Congestion Effects in Transport Modelling and Forecasting

Abstract: Transport investments and policies are increasingly turned towards dealing with transport congestion rather than with shortening the potential free flow travel time. However, appraisal methodologies for projects meant to reduce congestion are relatively less well developed compared to methodologies for projects aiming to reduce travel times. Static assignment models are for instance incapable of predicting the build-up and dissipation of traffic queues and capturing the experienced crowding caused by uneven on-board passenger loads. Despite of the availability of dynamic traffic assignment and despite of fairly concrete ideas of how integration with demand models could take place, only few model systems have been developed for real applications.The predicted reduction of traffic volume across the Gothenburg congestion charge cordon in the peak, 11%, turned out to be an accurate estimate of the observed reduction, 12%. The reduction in the off-peak, however, was overpredicted, as it was also in the Stockholm case. To analyse congestion charges in Stockholm it is necessary and fully possible to integrate DTA with the demand model. In the performed tests it could be seen that both tested models had problems replicating the flow on the main bypass early in the morning but otherwise performed well. A case study of a metro extension in Stockholm demonstrated that congestion effects constitute more than half of the total benefits and that these effects are excessively underestimated by a conventional static model. Effects of various operational measures can be analysed with BusMezzo and the results have been validated against observed data. The findings indicate that all three tested measures in a case study (boarding through all doors, headway-based holding and bus lanes) had an overall positive impact on service performance and that there are synergetic effects.Using a continuous VTT distribution and hierarchical route choice was demonstrated as a successful method of modelling the multi-passage rule implemented in Gothenburg congestion charges and was shown to give realistic predictions of route choice effects. First results from integration of DTA with a travel demand model for the Stockholm region show that even without systematic calibration the DTA is in reasonable agreement with observed traffic counts and travel times. The presented experiments did not reveal a striking difference between using a macroscopic and a microscopic assignment package. While travel time savings are often the only benefit included in public transport project appraisals, the best practice assigns weighted value of time to average load/capacity measures. However, failure to represent dynamic congestion effects may lead to substantial underestimation of the benefits of projects primarily designed to increase capacity rather than reduce travel times. The impact of small operational measures should not be underestimated. These measures are relatively cheap compared to investments in new transit infrastructure and large societal gains can therefore be achieved by their implementation.