Developing a Sweet Corn Simulation Model to Predict Fresh Market Yield and Quality of Ears

Abstract

Current production of sweet corn (Zea mays L.) in the United States is 4.0 million Mg with a value of $807 million. The fresh market component amounts to three-fourths of this value with California, Florida, and Georgia harvesting half of the U.S. fresh market production. Existing maize simulation models have limited potential to assist sweet corn production as a result of the distinctive nature of the marketed end product (i.e., fresh market ears versus dry mature kernels). The purpose of this study was to develop a sweet corn simulation model. The Cropping System Model-Crop-Environment Resource Synthesis CSM-CERES-Maize simulation model, version 4.0, was modified to improve the simulation of ear growth, to predict ear fresh market yield, and to predict fresh market ear quality according to U.S. standards. A field experiment conducted in Florida in 2003 was used for model development. Five nitrogen fertilization levels (0, 67, 133, 200, and 267 kg·ha−1 N) were applied to a sh2-based commercial hybrid with a Bt gene sown at 8.2 plants/m2. Three additional experiments conducted in 2002, 2004, and 2005 provided independent data to evaluate the new model. In 2002, the treatments and hybrid were the same as mentioned, but the population density was 5.5 plants/m2. A yellow sh2-based hybrid with a Bt gene was planted at 6.1 plants/m2 in 2004. In 2005, a bicolor sh2-based hybrid with a Bt gene was planted at 8.1 plants/m2. The 2004 and 2005 experiments had 100% and 150% of the Florida N recommendations applied to the crop. Results indicated that the new model was able to simulate adequately crop and ear growth of sweet corn. The ear dry weight simulation was improved as indicated by 30% reduction of root mean square of the error (RMSE) when the new model was compared with the original CSM-CERES-Maize. Total ear fresh weight yield and marketable yield were also simulated reasonably well with RMSE values of 3367 and 3502 kg·ha−1, respectively. The simulation of ear quality was consistently overpredicted at intermediate levels of N fertilization, indicating the need to further examine the impact of limited N on ear quality.