UC Berkeley

This thesis develops methodologies to evaluate the performance of various components of the National Airspace System (NAS) that are related to air traffic management (ATM). The NAS is a multifaceted network comprising aircraft, airports, air traffic managers and controllers, passengers, and navigational aids. Air traffic management is, in broadest terms, the process of shifting flight schedules to reduce demand capacity imbalances that must to be handled by tactical control and aircraft maneuvers. ATM performance is influenced by a mix of external weather uncertainties and internal operational errors, making the system's management unpredictable and complex. As U.S. air traffic is expected to grow significantly, further complicated by escalating weather uncertainties and operational challenges, there is an urgent need for a comprehensive performance analysis of ATM in the NAS. Such analysis is essential to identify key performance indicators, define metrics for trend assessments, and uncover underlying patterns that can inform future improvements. This thesis is structured into three parts, distributed over four chapters, each focusing on different aspects of NAS operations.

Part I-Chapter 1 focuses on analyzing Actual Airborne Time (AAT) between the U.S. and China. A major goal of ATM is to shift delay in the air to delay on the ground. Since Chinese ATM is in its nascent stages, a comparison of AAT in the two nations sheds light on how more advanced ATM can reduce airborne times. This chapter conducts the first comparative empirical study on AAT, examining how variables such as origin-destination distance, hub status, traffic conditions, and weather influence flight times in both countries. Econometric models reveal that AAT is strongly correlated with distance, with flights in China experiencing longer times under similar distances, suggesting inefficiencies in Chinese airspace management. The chapter also explores the potential for efficiency gains if Chinese airspace operations were aligned with U.S. standards, projecting significant reductions in AAT and associated cost savings.

Part II-Chapter 2 addresses the supply side by investigating runway and terminal airspace capacities, crucial bottlenecks in the NAS. Focusing on JFK Airport, this chapter presents a statistical model of controlled interarrival times (CIT) at landing thresholds, incorporating factors such as weather, runway utilization, and aircraft characteristics. A novel conditional probability model is introduced, capturing CIT behaviors under varying arrival demands. This model aids in understanding and enhancing landing safety and efficiency by depicting the probability distributions of CIT under different operational scenarios.

Part III evaluates the impact of TMIs designed to balance flight demand and airspace capacity across different components of the NAS. Chapter 3 evaluates Miles-in-Trail (MIT), a common TMI that regulates spacing between aircraft to manage capacity-demand imbalances. A case study at ORD Airport illustrates that eliminating MIT could reduce overall queuing delays but increase arrival delays, indicating a trade-off between local congestion and system-wide throughput. Chapter 4 delves into the efficiency of Ground Delay Programs (GDPs), which are implemented to address demand-capacity imbalances at arrival airports by converting potential airborne delays into pre-departure ground delays. This chapter introduces a deterministic queuing model to simulate scenarios absent of GDPs, focusing on excess delay and wasted slots. It applies the methodology to 1204 GDPs from 33 airports in 2019. The results indicate that removing GDPs could reduce excess delays by approximately 70%, albeit at the cost of increased airborne holding. The primary inefficiencies in GDPs stem from inaccurate program rate settings and flight time deviations, suggesting that significant improvements in GDP efficiency could be achieved by refining these parameters.