The CERES-Maize model was tested in a Mediterranean environment in Iran for simulating the response of three maize hybrids (DC370, ZP677, and SC704) to water supply (full irrigation, mid drought stress, and severe drought stress) and nitrogen rates (0, 120, 180, and 240 kg N ha -1 ). Grain yields of maize hybrids were adequately simulated by the model (differences between simulated values and observations were less than 10%). However, CERES-Maize underestimated grain yield in most of

more »

... Mismatches between observations and predictions increased with water stress. CERES-Maize predicted grain yield of hybrids well (normalized RMSE = 10.7, 7.0, and 9.0%, respectively). CERES-Maize model simulated the variation of biological yield of maize hybrids equal to the average measured biological yield, although the model almost underestimated biological yields in response to different levels of water and nitrogen supply. Validation trials gave an nRMSE for DC370, ZP677, and SC704 of 8.8, 12.2 and 14.3%, respectively. The field experiment and model output results also revealed that decrease of irrigation application increased water use efficiency of maize hybrids due to higher grain yield per unit of water supplied. Also, increase of nitrogen fertilizer gradually increased WUE of all maize hybrids. For the model validation, the nRMSE for WUE of three maize hybrids was 7.91, 6.12, and 9.14%, respectively indicating that the CERES-maize model predict WUE of studied hybrids reasonably well. Obtained results suggested that the CERES-Maize model can be utilized to evaluate the effects of N and irrigation management on maize under same climatic conditions.