Planning policy documents increasingly promote the ‘polycentric city’ as a model for sustainable urban development, not the least in Finland. There seems to be a widespread assumption that polycentric intra-metropolitan settlement patterns with compact, transit-oriented cities and neighborhoods will reduce travel distances, discourage car driving and promote the use of public and non-motorized modes of travel. But how well do these assumptions fit with state-of-the-art knowledge about the influences of land use on travel?
Will polycentric urban development reduce travel distances?
The short answer is to this question is no. Residents travel on average longer distances for commuting as well as most other trip purposes the farther away from the main center of the urban region they live – up to a turning point beyond which travel distances tend to be reduced if the residential distance to the city center increases even more (Figure 1). This has to do mainly with the greater concentration and density of workplaces, stores and other facilities in the inner than in the outer parts of the metropolitan areas. The more or less radial structure of the main road and rail networks and the fact that the point of gravity of all trip destination addresses in the region is usually located close to the main city center also contribute to this.
Figure 1: Curve showing how commuting distances tend to vary with among workforce participants living at different distances from the city center of Oslo. N = 1160. Source: Data from the RESACTRA project.
Proponents of polycentric urban development might argue that travel distances among suburbanites could be reduced if more workplaces and service facilities were decentralized. While it is true that non-specialized facilities such as grocery stores, primary schools and kindergartens should preferably be located close to the residential neighborhoods they are intended to serve, decentralization of specialized facilities will only lead to longer travel distances. This is especially so for specialized workplaces. Decentralizing of workplaces to a suburban sub-center may reduce commuting distances among the local residents, but most companies recruit their employees from a much larger geographic area than the local neighborhood. There is little overall variation between workplaces located at different distances from the city center in average commuting distances, although employees of workplaces close to a second-order center do commute shorter distances than employees at other peripheral workplaces.
Will polycentric urban development discourage car driving?
In most city regions, decentralizing residential development to polycentric nuclei contributes to more car driving than if the new dwellings were built as inner-city densification. Although suburban residents living close to a second-order center travel less by car than suburbanites living far away from a second-order center do, especially if they live in the outer suburbs, their car travel distances are substantially longer than those of their inner-city counterparts (Figure 2).
Figure 2: Weekly travel distance by car for commuting and seven non-work purposes among workforce participants living at different distances from the city center of Oslo and from the closest second-order center. N = 1061. Source: Data from the RESACTRA project.
Increasing the jobs-housing balance in the suburbs by decentralizing jobs will not decrease car driving, but increase it. Accessibility by car is usually much easier in the suburbs and exurbs than in the inner city. In the suburbs, roads are normally wider and less congested, with fewer light-regulated crossings and easier parking conditions. As a result, the proportion of employees and visitors who travel by car to outer-area jobs and service facilities is much higher and overall distances commuted by car longer than for workplaces and facilities located in the inner city, even for job locations close to second-order centers (Figure 3).
Figure 3: Proportions of regular car commuters (to the left) and mean weekly distance commuted by car (to the right) among employees of workplaces located at different distances from the city center of Oslo and from the closest second-order center. (Please notice that no workplaces located closer than 2 km from the city center of Oslo are located less than 1 km from the closest second-order center.) N = 1160. Source: Data from the RESACTRA project.
Will polycentric urban development promote the use of public and non-motorized modes of travel?
Mirroring the different shares of car commuting among employees of workplaces differently located, the proportions of regular public transport commuters are much lower at suburban than at centrally located workplaces (Figure 4). This is true also when the suburban workplace is located close to a second-order center, although such proximity contributes to somewhat higher shares of public transport commuters than among the employees of other peripheral workplaces.
Figure 4: Proportions of regular public transport commuters among workforce participants living at different distances from the city center of Oslo and from the closest second-order center. N = 1160. Source: Data from the RESACTRA project.
A similar pattern exists for non-motorized travel. Measured as the share of total travel distance, there is a steep center-periphery gradient, also among residents living near a second-order center (Figure 5). Living close to a second-order center still contributes to higher proportions of non-motorized travel among the suburban residents.
Figure 5: Non-motorized proportion of total travel distance for commuting and seven non-work purposes among workforce participants living at different distances from the city center of Oslo and from the closest second-order center. N = 1061. Source: Data from the RESACTRA project.
Polycentric intra-metropolitan settlement patterns with compact, transit-oriented cities and neighborhood perform better, judged from a sustainable mobility perspective, than ordinary outward urban spatial expansion does. However, polycentric development is far less favorable than densification close to the main city center if the aim is to reduce travel distances, discourage car driving and promote public and non-motorized travel. Polycentric intra-metropolitan urban development is likely to increase car travel per capita instead of reducing it.
Professor Petter Næss, Norwegian University of Life Sciences. E-mail: firstname.lastname@example.org