This investigation used a modified parameterization of the Agricultural Land Management Alternative with Numerical Assessment Criteria (ALMANAC) model to improve simulated growth and biomass yield of upland switchgrass (Panicum virgatum L.) ecotypes in northern U.S. locations. Leaf area development, biomass accumulation, and N utilization of upland ecotypes were parameterized by field evaluations from Montreal, QC, Canada, and sites throughout the northern U.S. Great Plains. Resulting ALMANAC simulations were validated against measured yields from 66 location–years of switchgrass production across 13 sites in Minnesota, North Dakota, and South Dakota. As contrasted to the model defaults, the modified parameterization reduced RMSE of annual simulated yields from 3.77 to 2.62 Mg ha−1 and improved percentage bias from −16 to 13%. Model performance was most improved in environments with no N fertilization, where ALMANAC simulated annual yields with an RMSE of 1.73 Mg ha−1 and percentage bias of −0.5%. Relative to the default ALMANAC parameterization, the modified parameterization also simulated a longer growing season and extended the median simulated maturity date from 1 to 28 August, greatly improving the estimation of switchgrass phenology within the study region. Sensitivity analyses revealed that simulated switchgrass yield was unaffected by modifications of runoff curve number and was most affected by modifications of radiation use efficiency. The other seven parameter modifications each had a median yield impact of 0.57–1.4 Mg ha−1. This work provides an improved characterization of upland switchgrass ecotypes in northern U.S. locations for future ALMANAC users.
Bibliographical notePublisher Copyright:
© 2021 The Authors. Agronomy Journal © 2021 American Society of Agronomy.