Simulating source-limited and sink-limited kernel set with CERES-Maize

CERES-Maize simulates kernel set as a sourcelimited process based on average plant growth rate during the lag phase after flowering. Yet the number of kernels formed by maize (Zea mays L.) also depends on timely interaction between male and female flowers, which can limit formation of reproductive sinks under some conditions. Failure to account for sink-limited kernel set may contribute to simulation error observed under conditions that affect dynamics of pollen shed or silking, but do not alter crop growth rate. We developed algorithms for a Flowering Model to simulate sink-limited kernel set from flowering dynamics. This model was calibrated against kernel production in hybrid seed production fields and then linked to CERES-Maize. The Modified CERES-Maize was calibrated against two years of field data involving three hybrids, eight population densities, and seven N levels. Integrating the capacity to simulate sinklimited kernel set with source-limited kernel set increased simulation accuracy dramatically, relative to original CERES-Maize. For 13 commercial fields tested, Modified CERES-Maize decreased simulation error for kernels per plant from 17.1 to 2.3%, improved P between measured and simulated values from 0.77 to 0.87, and decreased simulation error indicators mean error, root mean square error, and mean square deviation by 85, 40, and 64%, respectively. Modified CERES-Maize accounts for a much greater range of variability in the biological processes controlling kernel set.