WPR15MA243

The local controller's (LC) failure to properly identify the aircraft in the pattern and to ensure control instructions provided to the intended Cessna on downwind were being performed before turning Eagle1 into its path for landing. Contributing to the LC's actions was his incomplete situational awareness when he took over communications from the LC trainee due to the high workload at the time of the accident. Contributing to the accident were the inherent limitations of the see-and-avoid concept, resulting in the inability of the pilots involved to take evasive action in time to avert the collision.

The Cessna 172 (N1285U) was conducting touch-and-go landings at Brown Field Municipal Airport (SDM), San Diego, California, and the experimental North American Rockwell NA265-60SC Sabreliner (N442RM, call sign Eagle1) was returning to SDM from a mission flight. SDM has two parallel runways, 8R/26L and 8L/26R; it is common in west operations for controllers to use a right traffic pattern for both runways 26R and 26L due to the proximity of Tijuana Airport, Tijuana, Mexico, to the south of SDM. On the morning of the accident, the air traffic control tower (ATCT) at SDM had both control positions (local and ground control) in the tower combined at the local control position, which was staffed by a local controller (LC)/controller-in-charge, who was conducting on-the-job training with a developmental controller (LC trainee). The LC trainee was transmitting control instructions for all operations; however, the LC was monitoring the LC trainee's actions and was responsible for all activity at that position. About 13 minutes before the accident, the N1285U pilot contacted the ATCT and requested touch-and-go landings in the visual flight rules (VFR) traffic pattern. About that time, another Cessna 172 (N6ZP) and a helicopter (N8360R) were conducting operations in the VFR traffic pattern, and a Cessna 206 Stationair (N5058U) was inbound for landing. Over the next 5 minutes, traffic increased, with two additional aircraft inbound for landing. (Figure 1 in the factual report for this accident shows the aircraft in the SDM traffic pattern about 8 minutes before the accident.) The LC trainee cleared the N1285U pilot for a touch-and-go on runway 26R; the pilot acknowledged the clearance and then advised the LC trainee that he was going to go around. The LC trainee advised the N1285U pilot to expect runway 26L on the next approach. At that time, three aircraft were using runway 26R (Global Express [N18WZ] was inbound for landing, N6ZP was on a right base for a touch-and-go, and a Cessna Citation [XALVV] was on short final) and three aircraft were using runway 26L (N1285U was turning right downwind for the touch-and-go, a Skybolt [N81962] was on a left downwind for landing, and N8360R was conducting a touch-and-go landing). After N1285U completed the touch-and-go on runway 26L, the pilot entered a right downwind for runway 26R. Meanwhile, Eagle1 was 9 miles west of the airport and requested a full-stop landing; the LC trainee instructed the Eagle1 flight crew to enter a right downwind for runway 26R at or above an altitude of 2,000 ft mean sea level. At this time, about 3 minutes before the accident, the qualified LC terminated the LC trainee's training and took over control of radio communications. From this time until the collision occurred, the LC was controlling nine aircraft. (Figure 2 and Figure 4 in the factual report for this accident show the total number of aircraft under ATCT control shortly before the accident.) During the next 2 minutes, the LC made several errors. For example, after N6ZP completed a touch-and-go on runway 26R, the pilot requested a right downwind departure from the area, which the LC initially failed to acknowledge. The LC also instructed the N5058U pilot, who had been holding short of runway 26L, that he was cleared for takeoff from runway 26R. Both errors were corrected. In addition, the LC instructed the helicopter pilot to "listen up. turn crosswind" before correcting the instruction 4 seconds later to "turn base." (Figure 2 in the factual report for this accident shows the aircraft in the traffic pattern about 2 minutes before the accident.) About 1 minute before the collision, the Eagle1 flight crew reported on downwind midfield and stated that they had traffic to the left and right in sight. At that time, N1285U was to Eagle1's right, between Eagle1 and the tower, and established on a right downwind about 500 ft below Eagle1's position. N6ZP was about 1 mile forward and to the left of Eagle1, heading northeast and departing the area. Mistakenly identifying the Cessna to the right of Eagle1 as N6ZP, the LC instructed the N6ZP pilot to make a right 360° turn to rejoin the downwind when, in fact, N1285U was the airplane to the right of Eagle1. (The LC stated in a postaccident interview that he thought the turn would resolve the conflict with Eagle1 and would help the Cessna avoid Eagle1's wake turbulence.) The N6ZP pilot acknowledged the LC's instruction and began turning; N1285U continued its approach to runway 26R. However, the LC never visually confirmed that the Cessna to Eagle1's right (N1285U) was making the 360° turn. Ten seconds later, the LC instructed the Eagle1 flight crew to turn base and land on runway 26R, which put the accident airplanes on a collision course. The LC looked to ensure that Eagle1 was turning as instructed and noticed that the Cessna on the right downwind (which he still mistakenly identified as N6ZP) had not begun the 360° turn that he had issued. The LC called the N6ZP pilot, and the pilot responded that he was turning. In the first communication between the LC and the N1285U pilot (and the first between the controllers in the ATCT and that airplane's pilot in almost 6 minutes), the LC transmitted the call sign of N1285U, which the pilot acknowledged. N1285U and Eagle1 collided as the LC tried to verify N1285U's position. A postaccident examination of both airplanes did not reveal any mechanical anomalies that would have prevented the airplanes from maneuvering to avoid an impact. Local Controller Actions In a postaccident interview, the LC stated that his personal limit for handling aircraft was four aircraft on runway 26R plus three aircraft on runway 26L (for a total of seven). From the time the LC took over local control communications from the LC trainee (3 minutes before the accident) until the time of the collision, the LC was in control of nine aircraft. Thus, the LC had exceeded his own stated workload limit. Research indicates that the cognitive effects of increasing workload may include memory deficits; distraction; narrowing of attention; decreased situational awareness; and increased errors, such as readback errors or giving instructions to the wrong aircraft. (Mica Endsley and Mark Rodgers's 1997 report, Distribution of Attention, Situation Awareness, and Workload in a Passive Air Traffic Control Task: Implications for Operational Errors and Automation [FAA Report No. DOT/FAA/AM-97/13], details the cognitive effects of increasing workload.) To resolve the increasing workload, the LC had two options. He could have directed traffic away from SDM or split the local control/ground control positions, but he did neither. The LC trainee was qualified to work the ground control position, and the SDM ATCT had three controllers in the facility, which was the normal staffing schedule for that day and time. As a result of the high workload, the LC made several errors after taking over the position from the LC trainee, including not responding promptly to a departure request from the N6ZP pilot and incorrectly instructing a helicopter pilot to turn to crosswind before correcting the instruction to turn base. The LC also did not provide traffic and/or sequence information with the instructions for the N6ZP pilot to turn 360° right. If the LC had done so, the N6ZP pilot might have reminded the controller that he was departing the airspace or requested clarification per 14 Code of Federal Regulations (CFR) 91.123(a), "Compliance with ATC [Air Traffic Control] Clearances and Instructions." In addition, if the Eagle1 flight crew had heard their aircraft called as traffic to another aircraft, it may have helped their visual search or prompted them to seek more information about the location of the conflicting traffic. The LC's stress amid the high workload was evidenced in his "listen up. turn crosswind" instruction to the helicopter pilot, after which th

Aircraft 2 ***This report was modified on October 31, 2016. Please see the docket for this accident to view the original report.*** On August 16, 2015, about 1103 Pacific daylight time, a Cessna 172M, N1285U, and an experimental North American Rockwell NA265-60SC Sabreliner, N442RM (call sign Eagle1), collided in midair about 1 mile northeast of Brown Field Municipal Airport (SDM), San Diego, California. The pilot (and sole occupant) of N1285U and the two pilots and two mission specialists aboard Eagle1 died; both airplanes were destroyed. N1285U was registered to a private individual and operated by Plus One Flyers under the provisions of 14 Code of Federal Regulations (CFR) Part 91 as a personal flight. Eagle1 was registered to and operated by BAE Systems Technology Solutions & Services, Inc., for the US Department of Defense as a public aircraft in support of the US Navy. No flight plan was filed for N1285U, which originated from Montgomery-Gibbs Executive Airport, San Diego, California. A mission flight plan was filed for Eagle1, which originated from SDM about 0830 and was returning to SDM. Visual meteorological conditions prevailed at the time of the accident. On the morning of the accident, the SDM airport traffic control tower (ATCT) had all control positions (local and ground control) in the tower combined to the local control position. The position was staffed by a qualified local controller (LC)/controller-in-charge (CIC) who was conducting on-the-job training with a developmental controller (LC trainee) on the local control position. The LC trainee was transmitting control instructions for all operations; however, the qualified LC was closely monitoring the LC trainee's actions and was responsible for all activity at that position. According to air traffic control (ATC) radar and voice communications data, the pilot of N1285U contacted the SDM ATCT at 1049:44 and requested touch-and-go maneuvers in the visual flight rules (VFR) traffic pattern. N1285U was inbound about 6 miles to the northeast of SDM, at an indicated altitude of 2,600 ft. About that time, another Cessna 172 (N6ZP) and a helicopter (N8360R) were conducting operations in the VFR traffic pattern, and a Cessna 206 Stationair (N5058U) was inbound for landing after carrying parachutists to a local drop zone about 5 nautical miles (nm) east of the field. Between about 1049 and 1054, N6ZP and the helicopter continued to conduct approaches, N5058U landed on runway 26L, a Skybolt (N81962) reported west of SDM for landing on runway 26L, and a Cessna Citation (XALVV) reported straight in for landing on runway 26R. At 1052:57, the LC trainee cleared the pilot of N1285U for a touch-and-go on runway 26R, which the pilot acknowledged. At 1054:46, when N1285U was on final approach of the first approach to runway 26R, the pilot advised the LC trainee that he was going to go around. The LC trainee acknowledged the transmission and instructed the pilot to follow "a Cessna" (N6ZP) on the right downwind. At 1056:31, the LC trainee advised the pilot of N1285U to expect runway 26L on the next approach, which the pilot acknowledged. At that time, three aircraft were using runway 26R (Global Express [N18WZ] was inbound for landing, N6ZP was on a right base for a touch-and-go, and XALVV was on short final) and three aircraft were using runway 26L (N1285U was turning right downwind for the touch-and-go, N81962 was on a left downwind for landing, and N8360R was conducting a touch-and-go operation). Figure 1 shows the aircraft in the SDM traffic pattern about 8 minutes before the accident. Figure 1. Aircraft in the SDM traffic pattern about 8 minutes before the accident. At 1057:22, the LC trainee cleared the pilot of N1285U for a touch-and-go on runway 26L, and at 1057:27, the pilot acknowledged the clearance. At 1058:22, the LC trainee cleared the pilot of N6ZP for a touch-and-go on runway 26L. At 1058:29, the pilot of N6ZP stated, "…ah two six right cleared touch and go." After the pilot of N1285U completed the touch-and-go on runway 26L, the pilot turned the airplane right, crossing through the departure corridor of runway 26R, and entered a right downwind for runway 26R. At 1059:04, when Eagle1 was 9 miles west of SDM, the flight crew contacted the SDM ATCT and requested a full-stop landing. Throughout Eagle1's cockpit voice recorder (CVR) recording, the pilot, seated in the left seat, was communicating on the radio and responding to checklists, consistent with that pilot acting as the pilot monitoring and the copilot, seated in the right seat, acting as the pilot flying. The LC trainee instructed the Eagle1 flight crew to enter a right downwind for runway 26R at or above an altitude of 2,000 ft mean sea level (msl). At 1059:18, the pilot of N5058U reported holding short of runway 26L on taxiway C. (N5058U had landed on runway 26L at 1052:30 and was returning to runway 26L for takeoff.) The LC trainee mistakenly advised the pilot of N5058U to hold short of runway 26R. The pilot of N5058U clarified that he was holding short of 26L, and, at 1059:31, the LC trainee acknowledged the transmission. That was the last transmission from the LC trainee. At 1059:33, the qualified LC terminated the LC trainee's training and took over control of communications due to increased traffic. The LC trainee signed off the position but remained in the tower to observe operations. From this time until the collision occurred (about 1103), the LC was controlling nine aircraft. During the next 2 minutes, the LC made several errors that were either corrected by him or by the pilots under his control. At 1059:44, after the pilot of N6ZP completed a touch-and-go on runway 26R, he requested a right downwind departure from the area. The LC did not respond. At 1100:23, the LC instructed, "stationair five eight uniform two six right cleared for I'm sorry two six left cleared for takeoff." At 1100:29, the pilot of N5058U stated, "uh I'm sorry was that for five eight uniform?" The LC then cleared the pilot of N5058U for takeoff from runway 26L. At 1100:36, the LC transmitted, "helicopter six zero romeo there is a ces ah cen ah correction stationair just ahead they are going to the right runway base leg for two six left." At 1100:46, the pilot of N6ZP repeated his request for departure; the LC then approved N6ZP's departure request, and N6ZP departed the traffic pattern in a northeasterly direction. At 1100:53, the LC instructed the helicopter pilot, "helicopter six zero romeo listen up turn crosswind" before correcting the instruction 4 seconds later to "turn base." At 1101:15, the Eagle1 CVR recorded the copilot state, "got one on the runway," and at 1101:19, the Eagle1 CVR recorded the pilot comment, "wowww. he's like panicking" (with an emphasis on panicking). Figure 2 shows the aircraft in the SDM traffic pattern from about 1101 until the time of the accident. Figure 2. Aircraft in the SDM traffic pattern from about 1101 until the time of the accident. At 1101:49, the Eagle1 CVR recorded one of the mission specialists seated outside the cockpit ask "see him right there?" At 1102:14, while on the right downwind leg (and, according to radar data, while overtaking N1285U from behind and to the left) and abeam the tower, the Eagle1 flight crew reported to the ATCT that they had traffic in sight to the left and the right of their position. Radar data indicated that N6ZP was to the left of Eagle1 and heading to the northeast, and N1285U was between Eagle1 and SDM, on a closer-in right downwind leg. At 1102:32, the LC instructed the pilot of N6ZP, which he thought was the Cessna on right downwind, to make a right 360° turn over the airport and rejoin the downwind. Despite the fact that, at that time, N6ZP was 2.3 nm northeast of the airport and was departing the area, the pilot of N6ZP acknowledged the instruction and initiated a right turn. At the same time, Eagl --- Aircraft 1 Cockpit Voice Recorder The CVR was recovered from the Eagle1 wreckage and forwarded to the National Transportation Safety Board (NTSB) vehicle recorders laboratory in Washington, DC, for readout. The CVR had 30 minutes of analog audio on a continuous loop tape in a four-channel format: one channel for each of the two pilot stations, one channel for the cockpit observer station, and one channel for the cockpit area microphone (CAM). The magnetic tape was retrieved from within the crash-protected case and was successfully downloaded. The quality of the CVR audio information was degraded due to the erase mechanism not completely erasing the previous recordings, especially on the CAM channel. Timing on the transcript was established by correlating the CVR events to the common events recorded by SDM ATC. The CVR recording started at 1032:28 and ended at 1103:10. Due to the poor quality of the CVR recording, the SDM ATC transcript was used in conjunction with the CVR recording to clarify the flight crew's radio transmissions. Aircraft Performance and Cockpit Visibility Study The NTSB's investigation examined the ability of the N1285U and Eagle1 pilots to see and avoid the other aircraft. To determine approximately how each aircraft would appear in the pilots' fields of view, the position of the "target" aircraft in a reference frame attached to the "viewing" aircraft must be calculated. This calculation depends on the positions and orientation (pitch, roll, and yaw angles) of each aircraft, as well as the location of the pilots' eyes relative to the cockpit windows. Position and orientation information for both airplanes was estimated based on an analysis of the radar data, combined with models of each airplane's aerodynamic performance. For this study, the relative positions of the two aircraft were calculated beginning at 1100:06.0 and then at 0.05-second intervals up to the collision, which occurred at 1103:10.2. The time, location, and altitude of the collision were determined based on extrapolation of the radar data, the wreckage locations of both aircraft, and the time of the end of Eagle1's CVR recording. The locations of the structures and transparencies of Eagle1 in its copilot's (right seat) field of view, and of N1285U in its pilot's (left seat) field of view, were determined from the interior and exterior dimensions of representative airplanes, as measured using a laser scanner. The structural obscurations to each pilot's view were merged with the calculated relative position data and are discussed below. The study assumed a nominal pilot seating (and eye) position in each cockpit and evaluated a matrix of eye displacements from the nominal eye position. The variations in eye position indicated that pilot head movements can move the target airplane's positon in the field of view into and out of areas that are obscured from the pilots. For Eagle1, the visibility of N1285U from the copilot's seat is sensitive to the pilot's eye position relative to the top and left edges of the Eagle1 R2 window, and for N1285U, the visibility of Eagle1 is sensitive to the pilot's eye position relative to the post between the left door window and the windshield (see figure 6 for a top-down view of the Eagle1 forward fuselage with cockpit windows labeled). The description of the visibility from each aircraft that follows is based on the pilots' eyes at "nominal" positions, determined by the eye positions of persons of similar stature to the accident pilots seated in exemplar airplanes. The Aircraft Performance and Cockpit Visibility Study for this accident describes how the visibility from each airplane changes with variation in eye position and notes that head movements in several directions while scanning for traffic can make otherwise obscured aircraft visible. Eagle1's 10 cockpit windows (5 on each side of the airplane) are labeled L1 through L5 for the left windows and R1 through R5 for the right windows, as shown in figure 6. The Cessna windows are the windshield, left window, and right window. Figure 6. Top-down view of Eagle1 forward fuselage, showing labels used to identify cockpit windows. At 1100:06, N1285U was climbing through 540 ft over runway 26L after completing its touch-and-go operation, and Eagle1 was 6 nm west and 1 nm north of the SDM runway 26R threshold, descending through 2,380 ft. At 1100:29, Eagle1's CVR recorded the copilot comment "got one on short final" (likely Global Express N18WZ), and Eagle1 was descending through 2,260 ft, 1 nm north and 4.9 nm west of the runway 26R threshold. N1285U was climbing through 780 ft, along the extended centerline of runway 26L, and about 800 ft past the departure end of that runway. N1285U would have been located in Eagle1's R1 window. The other aircraft in the pattern would have been located in roughly the same area, except for the other Cessna 172, N6ZP, which was to the left in the R1 window. Eagle1 would have appeared in N1285U's windshield. The aircraft were 4.1 nm apart. At 1100:55, N1285U was climbing through 1,150 ft about 1,200 ft west of the departure end of runway 26R and began a right turn to cross over the extended centerline of 26R to enter the right traffic pattern for that runway. Eagle1 was at 2,190 ft, about 1 nm north and 3.6 nm west of the runway 26R threshold, about 2.4 nm from N1285U, and would have been located in N1285U's windshield about in line with the top of the instrument panel. N1285U would have been located in Eagle1's R1 window. At 1101:15.5, Eagle1's CVR recorded the copilot state "got one on the runway" (likely N18WZ). At this time, N18WZ may have been obscured behind the post separating Eagle1's R1 and R2 windows, if the copilot's eyes had been looking from the "nominal" position. Since the copilot saw and commented on N18WZ, however, he may have been leaning closer to the window to scan for traffic, bringing N18WZ more into view. N1285U would have appeared in Eagle1's R2 window and was 1.2 nm away. Eagle1 would have appeared in the N1285U pilot's field of view near the forward edge of the post separating the left window from the windshield. At 1101:24.6, Eagle1 was level at 2,100 ft, about 1 nm north and 2.4 nm west of the runway 26R threshold, when the pilot stated "I got twelve o'clock on a climb out." At this time, the Cessna 206 (N5058U) had recently departed from runway 26L. N1285U was 0.8 nm from Eagle1, climbing through 1,600 ft and turning from crosswind to right downwind for runway 26R, 0.5 nm north and 1.7 nm west of the runway 26R threshold. N1285U would have appeared in Eagle1's R2 window; N5058U would have appeared slightly below and to the left of N1285U. Eagle1 would have been hidden from the N1285U pilot's view behind the window post between the left window and windshield. At 1101:43.1, N1285U would have been located just to the left of the window post separating Eagle1's R2 and R3 windows, and N5058U would have been just to the right of this post (see figure 7a). Eagle1 would have appeared in the left window of N1285U, just below the wingtip (see figure 7b). Eagle1 and N1285U were both on the right downwind leg for runway 26R, with Eagle1 about 0.4 nm north of N1285U. Figure 7a. View from the copilot seat of Eagle1 at 1101:43.1, when the airplanes were 0.5 nm apart. Figure 7b. View from the pilot seat of N1285U at 1101:43.1, when the airplanes were 0.5 nm apart. At 1101:49.0, Eagle1's CVR recorded one of the mission specialists seated outside of the cockpit ask "see him right there?" Eagle1 was at about 2,040 ft, 1.1 nm north and 1.4 nm west of the runway 26R threshold. N1285U was level at 1,700 ft, 0.7 nm north and 1.3 nm west of the threshold. The airplanes were about 0.5 nm apart, and N1285U would have appeared in Eagle1's R3 window near the right edge of the post separating the R2 and R3 windows. Eagle1 would have been obscured from the N1285U pilot's v