欢迎访问《中国农学通报》,

中国农学通报 ›› 2020, Vol. 36 ›› Issue (20): 51-58.doi: 10.11924/j.issn.1000-6850.casb20191000776

• 资源·环境·生态·土壤·气象 • 上一篇    下一篇

基于Sentinel多源数据的晋南农田地表土壤水分反演

刘正春1, 冯美臣2, 徐立帅1, 荆耀栋1, 毕如田1()   

  1. 1山西农业大学资源环境学院,山西晋中 030801
    2山西农业大学旱作农业工程研究所,山西晋中 030801
  • 收稿日期:2019-10-30 修回日期:2019-12-02 出版日期:2020-07-15 发布日期:2020-07-20
  • 通讯作者: 毕如田
  • 作者简介:刘正春,女,1986年出生,山西应县人,讲师,硕士,研究方向:农业遥感。通信地址:030801 山西太谷铭贤南路1号 山西农业大学资源环境学院,Tel:0354-6288226,E-mail:chun_5560@qq.com。
  • 基金资助:
    山西省重点研发计划项目“旱作小麦水分高效利用及化肥精准施用技术研究”(201803D221005-2);山西省软科学研究项目“山西省新型城镇化空间结构优化研究”(2016041027-1)

Soil Moisture Retrieval of Farmland in Southern Shanxi: Based on Sentinel Multi-source Data

Liu Zhengchun1, Feng Meichen2, Xu Lishuai1, Jing Yaodong1, Bi Rutian1()   

  1. 1College of Resources and Environment, Shanxi Agricultural University, Jinzhong Shanxi 030801
    2Institute of Dry Farming Engineering, Shanxi Agricultural University, Jinzhong Shanxi 030801
  • Received:2019-10-30 Revised:2019-12-02 Online:2020-07-15 Published:2020-07-20
  • Contact: Bi Rutian

摘要:

通过遥感数据反演农田地表土壤水分,对农作物干旱、长势监测及估产有重要指导作用。以山西省闻喜县冬小麦种植区为研究区,利用水云模型去除植被影响,建立土壤直接后向散射系数与土壤含水量的关系,反演闻喜县2018年3月19日冬小麦种植区土壤水分。结果表明:协同Sentinel-1微波和Sentinel-2光学影像数据能够去除植被影响提高土壤水分反演精度,VV极化决定系数R2提高0.0914,均方根误差RMSE减少0.0895%。闻喜县中部河谷平原以及东南和西北部的台地土壤处于轻度干旱状态,西南部丘陵和东部山地土壤处于中度干旱状态。反演的土壤水分空间分布结果与地形起伏、灌溉条件和地力等级有较好的空间一致性,地形起伏低、灌溉条件良好的农田土壤水分含量高。

关键词: 土壤水分, 冬小麦春旱, 水云模型, Sentinel数据, 地力等级

Abstract:

Using remote sensing data to retrieve soil moisture of farmland is essential for monitoring agricultural drought and crop status, and predicting crop yield. Taking the winter wheat planting area in Wenxi of Shanxi as the study area, we wiped off vegetation influence by “water-cloud model”, established the relationship between soil backscattering coefficient and soil moisture to retrieve the soil moisture of the winter wheat planting area of Wenxi on March 19 th 2018. The results showed that: fusion image of microwave (Sentinel-1 image) combined with visible spectrum remote sensing image (Sentinel-2 image) could remove vegetation influence to improve the retrieval accuracy of soil moisture, VV polarization determination coefficient R 2 was increased by 0.0914, and the RMSE was decreased by 0.0895%; the soil at middle valley plain and southeastern and northwestern tablelands of Wenxi was mildly drought, while it was moderately drought at the southwestern hills and eastern mountain lands. The spatial distribution of soil moisture retrieved is highly consistent with hypsographic condition, irrigation condition and farmland productivity grade, with low-lying terrains and favorable irrigation condition corresponding to regions of high soil moisture.

Key words: soil moisture, winter wheat spring drought, water-cloud model, Sentinel data, productivity grade

中图分类号: