Thursday, June 29, 2006

Service frequency theory and a transport system's role

Along with routes and operating hours, service frequency is a major determinant of its attractiveness to passengers and its suitability for various trips. The following is an attempt to assign capability to specific service frequencies that may be found on a network.

5 min: turn up and go providing instant service just like turning on a tap. Go anywhere any time with no timetables needed. It's stress free (not worth running if you see a service arriving) and connections are always fast. Shelters, seats and other facilities become less important. Operationally, maintaining even headways is more important than time adherance. Very high potential to attract people out of their cars.

10 min: Turn up and go, go anywhere any time, timetables less important. Main passenger information item is a network map. Extremely suitable for peak period travel. Car-competitive and good potential for modal shift.

15 min: Sort of turn up & go. Clockface. Requires no prior planning and allows spontaneous changes of plan. Transfers are timed but no big deal if connections are missed. Little need to live life around timetables. Suitable for work trips, leisure trips and short errands. An acceptable minimum standard for complex capital-intensive urban rail systems with many lines and transfer combinations. People cannot plausibly claim a 'lack of service' as a reason for non-use. Semi-car competitive - car travel is no more than twice as fast and can be slower.

30 min: Clockface. Requires some prior planning. Transfers are timed. Missed connections a worry when transferring. Waiting becomes a large component of trip time and car travel is 2-3 times quicker. 'Lack of service time wishes to travel' may be valid. However, provided connections aren't required or are seamless, suitable for leisure trips with flexible finish times. Can also work for medium-long length shopping trips (>30 min). Limited potential for mode shift from car owners.

60 min: Clockface. Transfers must be timed. Inflexible. Does not allow spontaneous changes of plan. Timetables far more important than network maps. Can work with long shopping trips or if impeccably planned. If not planned, waiting forms majority of trip time and car travel is 3-10 times quicker. Use on an electrified suburban system may represent poor utilisation of sunk capital and infrastructure. 'Lack of service at time one wishes to travel' a valid argument. Negligible potential for mode shift from car owners.

120 min: Require extensive planning of forward and return trips before trip is started, and thus reference to paper or online timetables. Transfers must be timed. Extremely inflexible - must plan life around timetables and can do little else during day. Lack of service excuse almost always valid. Not used by car owners except in emergencies.

2-3 services per day: May be adequate for low-density rural areas, but represents a 'charity' service inappropriate for suburban areas. Fulfills a welfare role for people without cars. Transfers are limited and often fortuitous. Extremely inflexible.

I should add that the above times apply for routine trips within an urban area. I'd expect similar principles to apply for other types of travel. However in all cases the proportional travel time advantages of increasing frequency fall once in-vehicle journey time exceeds several times the service frequency.


melbourne train girl said...

Thanks for the link!

Anonymous said...


A very useful analysis particularly considering the lack of holistic improvement for bus and rail frequencies within Melbourne.

For example, weekday off-peak rail services (30 minute frequency) beyond Ringwood are actually less frequent than weekend services (20 minutes) resulting in the need for trip planning despite the roles of Ringwood and Croydon as 'activity centres'.

Alex Makin