Abstract: Traditional irrigation decision-making methods based on the AquaCrop model have limited capability for dynamic response, making it difficult to promptly adjust irrigation schedules according to short-term meteorological variations. An irrigation decision-making method integrating the AquaCrop model with model predictive control (MPC) was developed. The AquaCrop model was used to determine irrigation upper limits for different growth stages of summer maize. Soil water balance, surface runoff, deep percolation, and short-term rainfall forecast information were incorporated into the MPC decision-making framework. An AquaCrop-MPC dynamic irrigation decision-making model was constructed, and irrigation amounts were determined through rolling-horizon optimization. A simulation experiment was conducted for summer maize grown in silt loam soil in Shangqiu City, Henan Province. Under total water supply levels of 100, 130, and 160 mm, the AquaCrop-MPC method increased simulated yield by 0.4%, 0.3%, and 0.1%, respectively, reduced irrigation water use by 36.0%, 42.3%, and 40.0%, respectively, increased irrigation water productivity by 56.9%, 74.0%, and 66.9%, respectively, and improved water use efficiency by 8.3%, 9.1%, and 8.3%, respectively, compared with the preset irrigation schedules determined by AquaCrop. The proposed method dynamically adjusted irrigation amounts using short-term rainfall forecast information and reduced irrigation water use while maintaining simulated crop yield. These findings provided a scientific reference for irrigation decision-making of summer maize under limited water supply conditions.

Key words: AquaCrop model, model predictive control, smart irrigation, irrigation decision-making, summer maize