Evaluation of seven satellite-based and two reanalysis global terrestrial evapotranspiration products
Evaluation of seven satellite-based and two reanalysis global terrestrial evapotranspiration products
复制标题
DOI：
10.1016/j.jhydrol.2024.130649
复制
发表时间：
2024-01
影响因子：
6.4
通讯作者：
Zijing Xie;Yunjun Yao;Qingxin Tang;Meng Liu;Joshua B. Fisher;Jiquan Chen;Xiaotong Zhang;Kun Jia-K
中科院分区：
文献类型：
--
作者：
Zijing Xie;Yunjun Yao;Qingxin Tang;Meng Liu;Joshua B. Fisher;Jiquan Chen;Xiaotong Zhang;Kun Jia-K
虽然已经对不同类型的全球陆地蒸散量（ET）产品进行了全面的评估，但卫星遥感技术促进了几种可用的全球ET产品的开发，随着产品的不断发展，需要重新评估。最近，我们制作了长期的全球陆地卫星（GLASS）ET产品，但缺乏与其他ET产品和EC观测在全球范围内的比较和评估。在这项研究中，我们评估的精度和不确定性的七个卫星为基础的（GLASS-AVHRR，GLASS-MODIS，BESS，FLUXCOM，GLEAM，MOD 16，和PML_V2）和两个再分析（ERA 5和MERRA 2）全球陆地ET产品在多个尺度上选择最合适的ET产品和开发大尺度ET模式。在点尺度上，通过与230个全球通量塔的现场观测结果进行直接比较，评估了其准确性。结果表明，没有一个ET产品可以为所有土地覆盖类型提供最准确的ET估计，尽管GLASS-MODIS [决定系数（R2）为0.51，Kling-Gupta效率（KGE）为0.68]和FLUXCOM [R2为0.51，KGE为0.66]优于其他七个产品[0.54 =< KGE =< 0.66，0.35 =< R2=< 0.50]。在流域尺度上，通过36个大型流域评估了ET产品的准确性。所有ET产品与水分平衡ET（WBET）的R2值均大于0.88，但9种ET产品的精度有一定差异。应用三角帽法和比较分析法对9种ET产品在像素级上的不确定度进行了评定。TCH结果显示，GLASS-AVHRR和GLASS-MODIS是相对不确定度最低的两个产品，而MERRA 2具有最大的相对不确定度。我们的研究结果表明，没有一个单一的ET产品在所有方面都表现最好。科学研究中ET产品的选择应考虑其在空间尺度上的性能差异以及土地覆盖和气候条件的影响。
Although comprehensive evaluation of different types of global terrestrial evapotranspiration (ET) products has been conducted, the satellite remote sensing techniques have prompted the development of several available global ET products, warranting a reassessment as products continue to evolve. Recently, we produced the long-term Global LAnd Surface Satellite (GLASS) ET product, but there is a lack of comparison and evaluation with other ET products and EC observations on a global scale. In this study, we evaluated the accuracy and uncertainty of seven satellite-based (GLASS-AVHRR, GLASS-MODIS, BESS, FLUXCOM, GLEAM, MOD16, and PML_V2) and two reanalysis (ERA5 and MERRA2) global terrestrial ET products at multiple scales for selecting the most suitable ET products and developing large-scale ET models. At the point scale, their accuracy was evaluated through direct comparison with in situ observations from 230 global flux towers. The results indicate that no single ET product can provide the most accurate ET estimates for all land cover types, although GLASS-MODIS [coefficient of determination (R2) of 0.51, Kling–Gupta efficiency (KGE) of 0.68] and FLUXCOM [R2of 0.51, KGE of 0.66] outperform the seven other products [0.54 =< KGE =< 0.66, 0.35 =< R2=< 0.50] at all sites. At the basin scale, the accuracy of ET products was assessed through 36 large river basins. The R2values between all ET products and the water balance-derived ET (WBET) are >0.88, while the accuracies of the nine ET products differ in some sense. The three-cornered hat (TCH) method and comparison analysis are applied to assess the uncertainty of nine ET products at the pixel level. The TCH outputs reveal that GLASS-AVHRR and GLASS-MODIS are the two products with the lowest relative uncertainty, while MERRA2 has the largest relative uncertainty. Our results imply that there is no single ET product performing best in all respects. The selection of ET products for scientific research should consider their performance differences in spatial scale as well as the influence of land cover and climate conditions.