LIMITS TO TRAFFIC GROWTH IN CONGESTED URBAN NETWORKS

Traffic performance measures under conditions of increasing demand have been assessed on five networks of differing characteristics using the traffic assignment model CONTRAM. The time-dependent nature of the modelling has enabled the circumstances to be identified when the network is effectively saturated. Monitoring of queue lengths in each time interval has also indicated the levels of traffic demand at which queues fail to recover (or reduce) by the end of the modelling period. This indicates queueing instability and, with extended modelling to include subsequent off-peak periods, maximum traffic growth consistent with longer term queue stability could be identified. While the above criteria are useful in practice, the limit of traffic growth in a network is more likely to depend on behavioural responses to traffic conditions and the availability of alternatives. These considerations have led to a hybrid approach to saturation level prediction being developed in this study. This consists of initially stopping traffic growth on Origin-Destination (O-D) movements when average travel speed on set routes on those movements fall below minimum tolerable values. A methodology is described on three networks, and is shown to be practicable. Using 15 kph minimum speeds for illustrative purposes, the gradual redistribution of traffic demands in the network with increasing growth caused network saturation to be predicted a few years later in the evaluation period. This study has highlighted a number of areas of uncertainty and further research is recommended into: (1) methodology application; (2) sensitivity testing; (3) peak spreading; (4) the use of alternative methodologies such as SCOOT or the use of matrix manipulation methods; and (5) behavioural responses.