Forecasting temptations: The spatial dimension of traffic modelling There is a clear relation between spatial structure and traffic behaviour. In existing traffic models, this is not reproduced well. Differences in density, centre distance or function mix lead to differences in traffic performance and modal split of 50% and higher. Furthermore indirect effects occur by location preferences of different social groups. To integrate this in traffic models, the parameters of all steps have to be estimated in dependence of the spatial structure: (1) perceived distances instead of synthetic resistance matrix, (2) allocation of social groups by spatial factors, (3) estimating of trip production parameters (trips per inhabitant etc.) per spatial and social group discretely and (4) determination of the preferential distance and transport mode dependent of these groups. The pilot model of the Arnhem-Nijmegen region according to this approach is promising. The application of traffic saving scenarios on this model leads to small car traffic reductions. However, less car dependent structures are advisable for other reasons: more choice, better accessibility, more participation of people without car and the possibility for a car reducing policy in future. To examine this, a fictive cost increase for an increase of car resistance is a good indicator. (Author/publisher)
Abstract