The modeling of freight travel has gained attention in the last two decades due to the importance of freight transportation to economic growth, interstate and international trade and its impacts on the transportation network. The Utah Department of Transportation (UDOT) commissioned the development of a statewide freight model to assist in statewide long-range planning and to better assess roadway impacts of trucks. This freight model is integrated with the Utah Statewide Travel Model (USTM), which incorporates the State’s metropolitan model networks, socioeconomics and trip tables.
This presentation covers the methodology along with calibration and forecasting results and discusses lessons learned to assist other regions seeking to develop freight demand modeling approaches. The methodology includes a discussion of the data analysis, the development of a dynamic in-stream merging of the national and Utah traffic analysis zone (TAZ) and highway network data, the short-haul commercial vehicle forecasting, and the long-haul commodity flow forecasting. The short-haul model produces light, medium and heavy commercial vehicle trips and trips are distributed at the TAZ level inside Utah. The long-haul model forecasts tonnage by 12 commodity groups utilizing Transearch commodity flow data. Multivariate and multi-tier regression analyses using some advanced outlier-detection methods are employed to determine the relationships between employment type and new tonnage produced or attracted to a zone. Commodities are distributed nationally to Bureau of Economic Analysis (BEA) zones outside of Utah and to TAZs inside Utah and external flows are estimated using a sub-area extraction from the merged national-Utah network. Truck and truck-rail intermodal shares are determined from Transearch data. The Vehicle Inventory and Use Survey (VIUS) is used to convert the commodity flows from tonnage to trucks and to provide empty truck return rates. Both short and long-haul data are distributed via a gravity model and multimodal commodity flows are split by mode. Assigned truck trips are validated to statewide truck counts, provided by UDOT, with attention given to primary freight corridors. Average trip lengths by truck type were reviewed for reasonableness. The successful validation of each model step suggests that the methodology can be transferred to other states with relative ease.