Optimization of water and fertilizer coupling system based on rice grain quality

Nowadays attentions have been increasingly paid to the quality of grain in the production of rice. A framework for understanding the effects of water and fertilizer on rice yield and grain quality was proposed through integrating techniques of experiment, simulation and optimization. Firstly, experiments with varied combinations of water and fertilizer were conducted, during which ten quality traits of rice grain were measured. Based on experimental results, the relationships of water nitrogen-yield and water nitrogen-quality were simulated. The principal components analysis method was adopted to identify the main traits, based on which a comprehensive quality model of rice grain was established. On this basis, a multi-objective quadratic model for supporting the use of both water and fertilizer was developed and solved by fuzzy goal programming under eight scenarios. The framework combined principal component analysis with water nitrogen-yield and water nitrogen-quality models to establish a comprehensive quality model, so that the amounts of both water and fertilizer could be optimized simultaneously. Different quality traits showed different responses to the changes in water consumption and nitrogen application rates, and quadratic functions could well describe the relationships of water nitrogen-yield and water nitrogen-grain quality under the tested scenarios. Under different scenarios, optimal allocation amounts of water and fertilizer were different. The results showed that, irrigation amount and fertilizer application rates could be adjusted within certain limits for improving crop yields and grain quality. The optimal nitrogen application rate ranged from 80 to 140 kg/ha, and the water consumption rate ranged from 5000 to 8000 m(3)/ha. Larger amounts of water and fertilizer inputs could be counter-productive.