Distance-measure impacts on the calculation of transport service areas using GIS

Access coverage is important in public transit planning, as this is the means by which service is provided to riders. In fact the proximity of demand (population and employment) to stops or stations on the network to a great extent explains its greater or lesser usage by potential users. Coverage of service areas can be delineated by GIS through the creation of buffers around transit facilities based on Euclidean (straight-line) distance. A second method is based on calculations of distances along a street network (network distance). The choice of the distance calculation method affects significantly the final results in terms of population covered. This paper assesses the overestimation of the straight-line-distance method, which is the most widely used in coverage analysis, by comparing it with that of network distances. It investigates systematically the factors influencing this overestimation, such as the density of stops or stations, the coverage distance thresholds and the characteristics of the area analysed (street-network design, barriers, and population distribution in the neighbourhood of the bus stop or station). Finally, it concludes that the network-distance method provides systematically better estimates of transit ridership than the Euclidean distance method.