Rice irrigation water efficiency improvement: An AquaCrop-based optimization modeling approach

Rice is the cereal crop that consumes the largest amount of water in the world. To improve the water use efficiency of rice farming and promote the sustainable development of rice production, this study proposes a novel AquaCrop-based optimization modeling approach. First, an AquaCrop-based multiobjective model is constructed to generate dynamic water production functions (WPFs). The NSGA-II algorithm is employed to solve the model, and quadratic dynamic WPFs for rice under different hydrological years (wet, normal and dry) are generated based on the obtained Pareto fronts, which can improve the modeling efficiency and help reflect the complicated reality of water flows. Second, a fuzzy credibility-constrained stochastic multiobjective programming (FCCSMOP) model is developed for irrigation water allocation based on the acquired WPFs. The model addresses the randomness of precipitation and runoff, the fuzziness of decision-makers' preferences, and conflicting objectives of water savings, benefit increments, and water demand during the critical period of rice growth. The proposed simulation-optimization approach was applied to the Changgang irrigation district in northeast China. Optimal irrigation amounts under different decision-making preference scenarios (including conservative, moderate and positive) and different hydrological years were simultaneously generated. The following conclusions can be drawn: (1) The AquaCrop-based multiobjective method is capable of generating WPFs of different hydrological years with sparse discontinuous field experiments; (2) the decision-making preferences of decision-makers significantly impact the economic benefits, water use efficiency, and practicality of schemes; and (3) the FCCSMOP model successfully makes tradeoffs among the three objectives under multiple uncertainties and can further improve the system efficiency and increase planting income in the Changgang irrigation district when its irrigation facilities become optimal.

Automated summary

This study proposes an AquaCrop-based optimization modeling approach to improve the water use efficiency of rice farming. A multiobjective model is constructed to generate dynamic water production functions (WPFs) for rice under different hydrological years. Based on the obtained WPFs, a fuzzy credibility-constrained stochastic multiobjective programming (FCCSMOP) model is developed for irrigation water allocation. The proposed approach is applied to the Changgang irrigation district in northeast China and shows promising results in improving system efficiency and increasing planting income.