The National Airspace System (NAS) in the United States had an inventory of 5,156 big jets at the end of December 2002, of which 4,085 were narrow bodies, and 1,071 were wide bodies. In addition, there were 1,180 regional jets and 660 turboprops in the system at that time. Empirical research reveals that there is a critical link between the flow of scheduled passenger services and the choice of aircraft used by the airlines in serving market pair demand. This relationship can be empirically retrieved without detailed knowledge of airlines' behavior and used for analyzing traffic patterns in the NAS. Using the T100 segment data from the first two quarters of 2004, a multinomial qualitative choice model is developed in this paper. This framework establishes empirical linkages among aircraft choice, and passenger flows, distance, types of airport hubs, network and time of the year. Estimated models demonstrate that both passengers and distance play important roles in selecting types of aircraft. Overall, the model is capable of predicting exact choices 51% of the time; with some flexibility of making a one-off mistake, the model is capable of making almost nine out of 10 choices correctly. Using the estimated coefficients from the qualitative choice model and varying assumptions (number of passengers, in particular), forecasts of aircraft operations by market segments and the fleet mix can easily be generated. These forecasts can then be used to understand the performance of the U.S. NAS.
