Crop residues are receiving attention as a potential feedstock for lignocellulosic biofuels, but sustainable residue harvest may be limited by the need to control soil erosion and maintain soil organic carbon (SOC). Little attention has been given to the potential for residue harvest in the semi-arid Great Plains, largely due to assumptions of low production levels and the strong erosive forces of wind. Due to expanding interest in growing dedicated biofuel crops on marginal lands, this study examines the capabilities and constraints of harvesting agricultural residues from semi-arid dryland production systems. Long-term production levels of grain and stover for wheat, corn, and grain sorghum were examined at three no-till dryland cropping sites and a modeling approach was used to evaluate the impacts of stover removal on wind and water erosion, soil organic carbon, nitrogen dynamics, and grain yields. Under current wheat-corn-fallow management, virtually all stover must remain in order to control wind erosion and maintain soil organic carbon. However, if dedicated non-grain bioenergy crops were grown on an annual basis, there could be 2500-2700 kg ha-1 of harvestable biomass yearly while still retaining enough residues to maintain SOC. Total biomass production of a dedicated non-grain energy crop could be higher than the biomass production of the grain crops examined, namely because energy is not diverted to grain, and non-grain crops are not as sensitive to the timing of water deficits.