基于数据融合的高时空分辨率作物蒸散发反演与高效精细化灌溉决策

Data Fusion-based High Spatiotemporal Resolution Retrieval of Crop Evapotranspiration and Efficient Refined Irrigation Decision-making for Irrigation Areas

摘要:

高效利用水资源和精准灌溉管理是提升农业生产效率的关键，而蒸散发作为灌溉水管理中的核心参数，其传统估算方法因时空分辨率低等问题，制约了精细化灌溉管理的实施。提出了一种集成遥感技术、数据融合和多目标优化的高时空分辨率作物蒸散发反演与精细化灌溉管理方法。通过时空融合模型获取高精度地表变量（NDVI、反照率和地表温度），实现田间尺度（30m×30m）逐日作物蒸散发（ET）精准反演，并构建基于非支配排序遗传算法（NSGA-Ⅱ）的灌溉制度优化模型，为灌区差异化灌溉提供决策支持。结果表明，该方法充分考虑田间时空变化，水分供给精准契合作物需水特性。通过优化灌溉策略，在非关键生育期灌水量显著降低57mm（30%），增产423.23kg/hm2(4.6%)，达到节水增产目标。结合温度植被干旱指数（TVDI）的区域差异化水资源管理结果显示，在湿润区（TVDI为0~0.1），作物产量较高（>9500kg/hm2），但灌水量也较大。通过优化灌溉策略最大灌水量减少15%（83.17mm），避免了水资源浪费并保持高产；而在干旱区（TVDI大于0.4），灌水量不足导致产量较低（7500~8500kg/hm2），优化灌溉策略后灌水量增加37.15mm，产量提升2171.88kg/hm2，有效避免了干旱风险。整体而言，该优化方法使区域整体产量提高4.6%，灌水效率提升14%，灌水量减少11%。

Abstract:

Efficient water resources utilization and precision irrigation management are critical for improving agricultural productivity. Evapotranspiration (ET) estimation, a pivotal parameter in irrigation water management, has traditionally been limited by low spatiotemporal resolution, thereby constraining the implementation of precise irrigation practices. A framework combining remote sensing, data fusion, and multi-objective optimization for highresolution evapotranspiration (ET) estimation and irrigation management was presented. The framework used a spatiotemporal fusion model to generate accurate surface variables (NDVI, Albedo, land surface temperature), enabling daily field-scale (30 m×30 m) ET estimation. It also integrated the non-dominated sorting genetic algorithm Ⅱ (NSGA-Ⅱ) to optimize irrigation strategies for different districts. Results showed that the framework effectively addressed spatiotemporal variability, providing precise irrigation to meet crop water needs. Optimized irrigation reduced water use by 57mm during non-critical growth stages, increased crop yield by 423.23kg/hm2, achieving both water savings and yield enhancement. Analysis with the temperature vegetation drought index (TVDI) revealed spatial differences: in humid zones (00.4), where insufficient irrigation reduced yields (7500~8500kg/hm2), increased irrigation by 37.15mm, boosted yields by 2171.88kg/hm2, alleviating drought risks. Overall, the framework improved agricultural water management, increased regional yield by 4.6%, enhanced irrigation efficiency by 14%, and reduced irrigation volume by 11%. This decision-making framework at a 30m grid scale offered valuable insights for sustainable precision agriculture.