Model concepts to express genetic differences in maize yield components

Maize (Zea mays L.) grain yield is closely related to kernel number per unit area. The quantification of genetic differences among maize cultivars to kernel number plant(-1) (KNP) is critical for accurate yield simulation but remains one of the less accurate components of yield modeling. Our objective was to document the recently published KNP data and revise CERES Maize model (V3.5). The duration of a critical window for KNP simulation was 327degreesC days (227degreesC days before and 100degreesC days after silking-base temperature 8degreesC) when ears actively grew. The KNP was curvilinearly related to cumulative intercepted photosynthetically active radiation plant(-1) (CIPAR) during the critical window. Potential kernel ear(-1) and kernel produced per unit CIPAR were the genetic coefficients needed to simulate KNP. Apical ears produced maximum KNP at a plateau CIPAR of 64 MJ, and prolific hybrids produced secondary ears when CIPAR exceeded 64 MJ. The genetic differences in prolificacy in low plant density were expressed by another coefficient. Below a threshold CIPAR of 11 MJ, all plants were barren, and a barrenness coefficient expressed genetic differences among old and modern hybrids to produce KNP in high plant density. Sensitivity analysis with limited testing indicated that the revised model simulated yield reasonably well [root mean square error (RMSE) = 0.63 Mg ha(-1)] compared with the original model (RMSE = 1.25 Mg ha(-1)) across a wide range of plant densities. However, rigorous testing of the model will be required to gain greater confidence in the proposed concepts.