Predicting water and nitrogen requirements for maize under semi-arid conditions using the CSM-CERES-Maize model

Crop models can be useful tools for optimizing irrigation water and fertilizer management to improve crop productivity. The goal of this study was to assess the performance of the Cropping System Model (CSM)-CERES-Maize for its capability to simulate soil moisture content in relation to plant growth, development and grain yield and to determine optimum irrigation and fertilizer inputs under semiarid conditions. The model simulations were compared with the observed results from field trials that were conducted during 2009 and 2010 with a combination of three irrigation regimes (full irrigation, water deficit at vegetative and at reproductive stage) and five nitrogen (N) rates using a split plot design for a total of 15 treatments. To determine the most appropriate combination of nitrogen fertilizer and irrigation, a combination of three irrigation regimes and five N rates ranging from 100 to 300 kg N ha⁻¹ for a total of 15 scenarios were simulated for 35 years of historical daily weather data. The model was calibrated with an optimum treatment from the 2010 experiment, while the remainder of the data were used for model evaluation. The results showed that the model successfully predicted (R² = 0.98) soil moisture content throughout the growing season. The observed (calibrated) mean percentage differences (MPD) for the numbers of grains per ear, leaf area index (LAI) and total dry matter (TDM) were 5.98, 11.4, and 4.85%, respectively. The MPD was zero for yield, anthesis and maturity days. The normalized root mean square error (nRMSE) for grain yield was 10.4% and 11.4% in 2009 and 2010, respectively. Based on the economic analysis, the management scenario with an N fertilizer application rate of 300 kg N ha⁻¹ in three splits and total irrigation of 525 mm was dominant with the highest mean grain yield (7973 kg ha⁻¹) and a gross margin of US $ 548 ha⁻¹. The outcomes of this study can be used for determining the optimum water and N requirements for maize production under semi-arid conditions. The modeled genetic coefficients might be helpful for plant breeders to develop maize cultivars for semi-arid regions that may give the optimum yield under above recommended N and water management practices.