Accurate traffic forecasts for highway planning and design are crucial for ensuring that public dollars are spent wisely; however, there is only a small library of empirical studies that have examined traffic forecasting accuracy in the United States. Even with limited availability, these studies are important as they address three critical benefits: insight on observed inaccuracy levels to decision makers and the public, a demonstration of the value of advanced models and data techniques, and assistance in identifying new or advanced methods to improve traffic forecasting practice. Such studies are rare because of numerous challenges, including data availability and staff turnover, and because of a lack of consistency in accepted procedures for preserving detailed information on forecasting methods and data used in the analysis. These challenges are slowly being addressed as the importance of empirical accuracy reporting has grown. The need for the demonstrated value of advanced modeling and data techniques has also grown, as these techniques require significant resources. In traffic forecasting, departments of transportation in Wisconsin, Minnesota, and Ohio have conducted targeted reviews of some traffic forecasts within the past 6 years. Other fields have demonstrated the effectiveness of such reviews, most notably the National Oceanic and Atmospheric Administration (NOAA) through their highly successful Hurricane Forecasting Improvement Program. Building on these earlier, limited reviews, especially in the context of improving technology, there is a need to expand documentation and assessment of traffic forecasting experience to improve future applications.
The objective of this study was to develop a process to analyze and improve the accuracy, reliability, and utility of project-level traffic forecasts. To meet this objective, addressing the following components: (1) analysis of traffic forecasting accuracy and usefulness using information from various sources including, but not limited to, state departments of transportation (DOTs), metropolitan planning organizations (MPOs), counties, and other transportation agencies actively involved in forecasting travel demand in competitive modes; (2) assessment of transportation agency experience with respect to accuracy of various forecasting approaches; (3) identification of methods for improving flexibility and adaptability of available forecasting techniques to changing assumptions and input data; (4) consideration of risk and uncertainty in the forecasts; and (5) identification of potential methods to help the traffic forecasting industry improve forecasting usefulness and accuracy while improving their ability to communicate and explain these forecasts to affected communities. In the context of this study, the terms accuracy and reliability are meant to address how well the forecasting procedures estimate what actually occurred; utility is meant to encompass how well a particular projected outcome informs a decision; and project level is meant to include a single project or a bundle of related projects. The product of this research provides guidance to MPOs, state DOTs, and others to improve the accuracy, reliability, and utility of traffic forecasting methods as applied to transportation planning, design, and operation efforts—both short and long term.
To meet the objectives of this study, the research plan addressed the following:
1. Developing metrics and processes for evaluating traffic forecasts and procedures;
2. Evaluating traffic forecast plausibility, accuracy, utility, and sensitivity across several dimensions:
a. Forecast method; e.g., travel model, linear regression, trend analysis;
b. Size and location of forecast study area;
c. Project size, design, and type, including HOV and HOT;
d. Project stage and decision points;
e. Functional class, facility type, and volume; and
f. Forecast horizon—short, medium and long term.
3. Determining under what conditions forecasting accuracy improves, including agency experience;
4. Enumerating contributors to forecast inaccuracy, and suggesting mitigation methods;
5. Evaluating methods for improving communication of forecast risk and uncertainties;
6. Exploring optional forecasting procedures or methods from other industries or institutions; and
7. Providing guidelines in support of an ongoing review of forecast accuracy, including a recommended policy for documenting and retaining models, associated analytical techniques, and sufficient data to allow replication of earlier applications.