The purpose of this study was to forecast grain and oilseed shipments through the Mississippi River system and to analyze impacts of delay costs. The focus is on the world grain trade and expected changes in response to a multitude of evolving competitive pressures and structural changes. The model is a spatial optimization model of the world grain trade. Important parameters are forecasted and used to evaluate changes in flows through specific logistical channels. Projected import demands are based on consumption functions estimated using income and population and accounting for intercountry differences in consumption dependent on economic development. Each of the competing supply regions and countries were represented by yields, area potential that could be used in production of each grain, costs of production, and interior shipping costs, where relevant. Crucial in this model is the interior spatial competition between the U.S. Pacific Northwest and shipments through the U.S. Gulf, as well as inter-reach competition. Delay costs are the additional costs associated with shipping on the barge system and result from queuing and the added costs for shipments that are delayed. These are an important feature of barge shipping, particularly when shipment volumes are greater. In several of the reaches, grain flows are near the point at which positive delay costs are accrued. At higher volumes, delay costs escalate and ultimately become nearly vertical. The latter is an indicator of capacity, i.e., the level of volume at which the delay costs become perfectly inelastic. The results suggest exports from the United States increase from the base period to 2010, in part due to the assumption that the maximum area for plantings would increase and in part due to the fact that China�s corn exports are assumed nil in 2010 and beyond. U.S. corn exports decline the most, with a potential peak of 62 mmt to around 42 mmt. Wheat exports decline substantially, but soybeans increase through 2030. Exports from the United States are concentrated in the U.S. Gulf, which declines to 57 mmt after reaching a peak of 92 mmt in 2010. Exports from the Pacific Northwest (PNW) are 25 mmt in the base year and decline in later years. The results illustrate that the United States remains an important exporter of soybeans and this conclusion persists in other scenarios. Expansion would result in reduced delay costs on each or Reaches 1, 2, and 4 by about $0.44/mt, $1.04/mt, and $1.01/mt, respectively. Expanding lock capacity reduces delay costs, increases capacity and shipments by barge. Barge shipments increase by about +4 mmt by 2020. Thereafter, the change in barge shipments would be about +0.9 mmt to +2.5 mmt. There is substantive inter-reach competition and by 2020 shipments on Reach 1, 2, and 4 increase, but shipments on Reach 5 and 6 would decrease. Delay costs, in aggregate, are comprised of the lower delay costs that would occur at current capacity, plus the volume effect. The impact of expansions on delay costs are in the area of $61 million, inclusive of both direct effects. Most of this is accrued on Reach 4, followed by Reach 2 and 1. Expansion results in an increase in barge costs due to the increase in volume, a decrease in rail shipping costs, and a slight increase in ocean shipping costs. In total, the impact of expanding locks is a decrease in costs by about $52 million. The base case assumed Energy Information Administration (EIA 2005) projections of corn use in ethanol demand. The model was revised assuming the EIA (2006) estimates of ethanol produced from corn. Results are drastic. Exports from Argentina, Europe, and Eastern Europe increase and wheat exports from Australia increase and exports from the United States decline. Gulf exports decrease (65 to 51 mmt) and PNW changes only slightly to 14 mmt. Most of the decline is in corn and wheat shipments. In addition, there are major changes in flows within the United States. Most interesting is the increase in shipments to the Eastern and Western Corn Belts, reflecting the increase in domestic demand for ethanol use. Also interesting are the changes in flows from the Northern Plains which had previously exported most of its corn through the PNW. These are now shifted with a significant portion destined for domestic movements. The model was run assuming more stylized assumptions for some critical variables, mostly impacting the ability of corn production to expand to meet these competing demands. The results suggest the model is fairly robust in capturing these different assumptions. Most striking in making these comparisons are: 1) corn exports from the United States increase; 2) soybean exports from the United States decline to 28 mmt vs. 36 mmt for the 2010 case; and 3) wheat exports increase from each of the competitors and those from the United States decline, but by not as much as in our unrestricted high-ethanol case. There are important reasons for these differences. Most important are assumptions about the yield growth, the ability to expand corn acres, and differing assumptions on soybean production and exports from the United States vs. competitor countries.