Model Predictive Control of Adaptive Irrigation Decisions Incorporating Rainfall Intensity and Soil Properties

To mitigate the loss of surface runoff and deep percolation in the water-scarce area and enhance the utilization of rainfall resources, this study adaptively determines the soil water content threshold triggering such losses by incorporating rainfall intensity (RI) and soil properties (SP) based on the model predictive control (MPC) framework. These thresholds then serve as the target soil water content before rainfall, and a model predictive control incorporating RI and SP (RISPMPC) irrigation decision-making is proposed. We conducted irrigation simulation experiments in Ya’an City, Sichuan Province, across four RI levels and six soil texture types. The results were compared with those obtained from conventional model predictive control (CMPC), rule-based closed-loop irrigation decision (RBC), and a newly developed zone-based model predictive control (ZMPC). Results demonstrate that RISPMPC enhances the utilization of rainfall resources across different scenarios. In soils with strong infiltration capabilities, such as loamy sand, loam, and clay loam, RISPMPC reduces irrigation water use by 26%, 5%, and 3% compared to RBC, CMPC, and ZMPC, respectively. In contrast, for soils with poor infiltration capabilities, including silty soil, clay A, and clay B, RISPMPC’s water-saving efficiency strongly correlates with rainfall intensity levels, achieving maximum savings of 61%, 36%, and 34% compared to the same three methods. Furthermore, in all cases, RISPMPC demonstrates the highest maximum effective rainfall utilization rate (MERU). As soil infiltration capability decreases and rainfall intensity increases, the MERU gap between RISPMPC and the other three methods widens significantly, underscoring RISPMPC’s robustness in environments where rainwater utilization is challenging. Therefore, RISPMPC can improve the utilization efficiency of rainwater resources and effectively alleviate agricultural water scarcity issues.