Abstract
The forecasts produced by the mathematical models used in the urban transportation planning process are subject to uncertainty. At present there is no satisfactory method of estimating the uncertainty in the forecasts. A new method of estimating uncertainty is presented.