Airborne incidents : an econometric analysis of severity, December 19, 2014 : Final report

Final reportPDFTech ReportDOT-VNTSC-FAA-15-03FAA/AJI-33Air traffic controlAircraft separationEconometric modelsMultinomial logitsEconometricsChoice modelsLogitsVariablesUnited StatesJohn A. Volpe National Transportation Systems Center (U.S.)Biernbaum, LeePrice, MeganWishart, JacobJohn A. Volpe National Transportation Systems Center (U.S.)United States. Federal Aviation Administration. Air Traffic Organization. Office of Safety and Technical TrainingAGR-SAFETY AND SECURITY-SAFETY AND SECURITYNTL-AVIATION-AVIATIONNTL-AVIATION-Air Traffic ControlNTL-AVIATION-Aviation Economics and FinanceNTL-AVIATION-Aviation Planning and PolicyNTL-AVIATION-Aviation Safety/AirworthinessNTL-ECONOMICS AND FINANCE-ECONOMICS AND FINANCENTL-ECONOMICS AND FINANCE-Aviation Economics and FinanceNTL-SAFETY AND SECURITY-Aviation Safety/AirworthinessVolpe National Transportation Systems Center, Technical Reference CenterAirborne loss of separation incidents occur when an aircraft breaches the defined separation limit (vertical and/or horizontal) with another aircraft or terrain imposed by Air Traffic Control. Identifying conditions that lead to more severe loss of separation incidents can lead to policy implications and future areas of research. Previous research focused on qualitative approaches to analyzing such events, and tended to examine only the frequency of events. This report puts the severity of a loss of separation incident front and center and uses econometric techniques to examine the relationship between severity and conditional factors during the incident.The report utilizes report data from the Air Traffic Safety Action Program (ATSAP), with a concentration on terminal airspace incidents. A number of other FAA data sources were merged to provide a robust set of information at the time of event in terms of facility, weather, and other operational characteristics. The primary focus of this research was on the use of discrete choice, multinomial logit models to better understand the relationship between these different set of factors at the time of the event and the severity outcome