SIMULATION OF SOIL WATER CONTENT FOR IRRIGATION MANAGEMENT BASED ON ON-FIELD AND CERES-WHEAT SIMULATED DATA IN WINTER WHEAT IN THE NORTH CHINA PLAIN

Accurate forecasting of the soil water in crop growth is crucial for developing an irrigation management strategy and saving water resources. The CERES model can be used to quantitatively predict the soil water content (SWC), but its accuracy and stability should be improved. The objectives of this research were to assess the CERES-Wheat model for its ability to simulate the SWC in the winter wheat growth season and to study the possibility of improving simulation accuracy by building a modified model through introducing on-field data and simulated values of CERES-Wheat. Field experiment data of the varieties ShiJiaZhuang 8 and JiMai 22 from three to four irrigation treatments conducted over 4 years were used to compare precision of the CERES-Wheat model and the modified model for SWC in the North China Plain. On average across all soil layers between 20 and 120 cm and two winter wheat varieties, the NRMSE decreased from 20.3% with the CERES-Wheat model to 13.2% with the modified model. The results showed that the further improvement of SWC based on on-field data and simulated data of CERES-Wheat may be possible. Application prospects of dynamic water management of the modified model by introducing on-field data during the wheat growth season were also discussed. (c) 2019 John Wiley & Sons, Ltd.