西北干旱区农田灌溉事件遥感检测  

Detecting farmland irrigation in arid areas of Northwest China using remote sensing

作　　者：李文佳 王瑞燕[1] 徐家豪 王汝浩 李晓藤 LI Wenjia;WANG Ruiyan;XU Jiahao;WANG Ruhao;LI Xiaoteng(College of Resources and Environment,Shandong Agricultural University,Tai’an 271018,China)

机构地区：[1]山东农业大学资源与环境学院,泰安271018

出　　处：《农业工程学报》2024年第17期120-128,共9页Transactions of the Chinese Society of Agricultural Engineering

基　　金：山东省研究生教育质量提升计划项目(SDYAL21044);山东省自然科学基金项目(ZR2020MD003,ZR2022MD026)。

摘　　要：快速获取灌溉事件的时空分布信息,对于耕地质量监测保护和应对气候变化具有重要意义。该研究以中国西北干旱区甘肃省张掖市甘州区为研究区,通过对常规方法、基于MOD09GA光学遥感数据的近红外波段反射率法、分段PDI(perpendicular dryness index)差异阈值模型法、微波水分产品以及PDI差异阈值法与微波水分产品相结合的方法进行对比分析,研究适用于不同植被覆盖状况的全年度农田灌溉事件的遥感检测方案,并从时空维度对优选出的方法进行检验。结果显示,PDI差异阈值与归一化植被指数(normalized difference vegetation index,NDVI)呈强负线性相关(R2为0.879),由此构建了分段PDI差异阈值模型;研究结果表明,微波水分产品结合PDI差异阈值法检测灌溉事件的精度最高,该方法在干旱年的效果最优,在大满站无漏检测、无多检测,检测到党寨站点记录的2次灌溉事件,检测率OA为1,其次是平常年。在甘州区的区域反演结果显示,微波水分产品结合PDI差异阈值法制作的灌溉次数分布图与实际情况一致,甘州区灌溉能力较差的农田主要分布在离河流较远的张掖盆地的边缘,这些地方可能面临更大的气候变化风险。因此,在中国西北干旱区灌溉事件的检测应用中适宜采用干旱年份的微波水分产品结合PDI差异阈值法。该研究结果为实际灌溉信息的客观高效获取,开展耕地质量保护和应对气候变化提供了理论依据。The information of the temporal and spatial distribution of irrigation events is of great significance for protecting the cultivated land quality and coping with climate change.Irrigation information can traditionally be captured using farmland water conservancy construction and field investigation.However,there is a great variation in the time and space of irrigation information from the different regions,due to the various water sources,planting structures,irrigation technology,water transmission,and measurement facilities.At the same time,remote sensing can be expected to timely and accurately monitor the ground surface all day.The previous research focused mainly on the crop growing season without considering the influence of different vegetation coverage on the irrigation remote sensing signals.Therefore,it is very necessary to detect the year-round events of farmland irrigation using remote sensing,particularly for different vegetation coverage.In this study,the farmland irrigation events were detected in the arid areas of northwest China using remote sensing.The research area was taken as Daman Super Station in Ganzhou District,Zhangye City,Gansu Province,China.The comparison was also performed on the conventional,near-infrared band reflectivity with MOD09GA optical remote sensing data,segmented PDI differential threshold model,microwave moisture products,and the combination of PDI differential threshold and microwave moisture products.An optimal scheme was achieved in the remote sensing detection suitable for all-year farmland irrigation events with different vegetation coverage.The performance was finally tested from the space-time dimension.The results show that there was a strong negative linear correlation between the PDI difference threshold and NDVI.Thus,a segmented PDI difference threshold model was constructed in 2016;The highest accuracy of irrigation events was achieved in the microwave moisture products combined with the PDI differential threshold.There was a close relationship between the soil

关 键 词：遥感 农田 灌溉事件 干旱地区 检测 

分 类 号：S24[农业科学—农业电气化与自动化]