On November 12, 2022, about 1322 central standard time, a Boeing B-17G, N7227C, and a Bell P-63F, N6763, collided in flight during a performance at the Commemorative Air Force’s (CAF) Wings Over Dallas air show at Dallas Executive Airport (KRBD) in Dallas, Texas. The pilot, copilot, flight engineer, and two scanners on board the Boeing B-17G and the pilot of the Bell P-63F were fatally injured, and both airplanes were destroyed. No injuries to persons on the ground were reported. Both accident airplanes (and six other historic, former military airplanes that were airborne as part of the same performance) were operated by the CAF under the provisions of Title 14 Code of Federal Regulations (CFR) Part 91 and a certificate of waiver for the air show.

The Boeing B-17G was in the first position of five historic bomber airplanes flying as solo aircraft in trail, and the Bell P-63F was in the last position of three historic fighter airplanes flying in formation. The takeoffs, repositioning turns, and passes of the eight airplanes in the accident performance were directed in real time via radio by the air boss, who had primary responsibility for the control of air show operations.

Just before the accident, the bomber group and the fighter formation completed a pass in front of the crowd of spectators from show right to left (that is, right to left from the crowd’s perspective). The airplanes were setting up for the next pass when the accident occurred. This pass was intended to be from show left to right in front of the crowd, and the air boss issued directives for the fighter formation to pass off the left side of the bomber group airplanes and then cross in front of them. The position data showed that the flight path for the fighter lead and position 2 fighter airplanes passed the bomber airplanes off the bombers’ left side before crossing in front of the Boeing B-17G but that the Bell P-63F’s flight path converged with that of the Boeing B-17G. Video and photographic evidence captured by witnesses on the ground showed that the Bell P-63F was in a descending, left-banked turn when it struck the left side of the Boeing B-17G near the trailing edge of the left wing, then both airplanes broke apart in flight.

​See the photos on Flickr​.​

We found that the accident pilots had limited ability to see and avoid each other’s airplane due to flight path geometry, out-the-window view obscuration by aircraft structures, the attention demands associated with the air show performance, and the limitations of human performance that can make it difficult to see another aircraft.

We also found that the air boss’s deconfliction strategy for the accident performance, which relied on the air boss’s real-time, predictive assessment of airplane locations and the ability of the CAF pilots to see and avoid other airplanes, was ineffective because the flight paths of the Boeing B-17G and the Bell P-63F converged as each pilot maneuvered to set up for the pass.

We found evidence that the air show guidance provided by the Federal Aviation Administration (FAA) and the International Council of Air Shows Inc. (ICAS) did not adequately address the need to better mitigate the collision risks associated with air boss–directed performances involving multiple, dissimilar aircraft. We determined that a lack of administrative controls and a documented risk assessment process for ensuring air show aircraft separation directly contributed to the in-flight collision.

We found that, unlike the FAA regulatory requirements for pilots and air traffic controllers, air bosses are not subject to any recurrent evaluations, and there are no standardized communications terms for air boss–provided directives to ensure the clarity and brevity of radio communications. Further, the lack of guidance and required surveillance tasks for FAA inspectors assigned to air shows related to the direct observation of an air boss’s performance represents a missed opportunity for inspectors to detect and provide debriefing feedback to address performance-related safety issues.

We also found that the CAF’s lack of a strong, clearly defined safety risk assessment plan resulted in air show production decisions that were not systematically developed to determine acceptable levels of risk and were susceptible to influences unrelated to safety.

We determined that the probable cause of this accident was the air boss’s and air show event organizer’s lack of an adequate, prebriefed aircraft separation plan for the air show performance, relying instead on the air boss’s real-time deconfliction directives and the see-and-avoid strategy for collision avoidance, which allowed for the loss of separation between the Boeing B-17G and the Bell P-63F airplanes. Also causal was the diminished ability of the accident pilots to see and avoid the other aircraft due to flight path geometry, out-the-window view obscuration by aircraft structures, attention demands associated with the air show performance, and the inherent limitations of human performance that can make it difficult to see another aircraft. Contributing to the accident were the lack of FAA guidance for air bosses and air show event organizers on developing plans and performing risk assessments that ensure the separation of aircraft that are not part of an approved maneuvers package and the lack of FAA requirements and guidance for recurrent evaluations of air bosses and direct surveillance of their performance.

​As a result of this investigation, we made two new recommendations each to the FAA and to the ICAS. We recommended that they work together with other air show industry stakeholders to develop standardized terms to help ensure the clarity and brevity of air boss–provided directives to performers. We also recommended that they both work with the warbird community to establish standard operating procedures for air show event organizers and air bosses that include applying effective administrative controls to ensure air show aircraft separation, performing a safety risk assessment for each performance, and completing a daily debriefing with continuous feedback to the FAA and the ICAS to address any identified deficiencies.

We also made two new recommendations to the FAA to require recurrent air boss evaluations as part of the letter of authorization renewal process and to revise FAA Order 8900.1 to provide guidance and a job aid for FAA inspectors who evaluate an air boss’s performance, including requirements for the FAA inspector-in-charge of an air show to observe an air boss’s performance visually and on the air boss radio frequency and provide appropriate feedback during the air show debriefing. We made one new recommendation to the CAF to establish a safety risk assessment process for identifying and mitigating risks, tracking safety hazards, and routinely assessing trend data to include policies and procedures that specifically address the unique aspects of air show operations.