The Impact of Land Use Patterns on Travel Behaviour

The Impact of Land Use Patterns on Travel Behaviour


M Hanly, J Dargay, ESRC Transport Studies Unit, UK



This paper presents the results of the first stage of a project to examine the role of land use characteristics in determining travel behaviour and mobility. Land use characteristics include population density, level of rurality, urban size, distance from urban centre and the spatial distribution of housing, employment and facilities for education, shopping, services, leisure activities, etc. The number of trips made, the distance travelled and mode choice are all related to land use patterns. For example, higher population densities should lead to shorter, but more frequent, trips, and a greater use of public transport and non-motorised modes, while longer, but less frequent trips would be associated with more rural areas. Local access to jobs, shops, services, leisure activities and other facilities will result in shorter trips than when accessibility is low, but it is likely that the number of discretionary trips will be greater since the greater distances involved in areas with low accessibility will discourage unnecessary trips and encourage a greater efficiency in the trips actually made.

The influence of land use patterns on travel demand cannot be examined on an aggregate level, but only by analysing the travel behaviour of individuals. Because of the inter-relationships between the determinants of travel demand, it is necessary to base the analysis on a comprehensive model which includes the wide-range of factors which influence the individual?s travel behaviour. Thus, socio-economic factors, such as income, employment and household composition and the characteristics of transport supply, such as the availability of roads and parking facilities, access to cars and public transport, and prices of vehicles, fuel and public transport should also be considered. By investigating the relative importance of these factors on travel, inferences can be drawn regarding the impacts of land-use patterns and alternative policy measures. This is clearly of utmost importance for developing an effective strategy for addressing the external costs associated with private vehicle travel and for the formation of a sustainable transport strategy.

The analysis will be based on data for individual households obtained from the UK National Travel Survey (NTS). The NTS contains information on travel patterns (the number of trips, distance travelled and travel time by mode and journey purpose), vehicle ownership and a large range of socio-economic factors (income, household structure, employment, etc). In addition, there are a number of land-use indicators of the area of residence: population density, distinction between rural/urban area, size (in terms of population) of urban area. There are also a number of variables concerning access to services: walk and bus time to grocers, High St shops, chemist, doctor and hospital; and access to public transport: walk time to bus, frequency of bus and walk and bus time to railway station. Since these accessibility and frequency measures are as reported by the respondents, it is likely that they contain errors, particularly for those who do not walk or take the bus to use the services. It is thus necessary to take such measurement errors into account in the model estimation.

In the initial stage of the project, presented in this paper, models will be estimated for total mobility (the number of trips and the distance travelled) by all modes, as well as for and commuting and non-commuting trips, separately. The explanatory variables included in the models will be limited to the information available in the NTS, so that travel costs and certain transport supply variables will unfortunately be excluded. This should not pose too much of a problem, since we are concerned with trips by all modes, rather than for individual modes. The models will be estimated for data from surveys carried out at different points in time, in order to explore whether the influence of land-use and other variables has remained constant or has changed over time.

Further stages of this project will examine the implications of land-use patterns on mode choice. It is also intended to explore the possibility of combining the NTS data with additional land-use data using GIS. This project also contributes to an international comparative analysis of the influence of land-use characteristics on personal travel in the USA, Netherlands, and the UK. Since the comparability of data across countries is a limiting factor in the international analysis, the study for the UK presented in this paper is based a more comprehensive model than that used for the international comparison.


Association for European Transport