This project developed a demand-based forecasting model for rural and highway road transportation planning to assist decision-makers in predicting transportation demand flows of wheat in the Pacific Northwest. In order to be sensitive to policy changes, the model is based upon the disaggregate, individual shipment decisions of wheat elevators. The research found that the mode/market selection process significantly affects the usual regression estimates of mode/market demand flows. Therefore, estimates of the parameters on the major determinants of transportation demand were obtained by weighted least squares that were corrected for selectivity bias. Procedure was developed to aggregate the disaggregate predictions into total regional flows. This procedure retains the policy-sensitivity of the disaggregate model and is computationally practical in applications. Finally, the aggregate forecasting model was developed into an interactive computer program which can be used in a variety of policy-relation applications.