In forecasting it is usual to assume the temporal stability of the validated route choice parameters used in highway assignment models. Recent experience in the Greater Manchester area indicates that driver behaviour as measured by these parameters has changed markedly and that these changes can have a substantial effect on traffic forecasts. These effects can amount to a doubling of forecast traffic levels in some circumstances. These effects are discussed in relation to forecasts made for a particular motorway scheme in 1983 and shortly before the opening of the scheme in 1989. The two sets of forecasts are compared with each other and with the traffic flows actually realised on the scheme. The recent forecasts compare well with the observations, whereas the older forecasts are substantial underestimates. The differences between the forecasts are shown to be almost entirely explained by the change in the validated time/ distance ratio that defines the 'generalised cost' in the assignments for the two forecasts. The validations took place in 1981 and 1986 and in that period the ratio has halved. The change in time/distance ratio is much greater than might have been expected from the relative change in petrol prices and GNP, although it follows the same trend. The implications of these changes for future forecasting are considerable as the current National Road Traffic Forecasts assume changes in the relationship between fuel prices and GNP between the present day and the end of the century of a similar magnitude to the actual changes between 1981 and 1986.
Samenvatting