Evaluating urban form and location effects on cycling based on Danish micro-data.

Abstract

Extended abstract The paper analyses the environmental correlates of cycling based on Danish transportation and urban form micro-data. Established walkability factors such as density, connectivity and diversity are related to cycling, but access to retail concentrations/centers, public transportation level-of-service, as well a competition between walking and cycling depending on environmental features can be added Attractive conditions for using public transportation or walking relate to less cycling. Highly relevant in the context of cycling promotion the results points to some opposite effects of urban form on the probability of cycling and distance cycled. A high probability of cycling generally imply short distances to cover and less cycling – leading to non-uniform, non-monotonous relations between environmental indicators such as walkability and aggregate cycling. Promoting cycling can be seen as a potential means to reduce congestion and increase the sustainability of transportation, but also as a means to increase physical activity and public health. Consequently the factors conditioning cycling have received increasing attention in public policy as well as in research. The urban environment is one of the factors conditioning transportation need and choices and this has been the focus of numerous studies, often taking a congestion management or sustainability perspective on personal transportation. Promoting physical activity in the form of cycling and walking also relate to the urban environment, and the concept of ‘walkability’ has been added to research and practice to describe the degree to which the urban environment support active modes of travel. There is, however, still limited research on what is known about the interaction between land use, destinations, infrastructure and the willingness to cycle (McNeil, 2011). Cycling as travel mode or activity may be placed somewhere in between ‘street life’ which is highly sensitive to environmental qualities, and a more utility oriented mode of travel moving between origins and destinations. Following Forsyth and Krizek (2011) it can be argued that there is a distinctive view from the bicycle, a difference from the walking mode together with which cycling is often classified and analyzed – warranting further research into the factors preconditioning cycling. Additionally recommendations for research have pointed to a need for elaboration of the associations between the built environment and different forms of physical activity (Handy, 2005) – something which this article aim to contribute to by focusing on cycling. The aim of the paper is to analyze the urban form and location factors that precondition and explain cycling in Denmark. Important questions related to this, and of particular interest from a cycling promotion and urban planning perspective, is the significance of urban form and location factors and which trade-offs between cycling and other transport provisions that appear from the results. The paper is based on a detailed micro-level representation of urban form and location, including neighbourhood level as well as regional location aspects. A sample selection model of the probability of cycling and distance cycled is presented. Model results include a testing of the robustness of urban form correlates of cycling to model specifications including control variables with more or less assumed endogenity towards cycling outcomes. The model results points to a high probability of cycling being related to flat terrain, short distance to a retail concentration, population density, network connectivity within a 1,5 km ‘personal’ neighbourhood. Transit access and level-of-service measured as the availability of a train station within 1000 meter and the number of daily bus and train departures within 500 meter from the dwelling both contribute to lower probabilities of cycling indicating that public transportation to some degree is a competing alternative. Retail jobs per resident within a very convenient 500 meter walking range indicating main street and central area locations, as well as network connectivity within the 500 meter range also contribute to lower probabilities of cycling – indicating that walking is also a competing alternative towards the bicycle. A long cycling distance per day is related to short distance to a large retail concentration, low population density, and low network connectivity within walking range. Additionally the sample selection model results indicate that when the probability of cycling increase – the daily cycling distance decrease – most likely because the convenience of cycling as travel mode rely on short distances. The correlations between cycling and urban form and location factors are generally robust and sample selection model specifications with and without control for mode access (bicycle; car ownership; and public transport travel pass) in addition to other demographic and socio-economic variables produced only marginal changes to the set of statistically significant variables or their parameter estimates. The only exceptions were the urban form variable: number of retail jobs pr. resident within walking range which turned insignificant when the respondents access to a bicycle is included in the set of control variables This indicate that people living in very central locations more often do not possess a bicycle which again may be a combined effect of the convenience associated with walking, higher risk of bicycle theft (or more parking troubles), or of course transport/residence related preferences. Additionally the parameter estimate for the effect of population density upon cycling distance changed significantly as a result of including mode access variables in the analysis – suggesting the strongest effect of density when controlling for mode access. Although direct inclusion of residential self selection issues in the analysis has not been possible based on the available data, this result indicate that the inclusion of preferences will not necessarily lessen the effect of urban form and location aspects on cycling. From a cycling promotion for sustainable transportation or public health perspective – the results points to important tradeoffs between factors related to the probability of cycling and the daily distanced cycled. The effects of urban form and location aspects does not point in the same direction for both – leading to a question of the desirable cycling outcome if cycling is to be promoted through spatial planning and policies.