Switchgrass (Panicum virgatum L.) is undergoing development as a biomass crop to support conversion of cellulosic biomass to energy. To avoid the competition of biomass with food or feed crops, most commercialization proposals suggest that switchgrass should be grown exclusively on marginal lands that are not fit for food or feed production. The objective of this study was to investigate the potential for cultivar × environment interactions that would affect the methods and approaches for breeding and evaluating switchgrass cultivars, including both upland and lowland types, for high‐input versus low‐input types of environments. Biomass yield was measured on 14 cultivars that were present in 28 replicated field experiments representing seven regions, ranging from 75 to 100° W and spanning USDA Hardiness Zones 4 through 7. Region was the most important environmental factor interacting with cultivars, supporting the idea that the north‐central and northeastern United States should have independent switchgrass breeding programs. Cultivars interacted with soil phosphorus concentration in New Jersey and with depth of the A and B horizons in New York and showed mild interactions with rate of nitrogen fertilizer at several locations. Cultivar rank correlation coefficients between the two rates of nitrogen fertilization (100 vs. 0 kg N ha−1) ranged from 0.23 to 0.88, suggesting a possible benefit to breeding and selection without applied nitrogen fertilizer.

Development of switchgrass (Panicum virgatum L.) as a dedicated bioenergy feedstock requires intensive and extensive breeding programs that include careful and thoughtful consideration of appropriate target populations of environments (TPEs). The purpose of this study was to evaluate region (climate), soil quality, and N fertilization level as potential factors influencing the choice of TPE. A total of 45 switchgrass genotypes were evaluated in uniform field studies at six field sites defined as prime or marginal soils in New Jersey, South Dakota, and Wisconsin. Region and soil quality had strong interactions with genotype, but N fertilization had little impact on genetic variation or ranking of genotypes. Lowland genotypes were considerably more sensitive than upland genotypes to interactions with environmental factors, probably due to these field sites being outside of the traditional lowland adaptation zones. Genotype rankings were highly inconsistent across regions and soil types, indicating that breeding programs that target marginal soils should be located on soils that represent the appropriate TPE. Furthermore, interactions across the three regions suggest that breeding programs for the lowland ecotype should be subdivided into different sets of TPEs, which are largely a function of hardiness zone and annual precipitation. Lastly, even with negligible interactions involving N fertilization level, future definitions of TPEs should be based on minimal or no N fertilizer applications to allow breeders to select plants with greater N‐use efficiency, N‐scavenging ability, and N‐recycling efficiency.

No abstract