Long-term yield potential of switchgrass-for-biofuel systems

• Published in: Biomass & bioenergy 2006-03, Vol.30 (3), p.198-206
• Main Authors: Fike, John H., Parrish, David J., Wolf, Dale D., Balasko, John A., Green, James T., Rasnake, Monroe, Reynolds, John H.
• Format: Article
• Language: English
• Subjects: Applied sciences; Biofuels; Biomass; Cultivar selection; Cutting frequency; Energy; Exact sciences and technology; Natural energy; Nitrogen; Panicum virgatum; Panicum virgatum L; Precipitation; System management
• ISSN: 0961-9534; 1873-2909
• DOI: 10.1016/j.biombioe.2005.10.006

Limited information is available regarding biomass production potential of long-term (>5- yr-old) switchgrass ( Panicum virgatum L.) stands. Variables of interest in biomass production systems include cultivar selection, site/environment effects, and the impacts of fertility and harvest management on productivity and stand life. We studied biomass production of two upland and two lowland cultivars under two different managements at eight sites in the upper southeastern USA during 1999–2001. (Sites had been planted in 1992 and continuously managed for biomass production.) Switchgrass plots under lower-input management received 50 kg N ha −1 yr −1 and were harvested once, at the end of the season. Plots under higher-input management received 100 kg N ha −1 (in two applications) and were harvested twice, in midsummer and at the end of the season. Management effects on yield, N removal, and stand density were evaluated. Annual biomass production across years, sites, cultivars, and managements averaged 14.2 Mg ha −1. Across years and sites, a large (28%) yield response to increased inputs was observed for upland cultivars; but the potential value of higher-input management for lowland cultivars was masked overall by large site×management interactions. Nitrogen removal was greater under the higher-input system largely due to greater N concentrations in the midsummer harvests. Management recommendations (cultivar, fertilization, and harvest frequency), ideally, should be site and cultivar dependent, given the variable responses reported here.