Dynamic Adjustment Mechanism of Irrigation Quotas for Winter Wheat–Summer Maize Based on Transformer-Driven Crop Water Requirement Forecasting

Amid intensifying climate change and growing water scarcity, improving irrigation efficiency is essential for food security and sustainable water management. As a major diversion irrigation district in Shaanxi province, the Jiaokou irrigation district plays a crucial role in regional grain production. However, irrigation quotas based on traditional experience are poorly suited to climate variability and crop water requirements, leading to inefficient water use. To address this issue, a dynamic irrigation quota adjustment framework was developed based on daily meteorological data for winter wheat and summer maize from 1980 to 2023. The framework integrates a Transformer-based time-series prediction model, crop water demand frequency analysis, multi-objective optimization, and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method. Irrigation requirements for a typical year were obtained through frequency analysis to construct a quota decision table. Before irrigation, meteorological forecasts and water availability are used to select preliminary quotas, and post-irrigation adjustments rely on Transformer-based predictions and optimization results. The approach captures variations in crop water demand under different climatic conditions and achieves coordinated optimization of irrigation water use and yield. The findings provide a scientific basis for rational irrigation quotas and water pricing policies in the Jiaokou irrigation district.