Understanding the capacity of airport runway systems

Abstract

Runway systems are often the primary bottlenecks in airport operations. Thus, understanding their capacity is of critical importance to airport operators. However, developing this understanding is not straightforward because, unlike demand or throughput, runway system capacity (RSC) remains unobserved. Moreover, the complex interactions of the physical runway system infrastructure with underlying operating conditions (such as weather) and the airspace result in different capacities under different airport operational scenarios, thereby making the measurement of RSC more complicated. Both analytical and simulation-based approaches need extensive efforts for customization according to specific runway configurations. Analytical models with a moderate level of fidelity are often used to support strategic capacity decisions. In contrast, high-fidelity simulation-based approaches are more appropriate for accommodating wide-ranging operational scenarios and providing accurate RSC estimates to support short-term capacity decisions, though they tend to be resource-intensive. To that end, the availability of granular data on day-to-day runway operations facilitates the development of statistical model that can offer a standardized model specification with minimal customization and provide a precise estimation of RSC for short-term capacity decisions. However, the exercise is empirically challenging due to statistical biases that emerge via the above-mentioned interactions between air traffic flow and control at airports and in the airspace and RSC. This paper develops a novel causal statistical framework based on a confounding-adjusted Stochastic Frontier Analysis (SFA) to deliver estimates of RSC and its parameters that are robust to such biases and are therefore suitable to inform airport operations and planning. The model captures the key factors and interactions affecting RSC in a computationally efficient manner. The performance of the model is benchmarked via a Monte Carlo simulation and further by comparing the estimated capacities of five major multi-runway airports with their representative estimates from the literature. • First application of causal statistical modeling in empirical estimation of runway system capacity. • Addressed various confounding challenges in such empirical analysis. • Delivered a computationally-efficient runway system capacity estimation model. • Used a unique day-to-day operation data of five major multi-runway hub airports. • Results are relevant in airport planning and management.

Topics & Concepts