We evaluated the biogeochemical cycling and relative greenhouse gas (GHG) mitigation potential of proposed biofuel feedstock crops by modeling growth dynamics of Miscanthus × giganteus Greef et Deuter (miscanthus), Panicum virgatum L. (switchgrass), Zea mays L. (corn), and a mixed prairie community under identical field conditions. DAYCENT model simulations for miscanthus were parameterized with data from trial plots in Europe and Illinois, USA. Switchgrass, corn, and prairie ecosystems were simulated using parameters published in the literature. A previously unknown source of nitrogen (N) was necessary to balance the plant nutrient budget in miscanthus crops, leading us to hypothesize that miscanthus growth depends on N-fixation. We tested for nitrogenase activity by acetylene reduction of whole rhizomes and bacteria isolated from the rhizosphere and miscanthus tissue. Our results supported the hypothesis that biological N-fixation contributed to the N demand of miscanthus, a highly productive perennial grass. Corn agro-ecosystems emit 956 to 1899 g CO₂eq m⁻²y⁻¹ greater GHGs (including CO₂, N₂O, CH₄) to the atmosphere than the other biofuel crop alternatives because of greater N₂O emissions from fertilizer additions. Of the feedstock crops evaluated in this study, miscanthus would result in the greatest GHG reduction.