Evaluation of ceres-maize model for simulation of maize under different scenarios of irrigation and nitrogen fertilizer management

Crop modeling is a cheap, fast and powerful method to achieve the results of the effect of various factors on crop growth. Hence, crop models such as CERES-Maize have been developed to simulate plant performance. Given that the amounts of irrigation water and nitrogen fertilizer are two very important factors to improve corn yield; it is important to know the accuracy and error of the CERES-Maize model to simulate the yield of this crop under the mentioned treatments. Therefore, the present study was conducted at a 500-hectare farm of the Seed and Plant Breeding Research Institute located at 50.58° East longitude and 35.56° latitude on two corn cultivars (double cross 370 and single cross 260). For double cross 370, two factors including the amount of irrigation water at four levels (W1: 120, W2: 100, WI3: 80 and W4: 60 percent of water requirement) and nitrogen fertilizer at four levels (N1: 100, N2: 80, N3: 60 and N4: zero percent of nitrogen requirement) were considered. For single cross 260, four fertilizer levels (N1: 100, N2: 80 and N3: 60 and N4: 50 percent of nitrogen requirement) were studied. The results for both cultivars showed that the CERES-Maize model underestimated crop yield (0 ≥ MBE). The amount of error for simulating yield of double cross cultivar 370 and single cross cultivar 260 was 1.24 and 0.44 tons per hectare, respectively. The accuracy of CERES-Maize model for simulating these two cultivars was in the category of good (NRMSE = 0.13) and excellent (NRMSE = 0.06), respectively. The error of CERES-Maize model for double cross cultivar 370 and for irrigation treatments was in the range of 0.89-1.65 t/ha and for fertilizer treatments was in the range of 0.43-9.9 t/ha. No difference was observed between the model errors in two fertilizer applications. Therefore, the model was not sensitive to fertilizer apportionment, while changes in irrigation water and fertilizer amount had a great effect on its accuracy. Based on all the results, the CERES-Maize model is recommended for simulation of both corn cultivars, although its accuracy was higher for the single cross 260 cultivar.