Parameterization Of The CERES Grain Sorghum Model For Simulating Sweet Sorghum Growth and Dry Matter Yield.

Sweet sorghum (Sorghum bicolor) has recently garnered attention in the southeastern USA as a bioenergy crop. Sweet sorghum can yield as much ethanol per acre in the region as corn does in the Midwest with lower inputs.  Few studies have focused on sweet sorghum modeling to date. The purpose of this study was to parameterize the CERES grain sorghum model for simulating sweet sorghum growth and dry matter yield. We collected growth sampling data under optimal conditions over two consecutive years (2012 and 2013) for the cultivar ‘M-81E’, and used end of season yield data from a planting date study conducted at three locations across Florida for the same genotype for model testing. Initially, the grain sorghum model overestimated grain production and underestimated total biomass and stem yield for sweet sorghum. Based on comparative studies found in the literature, we increased the radiation-use efficiency, reduced the amount of partitioning allocated to roots, reduced the grain filling rate, and reduced the amount of assimilate retranslocated from the stem to the grain head to 0. After these changes the same data were simulated using the new model, the CERES grain sorghum model and the CERES maize model. The models were then compared using the mean square error of prediction criteria. The new model produced a better prediction of total biomass, stem weight and grain weight. The current model can be used to simulate growth and yield of sweet sorghum in different regions with different planting dates. Further studies are necessary to provide accurate yield predictions under suboptimal conditions and to estimate soluble stem sugars.