摘要
目前,灌溉面积主要通过调查统计方式获取,时效性和准确性较差,而在灌区用水管理中迫切需要掌握实际灌溉面积及其空间分布和动态变化等信息。为此,本文开展基于雷达遥感信息的灌溉信号探测识别及次灌面积提取方法研究。首先,使用水云模型结合实测土壤水分数据进行模拟,量化灌溉事件导致后向散射系数的变化。然后,提出一种基于时序差值和局部阈值法的实际灌溉面积提取方法,使用高频次的哨兵1号雷达卫星遥感数据开展华北平原灵寿县磁右灌区灌溉面积提取,并结合灌区实地调查资料对提取结果进行验证。结果表明:使用水云模型模拟显示春季小麦地块的灌溉事件可导致C波段的雷达后向散射系数变化超过1 dB,灌溉信号可被绝对辐射精度优于1 dB的雷达卫星探测;基于时序雷达后向散射系数分析显示,灌溉事件导致C波段雷达后向散射系数呈现明显的先增大后减小的时域特征,与降水事件不同,灌溉事件导致雷达后向散射系数呈现明显的局地变化特征。以本文提出的方法提取研究区的灌溉面积,结合实地调查样本检验,结果表明灌溉面积的总体提取精度为76.6%。小麦返青期第一次渠灌面积约为9.20万亩,小麦拔节期第二次渠灌面积约为10.64万亩。此外,由于井灌区和渠灌区存在明显的时差,基于雷达遥感信息可对不同水源的灌溉区进行探测,研究区井灌面积约为3.9万亩。基于高频次的雷达遥感信息可对灌区的灌溉事件进行探测,本文提出的方法可为区域次灌溉面积调查和流域水资源管理等应用提供技术支撑。
Currently, statistical investigation is the main approach to derive irrigated area, with limited timeliness and accuracy, while the information on actual irrigated area alongside its spatial distribution and dynamic changes are urgently demanded in water management within the irrigated area.In this regard, an optical and microwave-based method is proposed for detecting irrigation signal and extracting actual irrigated area.First, simulations were performed using water cloud model combined with in-situ soil moisture measurements to quantify the changes in backscatter coefficients due to irrigation events.Then, using backscatter coefficients from Sentinel-1 with high revisit frequency, a method based on time-series difference and local threshold is proposed, and conducted for extraction of irrigated area in Ciyou irrigation area of Lingshou County, North China Plain.The results show that irrigation events in spring wheat plots can lead to fluctuation of radar backscatter coefficients in C band by more than 1 dB,thus irrigation signals can be detected by radar satellites with an absolute accuracy superior to 1 dB.Time-series of radar backscatter coefficients analysis indicates that due to irrigation events, significant increase of C-band radar backscatter coefficients will be followed by obvious decrease.Unlike precipitation event, local features are prominent in irrigation-induced radar backscatter coefficients variation.Through field survey sample validation, the proposed method showed overall accuracy of 76.6% in extracting actual irrigated area.The areas of first canal irrigation during the wheat seedling establishment period and during elongation period are about 92,000 and 106,400 acres respectively.In addition, due to the obvious time difference between well irrigation area and canal irrigation area, the irrigation area of different water sources can be detected based on radar remote sensing, and the well irrigation area in the study area is about 39,000 acres.The proposed method can provide technical support for applications such as regional irrigation area surveys and basin water resource management.
作者
杨永民
顾涛
吴迪
龙爱华
邱建秀
刘宏鑫
YANG Yongmin;GU Tao;WU Di;LONG Aihua;QIU Jianxiu;LIU Hongxin(China Institute of Water Resources and Hydropower Research,Beijing 100038,China;Research Canter on Flood and Drought Disaster Reduction of the Ministry of Water Resources,Beijing 100038,China;China Irrigation and Drainage Development Center,Beijing 100054,China;School of Geography and Planning,Sun Yat-sen University,Guangzhou 510275,China)
出处
《水利学报》
EI
CSCD
北大核心
2022年第9期1039-1048,共10页
Journal of Hydraulic Engineering
基金
国家自然科学基金项目(41501415)
中国水科院重点专项(JZ0145B032017)
高分水利遥感应用示范系统(二期)项目(08-Y30F02-9001-2022)
第三次新疆科学考察伊犁河流域水资源利用及其影响考察评估(2021xjkk0400)。
关键词
灌溉面积
雷达遥感
后向散射系数
水云模型
irrigated area
radar remote sensing
backscatter coefficients
water cloud model
