Optimization of irrigation scheduling for spring wheat based on simulation-optimization model under uncertainty

Water scarcity is the major constraint to social-economic development in arid and semiarid regions, where irrigation needs to be scheduled properly for the main crops. In this study, a simulation-optimization model for crop optimal irrigation scheduling under uncertainty was developed to maximize the net benefit. The model integrated a water-driven crop model (AquaCrop) with the optimization model, and incorporated the generation technique for the interval values of hydrological parameters (i.e., precipitation and evapotranspiration) and crop market prices to deal with uncertainties in these variables. The water price was assumed constant. The model was calibrated based on field experimental data obtained in 2014 and validated using 2015 data. The field experiments involved spring wheat (Yongliang No. 4) at Shiyang River Basin Experiment Station in Wuwei City, Gansu Province of Northwest China. The model was then used to generate the optimal irrigation schedules under various irrigation amounts, irrigation events, initial soil water storage and crop market price under uncertainty. Results indicated that the model is applicable for reflecting the complexities of simulation-optimization under uncertainties for spring wheat irrigation water scheduling. The optimization results indicated that the optimal irrigation amount range was [185, 322] mm with the corresponding optimal net benefit of [1.05, 2.77] ×104 Yuan/hm2 and yield of [7.4, 7.6] kg/hm2 for extreme wet conditions in the basin (defined as the combination of the 5% frequency precipitation with 95% frequency evapotranspiration). For extreme dry conditions, the optimal irrigation amount range was [442, 507] mm with the optimal net benefit of [0.85, 2.64] ×104 Yuan/hm2 and the corresponding yield of [6.6, 7.4] kg/hm2. Results also showed that four irrigation events under higher initial soil water storage were more likely to produce the higher net benefit and the optimal net benefit would increase as the crop market price increases, as expected. This work can be used to guide irrigation management for local farmers.