After opening of new transport infrastructure often more traffic is observed than originally predicted. Such traffic is generally called induced traffic, pulled into the transport system because of the improved levels-of-service. A number of methods are used in practice to estimate this induced element. However, the definition of induced traffic varies widely between studies, and depends critically on boundaries used in the model or in the analysis. In this paper a functional framework is presented that distinguishes between explicit, distributive model components and implicit, generative elements. These elements may be viewed as the induced element in an overall analysis. It is clear that the model structure may indeed determine both the practical definition and the magnitude of induced traffic. Where particular responses cause significant changes in demand, these should be represented in an explicit, distributive manner, rather than through generative approaches, as induced traffic. Rather than rigid rules for the definition and estimation of induced traffic, a clear and explicit statement of boundaries used in the analysis for geographical coverage, behavioural responses and time horizon, should remove much of the current controversy in this area. (A)
Abstract