An economically efficient biorefinery could be expected to require a steady flow of feedstock throughout the year. The logistics of providing a daily flow of several thousand Mg of bulky biomass could be challenging. Hwang (2007) enhanced a model (Tembo et al. 2003; Mapemba et al. 2007; Epplin et al. 2007; Mapemba et al. 2008) designed to approximate a just-in-time system for delivery of feedstock.

Providing a continuous flow of feedstock could be a major challenge for a biofuel production system that relies exclusively on switchgrass for feedstock. Most published estimates of switchgrass production costs assume that switchgrass would be harvested once per year when yield per hectare is maximized resulting in a narrow harvest window. A wide harvest window could reduce the fixed costs of harvest machinery per Mg of feedstock relative to a narrow harvest window and reduce storage cost. However, switchgrass harvestable dry matter yield and fertilizer requirements differ across harvest month (Haque et al. 2009; Haque 2010). The maximum expected dry matter yield of switchgrass grown in the U. S. Southern Plains is obtained by harvesting in either September or October (Table 10). In most years, harvest during April, May, and June would damage the plants and result in lower expected yields in subsequent years. The downside of an extended harvest window is that the expected yield from harvest in July is approximately 80 percent of maximum, and if switchgrass is left to stand in the field, dry matter losses of 5 percent per month are expected from November through March. Established stands of switchgrass that are harvested in July are expected to require about 90 kg/ha/yr of nitrogen to achieve the plateau yield, whereas fields harvested from October through March are expected to require approximately 67 kg/ha/yr (Haque 2010).

Conventional budgeting is necessary but insufficient to determine if the cost savings from extending harvest over nine months are sufficient to offset the losses in harvestable yield. A more comprehensive modeling approach is required to fully evaluate these tradeoffs. To address these tradeoffs, models may be constructed that encompass the entire chain of economic activities from acquisition of land use to delivery of baled switchgrass. Modeling may be conducted to determine if the additional fertilizer cost and the lower yield from an extended harvest window can be offset by the reduction in cost resulting from fewer harvest machines. The model optimally selects the number of harvest machines and the quantity of biomass to harvest by month and county during the potential July through March harvest window.