Adaptation of CERES-Maize model for a better simulation of plant population density effects

Abstract The Crop Environment Resource Synthesis (CERES) models have been developed and utilized for the last 30 years at the field scale to optimize current management, estimate crop yield potential, and study input-use efficiency. Despite a large number of simulated variables, little work has been devoted to modeling the effect of plant population density on the behavior of individual plants. Our objective was to construct a modified version (CL version) of the CERES-Maize model to improve the simulation of yield in relation to plant density in the conditions of Belgium and Chile, focusing on its main component, yield/plant. The changes in the CL version are to be found in the algorithm, in the structure of the model, as well as in the methods of estimating the input parameters G2 and G3. Values, of the relative rate of improvement (RRI) in function of the root mean square error (RMSE) indicate an important improvement of the simulation of yield grain (37.4%), number of grains (22.8%) and harvest index (18.6%). The CL version reproduces more adequately the shape of the response of yield per plant to population density than the original version.