How different network disturbances affect route choice of public transport passengers. A descriptive study based on tracking

Public transport networks are affected daily by disturbances with different entities. While big service distur-bances (or disruptions) are rare, delays and cancelled runs are more frequent and they affect daily the passengers. Despite this, most of the work on passengers' behaviour in case of disturbances focuses on big disturbances and not on common and daily ones. In this work, we analyse how different network disturbances affect public transport passengers, regarding the chosen route and the travel cost. For this purpose, we exploit a large-scale travel survey based on GPS tracking, and AVL data of public transport operations in Zurich, Switzerland. We quantify the disturbances in the network, which are relevant for a specific passenger's trip, with a metric of service degradation. In that way, we can analyse passengers' route choices in case of different disturbances. In particular, our study evaluates the effects of disturbances on travel costs, comparing the passenger's chosen route with three expected behaviours ("timetable", "no information", "full information"), based on the available al-ternatives with and without disturbances. Our analysis identifies that small disturbances and delays have a significant effect on travel cost, although marginal effects on route choice, since 72% of times users stick to the most likely route in planned conditions. In contrast, "good disturbances", i.e. variations of the operations from the timetable, which generate less costly alternatives, have a significant effect on route choice, since only 55% of times users choose the most likely route. Moreover, we identified that passengers do not exploit new available alternatives, suggesting a need of better information in these cases.