Using the CERES-Maize model in a semi-arid Mediterranean environment. Evaluation of model performance

The CERES-Maize model was tested in a semi-arid Mediterranean environment during a period of 2 years under three different soil moisture conditions (well-watered and two limited irrigation regimes). In well-watered plots, growth and yield were adequately simulated by the model (differences between simulated values and observations were less than 10%). Results suggest that the absence of air humidity among the model inputs does not limit the CERES-Maize performance, even under dry-air conditions. On the contrary, under mild soil water shortage, CERES-Maize underestimated the leaf area index (LAI) (up to 26% for maximum LAI), above-ground biomass (up to 23%) and grain yield (up to 15%). Mismatches between observations and predictions increased with water stress level (by up to 46, 29 and 23% for maximum LAI, biomass and grain yield, respectively). It is suggested that the functions describing leaf growth and senescence and those calculating the soil water deficit functions should be modified to adapt CERES-Maize to Mediterranean environments.