Peak spreading and congestion: some preliminary analysis
Peak spreading describes the broadening of traffic flow profiles in peak periods which can occur in congested urban networks as traffic demand increases. There are two mechanisms which can give rise to the broadening. The first is drivers re-timing their journeys to avoid unacceptable conditions of congestion in peak periods; this is termed 'active' peak spreading in this report. The second mechanism is where increased delays in a network, due to increase in traffic demand without changes in the demand profile, cause longer journey times through the network, and hence the tail of the peak period is extended. This mechanism is termed 'passive' peak spreading here, and is the main subject of this report. The analysis described in this report is of specific features of 'passive' peak spreading on sample networks. The research was undertaken within a larger study for the Transport and Road Research Laboratory and published in a report entitled: 'Limits to traffic growth in congested urban networks' (see IRRD 826722). Analyses were undertaken on two congested urban networks of the magnitude of 'passive' peak spreading with increasing traffic demand. This involved the use of the CONTRAM traffic assignment model to quantify the change in peak traffic profiles on key links in the networks, as overall traffic demand was increased. The magnitude of this 'passive' peak spreading was quantified using a procedure developed by the Greater Manchester Transportation Unit (GMTU) for analysing and forecasting peak spreading in Manchester. Comparisons are made in the report between results obtained by GMTU which reflect mainly 'active' peak spreading and those obtained from the CONTRAM modelling. An example is given of two methods of adjusting peak profiles to account for 'active' peak spreading in future years. Method A follows procedures developed by GMTU while Method B applies to the situation where Method A might be biased (and should not therefore be used), due to the inclusion of 'passive' peak spreading effects. A discussion of the uncertainties and limitations of these methods is included.