Optimization of multi-objective irrigation schedule for rice based on AquaCrop model

【Objective】 In order to optimize the irrigation schedule for rice, conserve irrigation water, reduce nitrogen and phosphorus loss in paddy fields, and minimize the risk of non-point source pollution. 【Method】 A simulation and optimization model for rice irrigation system based on AquaCrop model and NSGA-II algorithm was developed. Field experimental data of rice were used to compare and study the optimization of irrigation systems with three objective combinations: maximizing yield, minimizing nitrogen and phosphorus loss, and minimizing irrigation frequency (Yield-Total Nitrogen and Phosphorus loss (Y-TNP), Yield-Irrigation frequency (Y-N), and Yield-Total Nitrogen and Phosphorus loss-Irrigation frequency (Y-TNP-N)). Yield stable-pollution reducing-efficiency improving irrigation systems that adapt to different precipitation patterns during different growth stages of rice were proposed. 【Result】 ①Compared with conventional irrigation, Y-TNP optimization resulted in a 2.14% decrease in yield and a 23.09% reduction in nitrogen and phosphorus loss; Y-N optimization could achieve a 1.76% decrease in yield and an approximately 53% decrease in irrigation frequency; Y-TNP-N optimization performed a 2.64% decrease in yield, a 22.83% reduction in nitrogen and phosphorus loss and a decrease of 2 in irrigation frequency. ②The yield stable-pollution reducing-efficiency improving irrigation system optimized with the Y-TNP-N objective achieved a stable yield of 7.74-7.78 t/hm2 under different typical years, significantly reducing irrigation amount and nitrogen and phosphorus loss. 【Conclusion】 The AquaCrop model can be applied to simulate the growth and development process of rice in the experimental area. The simulation-optimization coupled model proposed in this study can be used to optimize irrigation systems with the goal of stable yield-reducing pollution-improving efficiency under different precipitation patterns during different growth periods.