基于时序Sentinel-2影像的引黄灌区作物结构提取和供需水分析  

Extraction of Crop Structure and Analysis of Water Supply and Demand in Irrigation Area of Yellow River Diversion Based on Sentinel⁃2 Image

作　　者：孙斌[1] 毕春宁 薛建春[1] 毕华军[2] 孙力[2] 许建辉 李斌 SUN Bin;BI Chunning;XUE Jianchun;BI Huajun;SUN Li;XU Jianhui;LI Bin(School of Economics and Management,Inner Mongolia University of Science and Technology,Baotou 014010,China;Shandong Institute of Water Resources Science,Jinan 250014,China;Water Resources Bureau of Kenli District of Dongying City of Shandong Province,Dongying 257500,China)

机构地区：[1]内蒙古科技大学经济与管理学院,内蒙古包头014010 [2]山东省水利科学研究院,山东济南250014 [3]山东省东营市垦利区水利局,山东东营257500

出　　处：《人民黄河》2024年第7期131-137,共7页Yellow River

基　　金：国家自然科学基金资助项目(42061051);内蒙古自然科学基金资助项目(2021LHMS07002)。

摘　　要：在黄河流域用水指标严格控制的背景下,以山东省东营市垦利区引黄灌区为例,利用2022年时序Sentinel-2遥感影像构建作物生育期的NDVI时间序列,采用决策树分类方法提取灌区作物种植结构,基于垦利站气象资料和Penman-Monteith公式,分析了1973—2022年各作物的需水特性,利用遥感影像解译的各作物种植面积,计算了2022年灌区作物在不同降水保证率(5%、25%、50%、75%、95%)条件下的灌溉总需水量,结合2023年分配给灌区作物的灌溉水指标探究了灌溉水资源供需之间的平衡。结果表明:基于NDVI时间序列构建决策树分类方法可有效提取作物的种植结构,总体分类精度为85.07%,Kappa系数为0.819,能够满足作物灌溉需水量的研究。作物净灌溉需水量年际波动较大,水稻和冬小麦补充灌溉水量在所有作物中位列前两位,均值分别为913 mm和410 mm;处于雨季生长的夏玉米、夏大豆补充灌溉水量较小且灌溉需求均值较小。研究区2023年分配的灌溉水指标在降水保证率为50%时研究区灌溉水亏缺量为235.5万m^(3),在降水保证率为75%和95%时灌溉水亏缺量分别为1 754.5万m^(3)和2 261.5万m^(3)。Under the background of strict control of water consumption index in the Yellow River Basin,taking Kenli District of Dongying City,Shandong Province as an example,NDVI time series of crop growth period was built by using Sentinel-2 remote sensing images in 2022,and the crop planting structure in the irrigation area was extracted by decision tree classification method.Based on the meteorological data of Kenli Station and Penman-Monteith formula,the water demand characteristics of each crop during 1973 to 2022 were analyzed,and the area results of each crop interpreted by remote sensing images were used as the base.The total irrigation water demand of crops in 2022 under different precipitation guarantee rates(5%,25%,50%,75%and 95%)was calculated,and the balance between supply and demand of irrigation water resources was explored combined with the irrigation water index allocated to crops in 2023.The results show that the deci-sion tree classification method based on NDVI time series can effectively extract the planting structure of crops,the overall classification ac-curacy is 85.07%,and the Kappa coefficient is 0.819,which can meet the research of crop irrigation water requirement.The net irrigation water requirement of crops is fluctuated greatly from year to year,and the supplementary irrigation amount of rice and winter wheat ranked first and second among all crops,with an average of 913 mm and 410 mm respectively.For summer corn and summer soybean growing in rainy season,the amount of supplementary irrigation is smaller and the average irrigation demand is lower.The irrigation water deficit alloca-ted in the study area in 2023 is 2.355 million m^(3) at 50%guarantee rate,and 17.545 million and 22.615 million m^(3) at 75%and 95%guaran-tee rates.

关 键 词：Sentinel-2影像 种植结构 需水特性 灌溉水供需 引黄灌区 

分 类 号：S127[农业科学—农业基础科学] S274