Adapting the APEX Model to Simulate Detasseling in Inbred Corn for Hybrid Seed Production

Highlights The APEX model was adapted to simulate detasseling in inbred corn for hybrid seed production. The adapted model satisfactorily predicted detasseling effects on LAI, grain yield, and N content. An inbred corn model could be applied to evaluate best management practices for inbred corns. Abstract. Hybrid seed corn production comprises approximately 10% of the entire corn acreage in the U.S. Because of seed corn‘s high economic value, and to maximize yields, seed corn growers often over-irrigate or apply nitrogen (N) fertilizers equal to or in excess of those recommended for commercial hybrid corn. Detasseling female corn inbred lines during hybrid corn seed production is critical to ensure the purity of seeds. In addition to the removal of tassels, detasseling also results in the removal of several leaves, which may lead to reduced seed yields. The objective of this study was to adapt the Agricultural Policy/Environmental eXtender (APEX) model to simulate the detasseling of female inbred corns in hybrid seed production. An APEX inbred corn model was developed to simulate the effects of detasseling and leaf removals on the development of inbred corn, leaf area index (LAI), grain yield, and grain N content. Growth characteristics of inbred corn were parameterized in APEX using data from a field study conducted in Nebraska. Overall, the APEX inbred corn model successfully predicted the effects of detasseling on LAI, grain yield, and grain N content under the conditions of the field experiment. There was a significant correlation between simulated and measured LAI (Pearson r = 0.86 and R2 = 0.74 at p ≤ 0.05). The computed paired t-test and permutation test p-values indicated no significant differences between measured and simulated LAI. The mean simulation percent difference and percent bias (PBIAS) were respectively 4.2% and 4.7%, while measured and simulated LAI values had an average root mean square error (RMSE) of 0.14. The APEX model predicted grain yield with RMSE of 120 kg ha-1, mean simulation percent difference of 0.48%, and PBIAS of 0.26%. Like LAI, predicted grain yields exhibited significant correlation with field data (Pearson r = 0.99 and R2 = 0.97 at p ≤ 0.05). Similarly, computed paired t-test and permutation test p-values indicated no significant differences between measured and simulated grain yields. Grain N content was predicted with RMSE of 6.75 kg N ha-1, mean simulation percent difference of 1.46%, and PBIAS of 2.45%. Predicted and measured grain N content values were correlated (Pearson r = 0.81 and R2 = 0.65 at p ≤ 0.05), while the t-test and permutation test p-values indicated no significant differences between measured and predicted grain N content. Overall, detasseling effects were better predicted for grain yield than for LAI and grain N content as indicated by a Nash-Sutcliffe efficiency (NSE) of 0.92 compared to NSE values of 0.47 for LAI and 0.43 for grain N content. In conclusion, the hybrid seed corn industry could benefit from the application of inbred corn models that could allow growers to evaluate and identify optimal irrigation and N management practices for inbred corn, similar to the benefits that have been obtained with model simulation for commercial hybrid corn grain production systems.