Improving Average Travel Speeds Estimated by Planning Models

Modeling the impact of air quality requires estimating vehicle volumes and average travel speeds in order to estimate air pollutant emissions.  Relatively accurate estimates of average vehicle travel speeds can be obtained from existing traffic operations models.  However, when the travel demand impacts of the highway project extend over many miles of the freeway and arterial street network, it is not feasible to apply these detailed traffic operations models to the entire area of impact because of their extensive data requirements.  Planners then must resort to areawide planning models to forecast vehicle volumes and average travel speeds.  These traditional planning models, however, are not calibrated to produce accurate speed estimates.  Free-flow speeds and a speed-flow curve are input into those models and adjusted as necessary to obtain calibrated volume estimates.  Typically, the reasonableness of the final travel speeds is not checked once reliable volume forecasts have been achieved.  A postprocessor methodology that can be applied at the end of a typical planning model forecast process to improve the estimates of travel speeds output by a planning model is proposed.  The methodology uses an improved speed-flow curve and queueing analysis to obtain travel speed estimates that more closely approximate the average speeds estimated by typical operations models.  The proposed methodology was applied to a real-life highway network and the results compared to FREQ and TRANSYT-7F simulations for a 5-mi section of freeway and a 3-mi section of a four-lane divided arterial street.  The postprocessor significantly improved the original planning model estimates of average speed and delay.