Intelligent irrigation management system for arid and semi-arid regions under climate change.

In arid and semi-arid regions, optimal irrigation management requires a detailed and integrated understanding of the interactions among climate, soil, and crop physiology on a daily time scale. In this study, an intelligent system for determining irrigation timing and depth was developed by integrating actual and predicted climate data with a physical water–soil–plant model, and its efficiency was evaluated for three crops, wheat, tomato, and apple, under water-limited conditions in the Miandoab Plain (Lake Urmia Basin). The system accurately estimated the water requirements of crops by simulating soil moisture at different layers, calculating crop evapotranspiration (ETc), and dynamically updating soil evaporation coefficient (Ke) and crop transpiration coefficient (Kcb) based on leaf area index (LAI) and phenological stages. The results showed that the temporal patterns of temperature and ETc changes were fully consistent with the semi-arid climate, and an increase in temperature during the middle of the growing season led to a significant rise in ETc. The highest ETc values were observed in the apple orchard (up to approximately 29 mm/day), while tomato and wheat exhibited peak-decline and flowering-to-grain-filling patterns, respectively. The results showed that the intelligent system provided estimates comparable to those of the Penman–Monteith method, with no significant difference (p > 0.05). Analysis of Kcb and Ke coefficients indicated that, as leaf area expanded, the contribution of soil surface evaporation decreased, and crop transpiration became the dominant component of water consumption. Field validation demonstrated that the intelligent system predicted water needs more accurately than conventional methods, resulting in up to 41% reduction in water use, up to 12.6% increase in yield, and up to 87% improvement in water productivity, depending on the crop and irrigation system. Correlation analysis confirmed the strong relationship between temperature, LAI, and ETc, demonstrating the system’s capability to represent climatic and physiological factors simultaneously. Overall, the study shows that intelligent systems based on physical–physiological models can significantly enhance the accuracy of water demand forecasting, improve productivity, and promote sustainable irrigation management at both farm and basin scales, serving as a generalizable tool for climate-change-adapted agriculture in arid and semi-arid regions.