Using Commodity Flow Survey Data to Develop a Truck Travel-Demand Model for Wisconsin

Four similar truck travel-demand (TTD) models were developed for Wisconsin using 1993 Commodity Flow Survey (CFS) origin–destination (O-D) data and a limited number of truck classification count data. First, statewide zonal-level trip tables were developed using the CFS database. Then gravity models for four trip types were calibrated to match the trip-length frequency distributions of the CFS O-D trip tables. Finally, zonal trip productions and attractions were adjusted using an iterative procedure to match assigned traffic volumes with truck classification count data for 40 selected links. The four alternative TTD models differed only in the method used to assign external trips to the external stations. All of the models provided reasonable levels of goodness-of-fit to the 40 selected calibration links, as well as 104 additional count locations across the state. The primary limitation of the four models was that they substantially underestimated the truck trips that crossed the state boundary. The CFS data focused on longer-distance transport of major commodities; therefore, the local truck trips that used the external stations were underrepresented in the model. Simple regression models for estimating trip productions and attractions were developed using zonal population and employment as the independent variables. These improved models were possible because the CFS data made available more detailed initial estimates of trip productions and attractions.