Establishment and Initial Productivity of Native Perennial Grassland Species for Bio-energy.

Native perennial herbaceous grassland species have been identified as a potential feedstock for energy production. Prairie plants are ideal candidates because they are adapted to low nutrient environments, generate significant biomass, and provide a plethora of ecological services. Our objectives were to 1) evaluate the establishment and productivity of four native warm and cool season grasses planted in monocultures and mixtures at four locations and 2) document the botanical composition of mixtures and monocultures subject to annual harvesting. Native grasses, legumes, and non-leguminous forbs were cultivated alone and in mixtures including 1, 4, 8, 12, and 24 species. The plots were established in Minnesota during June 2006 and harvested in October through November 2007. There was successful establishment of 22 of the 24 species seeded in the highest diversity mixture. The overall botanical composition was similar for all locations. However, plant density and relative abundance varied by location. The dominance, as defined by greatest percent ground cover, of native grasses varied between locations during 2007, with Panicum virgatum (Lamberton, 98% cover), Elymus canadensis (Waseca and St. Paul, 86% and 88% cover, respectively) and Andropogon gerardii (Becker, 88% cover) all accumulating considerable biomass.  Annual aboveground yield differed among mixtures but showed a similar pattern at all locations, with legume polycultures always lowest and grasses highest.  Grass monocultures yielded the most aboveground biomass at the end of the 2007 growing season, with the highest yielding grass monoculture 22% greater than the closest yielding mixture. At Lamberton, monocultures of Panicum virgatum (848.46 g/m²) and Elymus canadensis (876.34 g/m²) yielded more total aboveground biomass than any other location. The conclusion from the first two years of research is that aboveground biomass yield and relative dominance of species is influenced greatly by site.