Emergent route choice behaviour, motorway and trunk road network: the Nantes conurbation

Road network models have traditionally characterised network performance in terms of an average travel time associated with each link in the network, which varies according to the level of traffic using the link. However, this characterisation ignores the influence of network geometry on route choice behaviour. Previous Space Syntax research in inner city contexts has used topological distance minimisation together with road capacity or compared metric, topological and geometric distance showing a high degree of association with vehicular flow level using spatially dense counts dataset. It has been shown that road network geometry strongly influence the traffic flow pattern through quantifiable properties of the network geometry, where topological distance minimisation and effective road width account for the majority of the variance in flows from street to street (r2 = 0.8). Such studies have focused on high-density urban environments and street network. The originality of this paper is to examine this relationship in the context of sparsely built suburban areas and motorways, trunk roads and feeder roads. The case of Nantes conurbation is used. Population density, activity locations, shop and supermarket locations and commuters’ journey length distribution are mapped and analysed qualitatively. The spatial model covers an extent of 570 km2 - including a 12 km diameter orbital motorway, several surrounding communities, and more than 700,000 inhabitants. This paper compares generic path selection link assignment using three different distance minimisation approaches: metric, topologic and angular. Using an automated vehicular counts (156) dataset on motorways, trunk and main roads, accounting for more than 3.500,000 vehicles/day, it was found that the correlation between topological closeness compounded with road capacity and vehicular flow levels increased with every step upwards in radius, peaking at r2 ~ 0.84 at radius-n. The study uncovered for the first time the role of axial entropy in relation to the motorway network. The results of this study suggest that road network geometry analysis can be a powerful predictor in low-density environments with a relatively sparse count dataset. The paper concludes with a discussion on the use of such analysis to strategically assess vehicular network design in relation to road network performance and interface, journey length reduction and road network legibility