One potentially significant impact of growing biofuel crops will be the sequestration or release of carbon (C) in soil. Soil organic carbon (SOC) represents an important C sink in the lifecycle C balances of biofuels and strongly influences soil quality. We assembled and analyzed published estimates of SOC change following conversion of natural or agricultural land to biofuel crops of corn with residue harvest, sugarcane, Miscanthus x giganteus, switchgrass, or restored prairie. We estimated SOC losses associated with land conversion and rates of change in SOC over time by regressing net change in SOC relative to a control against age since establishment year. Conversion of uncultivated land to biofuel agriculture resulted in significant SOC losses -an effect that was most pronounced when native land was converted to sugarcane agriculture. Corn residue harvest (at 25-100% removal) consistently resulted in SOC losses averaging 3-8 Mg ha À1in the top 30 cm, whereas SOC accumulated under all four perennial grasses, with SOC accumulation rates averaging o1 Mg ha À1 yr À1 in the top 30 cm. More intensive harvests led to decreased C gains or increased C losses -an effect that was particularly clear for residue harvest in corn. Direct or indirect conversion of previously uncultivated land for biofuel agriculture will result in SOC losses that counteract the benefits of fossil fuel displacement. Additionally, SOC losses under corn residue harvest imply that its potential to offset C emissions may be overestimated, whereas SOC sequestration under perennial grasses represents an additional benefit that has rarely been accounted for in life cycle analyses of biofuels.

Current biofuel feedstock crops such as corn lead to large environmental losses of N through nitrate leaching and NO emissions; second-generation cellulosic crops have the potential to reduce these N losses. We measured N losses and cycling in establishing miscanthus (), switchgrass ( L. fertilized with 56 kg N ha yr), and mixed prairie, along with a corn ( L.)-corn-soybean [ (L.) Merr.] rotation (corn fertilized at 168-202 kg N ha). Nitrous oxide emissions, soil N mineralization, mid-profile nitrate leaching, and tile flow and nitrate concentrations were measured. Perennial crops quickly reduced nitrate leaching at a 50-cm soil depth as well as concentrations and loads from the tile systems (year 1 tile nitrate concentrations of 10-15 mg N L declined significantly by year 4 in all perennial crops to <0.6 mg N L, with losses of <0.8 kg N ha yr). Nitrous oxide emissions were 2.2 to 7.7 kg N ha yr in the corn-corn-soybean rotation but were <1.0 kg N ha yr by year 4 in the perennial crops. Overall N balances (atmospheric deposition + fertilization + soybean N fixation - harvest, leaching losses, and NO emissions) were positive for corn and soybean (22 kg N ha yr) as well as switchgrass (9.7 kg N ha yr) but were -18 and -29 kg N ha yr for prairie and miscanthus, respectively. Our results demonstrate rapid tightening of the N cycle as perennial biofuel crops established on a rich Mollisol soil.

a b s t r a c tPerennial grass biofuels may contribute to long-term carbon sequestration in soils, thereby providing a broad range of environmental benefits. To quantify those benefits, the carbon balance was investigated over three perennial grass biofuel crops -miscanthus (Miscanthus × giganteus), switchgrass (Panicum virgatum) and a mixture of native prairie plants -and a row crop control (maize-maize-soy) in Central Illinois, USA, during the establishment phase of the perennial grasses (2008)(2009)(2010)(2011). The eddy covariance technique was used to calculate fluxes of carbon dioxide and energy balance components, such as latent and sensible heat fluxes. Whereas maize attained the highest maximal carbon uptake rates, the perennial grasses had significantly extended growing seasons, such that their total carbon uptake rivaled that of corn in the second growing season and greatly exceeded that of soy in the third growing season. To account for the removal of carbon through harvest, net ecosystem exchange of carbon (NEE) was combined with estimates of yields, resulting in the net ecosystem carbon balance (NECB). After 2.5 years, NECB for the maize/soybean plot was positive (a source of carbon), while the grasses were a sink of carbon. Continuous measurements over the next years are required in order to confirm whether miscanthus, switchgrass and prairie can sustain a long-term sink of carbon if managed for biofuels, i.e., if harvested annually.