Hugh Gunn, HGA Limited, Royston, United Kingdom, James Fox, RAND Europe, Cambridge, United Kingdom
Transportation Models at Regional and National Level - Current Practice and Limits to Model Transferability (assigned to theme
Formal transportation models, with demand and supply expressed in mathematical forms and capable of identifying equilibrium flow conditions, were first developed for urban and metropolitan regions. The Chicago Area study (CATS) in the U.S.A. was amongst the first, and was replicated (and enhanced) in Europe in the South-east Lancashire study in England (SELNEC). The focus of these models was to guide the planning of the post-war expansion of the cities and city-regions, in particular in an era of rapidly growing car-ownership. The demand for models of this type at a national scale came later. In the UK, the driving force was the need to develop a national motorway system, which led to attempts to build a mono-modal road-based national model (the RHTM). In France, an equivalent requirement to support the development of a national TGV network also led to a formal model (MATISSE). The desire for multi-modal transport solutions to national transport problems led to the development of the Dutch National Model (LMS). Later, in other countries, concern with environmental issues and energy consumption also led to national multi-mode modelling (eg the Norwegian National Model). These model systems are essentially variants of the approach developed in early urban region analyses, and codified in M.I.T. Working Note 3 – the now ‘traditional’ four-step model involving trip generation, modal split and distribution, and assignment. Most recently, in Europe, the availability (and success) of the national models is generating a wish to develop (intra-country) regional models which are compatible with the national model of that country. By ‘compatible’, it is meant that regional forecasts for a given future scenario (in terms of prices, incomes, jobs and population demographics, etc) are very close to the forecasts produced by the national model for that region. Such a compatibility is needed to allow a sensible outcome to be reached by the successive operation of the different models. Behind the need for this compatibility is the fact that different levels of public authority (supported by different planning groups) need to interact in determining and executing transport planning policy. As a simple example, broad budget allocations between public transport and road building usually take place at a national level, informed by national level modelling of the different networks. When such an allocation is reached, it is a matter for lower-level authorities to turn the broad plans into detailed ones, to turn budgets and accessibility targets into actual rail-lines, bus routes and road schemes. To do this, they need to have recourse to model systems at an appropriate level – regional models, say. This national/regional dialogue needs to be informed by forecasting models with certain key characteristics. For example, the general relationships between supply and demand, between incomes and car-ownership, between relative prices and modal split and so on, should be near-equivalent (if not identical) to those in the national model. And, since this may require iteration between the models, it is highly desirable that this ease of interaction applies to the modelling teams themselves, not just the models. This paper is somewhat eclectic in nature. It sets out to describe some of the integrated regional/national model systems which are in current use or current development, touching briefly on the equivalent higher-level problem of the integration of national models with pan-European models. However, later sections present more detailed experience with a particular family of operational spatial models of transport demand for overlapping or nested geographical areas. The family in question is built from model systems using disaggregate techniques, and the questions tackled are technical ones, concerning the spatial transferability of key relationships.
 | Conference organized through conf-vienna
(copyright Gunther Maier) |