller is totally responsible for sequencing.

Asked to provide a definition of tower visual separation, the R3 trainee was unable to state the specific definition beyond saying that the aircraft involved must be in sight of the tower and that the controller should try to keep them apart.

R3 Position Radar Instructor

The R3 instructor began working for the FAA at MAF in 1987. From 1991 until 1992, he was assigned to San Antonio, Texas. In November 1992, he returned to MAF. At the time of the incident he was providing on-the-job training on the R3 radar position.

Asked about the incident from the perspective of the radar position, the R3 instructor noted that he had discussed the incident with his trainee at the conclusion of the training session and noted some alternative ways of handling the situation on the training form for the session. The King Air could have gone first, and the radar position could have provided more space between the King Air and the T-1 flight, which would have prevented the incident. However, the way it was handled was in compliance with the facility's standard operating procedures. The R3 instructor said that he was talking to his trainee about the situation while they monitored the aircraft as it proceeded toward the airport. He commented to the trainee that the local controller was probably going to have to break up the flight of T-1s to accommodate the King Air. At that point he noticed that the second T-1 was turning base leg toward final. That was unexpected. The R3 instructor stated that he had no idea why the tower did not simply extend the downwind leg of the T-1s to allow the King Air to go in first. As noted, the handling of the aircraft by the radar position was in compliance with the local SOP, but the R3 instructor expected that the spacing was going to require some adjustment by the tower controllers. The R3 instructor did not think that the situation required intervention on his part.

The usual practice for the T-1s was that the pilots planned the spacing on their overhead breaks so that the lead aircraft was exiting the runway when the following aircraft was crossing the threshold. In the case of TONE97, the spacing extended beyond what he expected, and the two aircraft ended up over a mile apart, to the point that they had become a non-standard formation.

The R3 instructor said that the usual thought process when setting up aircraft for the tower was, “How would you like to be on the receiving end of this?”

The R3 instructor stated that classroom-type training should occur in classrooms, not on position during OJT. Pre-OJT training preparation was typically done by the supervisors, who try to do a good job. Asked about the possible use of the tower simulator in San Antonio, he said that the former acting facility manager tried to set that up for 3 years but was unable to get it done because of budget issues.

Instructors did know that they should intervene when necessary, but there was always a need to let trainees get in trouble and then work their way out of it. He did not think that there was a general reluctance to take control, but instead instructors were trying to balance jumping in too early with waiting until it was too late to prevent an incident.