当前位置: X-MOL 学术Remote Sens. Environ. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
In-situ and triple-collocation based evaluations of eight global root zone soil moisture products
Remote Sensing of Environment ( IF 13.5 ) Pub Date : 2021-03-01 , DOI: 10.1016/j.rse.2020.112248
Lei Xu , Nengcheng Chen , Xiang Zhang , Hamid Moradkhani , Chong Zhang , Chuli Hu

Abstract Root zone soil moisture (RZSM) is a vital variable for vegetation growth, drought monitoring and agricultural water management. Satellite remote sensing measures soil moisture at the surface layer, while RZSM is derived usually by model-based simulations. Here, we provide the first comprehensive evaluation of eight RZSM products at a global scale, including GLDAS NOAH, ERA-5, MERRA-2, NCEP R1, NCEP R2, JRA-55, SMAP level 4 and SMOS level 4 datasets. An in-situ validation based on the stations from the International Soil Moisture Network (ISMN) and a triple collocation (TC) evaluation are both conducted to assess the accuracy of these RZSM products. SMAP exhibits the median highest correlation and the median lowest RMSE with in-situ stations over North America. In the TC analysis, MERRA-2 shows the highest median correlation and the median lowest error standard deviation with the unknown truth, followed by GLDAS, SMAP, JRA-55 and ERA-5. A temporal pattern analysis indicates that SMOS has a dry bias relative to other datasets and NCEP R1 has larger seasonal variations relative to other datasets over Asia and North America. The TC analysis indicates that MERRA-2, SMAP, GLDAS, JRA-55, and ERA-5 have better performance relative to other datasets. SMAP is not as good as GLDAS, MERRA-2 and JRA-55 in RZSM estimation over forest areas. The possible factors influencing RZSM performance are discussed, including precipitation forcing, assimilated observations, radio frequency interference issue and validation methods. These results and conclusions may provide new insights for the improvement of model-based RZSM estimation.

中文翻译:

八种全球根区土壤水分产品的原位和三重搭配评估

摘要 根区土壤水分(RZSM)是植被生长、干旱监测和农业用水管理的重要变量。卫星遥感测量表层土壤水分,而 RZSM 通常通过基于模型的模拟得出。在这里,我们首次在全球范围内对八种 RZSM 产品进行综合评估,包括 GLDAS NOAH、ERA-5、MERRA-2、NCEP R1、NCEP R2、JRA-55、SMAP 4 级和 SMOS 4 级数据集。进行了基于国际土壤水分网络 (ISMN) 站点的原位验证和三重搭配 (TC) 评估,以评估这些 RZSM 产品的准确性。SMAP 与北美现场台站的相关性中值最高,RMSE 中值最低。在TC分析中,MERRA-2 显示了最高的中值相关性和中值最低的错误标准偏差与未知真相,其次是 GLDAS、SMAP、JRA-55 和 ERA-5。时间模式分析表明,SMOS 相对于其他数据集具有干偏差,而 NCEP R1 相对于亚洲和北美的其他数据集具有更大的季节性变化。TC 分析表明,相对于其他数据集,MERRA-2、SMAP、GLDAS、JRA-55 和 ERA-5 具有更好的性能。SMAP 在森林区域的 RZSM 估计方面不如 GLDAS、MERRA-2 和 JRA-55。讨论了影响 RZSM 性能的可能因素,包括降水强迫、同化观测、射频干扰问题和验证方法。这些结果和结论可能为改进基于模型的 RZSM 估计提供新的见解。
更新日期:2021-03-01
down
wechat
bug