مدل سازی رفتار انتخابی حالت دسترسی زمینی به فرودگاه: مطالعه موردی در استانبول، ترکیه  Mode choice behavior modeling of ground access to airports: A case study in Istanbul, Turkey

1. Introduction and background There has been a rise in international and domestic air travel (or mode share of air travel) over the past ten years with a growth rate of 5.6 per cent (IATA: International Air Transport Association, 2015). This was due to the surge in number of airlines and the decrease in ticket prices due to the competition among them. Similarly, in Turkey, in terms of the number of passengers, the overall increase in 2015 was 9.4 per cent and the increases in domestic and international traffic were 14.1 and 4.4 per cent, respectively. Also, an increase with 8.2 per cent was observed in the overall commercial aircraft traffic. The percentages of the increase in this category for domestic and international traffic have been recorded as 10.7 and 5.6, respectively (General Directorate Of State Airports Authority of Turkey., 2016). The growth of passenger traffic in air travel causes an increase in the ground traffic to airports. Therefore, understanding the issues in ground access traffic to airports and providing solutions for related problems are important. In this study, the ground access mode choice of Atatürk International Airport (IST) in Istanbul, Turkey was analyzed. IST is the largest airport in Turkey and the international hub for Turkish Airlines (THY). In 2015, passenger movements increased in IST by 8.0, 4.0 and 10.0 per cent in overall, domestic and international traffic categories, respectively. In terms of commercial aircraft traffic in IST, the increases for the same categories were 6.0, 0.3 and 9.0 per cent (General Directorate Of State Airports Authority of Turkey., 2016). In the studies regarding mode choice for airport access, discrete choice models (Ben-Akiva and Lerman, 1985) were used to estimate the shares. Among them, the most widely used model was Multinomial Logit (MNL), yet a small amount of studies used Binary Logit (BL) and Mixed Logit (ML) as well.