PL 63, Assessment of KY Traffic Forecasts

Abstract PL 63, Assessment of Kentucky Traffic Forecasts PI: Greg Erhardt The Kentucky Transportation Cabinet uses traffic forecasts for project planning and design purposes. Inaccurate forecasts can lead to sub-optimal outcomes for both. For example, if too much traffic is forecast, a road may be unnecessarily widened. Conversely, if too little traffic is forecast, a valuable project may be overlooked in the planning process. This research would assess the accuracy of the methods used to forecast traffic in Kentucky. We propose to compile a set of roadway projects that have opened in Kentucky in the past ~15 years, along with traffic counts before and after the project opened, and the change in speed as observed in commercial datasets (Streetlight/HERE). We will compare these observed project outcomes to: 1. The change in traffic volume and speed predicted by the current version of the Kentucky Statewide Traffic Model. 2. The change in traffic volume would be predicted by applying a standard growth rate to the base year traffic count. If we are able to compile a sufficiently large sample of projects, we will apply a statistical technique known as quantile regression to estimate the range of outcomes that can be expected around a forecast with 90% confidence. For example, we might estimate that for a forecast of 30,000 ADT, 90% of post-opening traffic counts fall within the range of 22,000 to 38,000 ADT.