Estimation of travel time variability for cars, buses, metro and door-to-door public transport trips in Santiago, Chile

The analysis of travel time variability (TTV) is attracting attention among policy makers due to the increasing awareness that users assign a high value to level-of-service attributes. In this paper, the TTV of cars and public transport trips is analysed. We estimate the effect of each trip stage on the TTV for complete door-to-door public transport trips, including access, waiting, transfer and in-vehicle time. We employ data from Santiago, Chile, in which surveyors performed predetermined trips and recorded each stage on several days between 2007 and 2011, which were complemented by recorded bus GPS data. We found that (i) bus waiting and in-vehicle times are highly significant in explaining total (door-to-door) TTV relative to metro (subway) travel times, whereas walking time is not significant; (ii) metro travel time is generally more stable but may be more skewed compared with the travel time of buses on a segregated right-of-way; and (iii) buses that travel in mixed traffic have not only a larger mean travel time but also a larger variability than buses that travel in bus lanes and segregated busways. Formal cost-benefit analysis should consider the effect of (total or partial) segregation of public transport operation on reducing travel time variability.