遥感和再分析蒸散发数据精度评估:基于GRACE和流域月水量平衡模型的比较研究  

Accuracy validation of remote sensing and reanalysis evapotranspiration data:A comparative study based on GRACE and monthly water balance model

作　　者：云兆得 胡庆芳[1,2] 王银堂 吴海燕[3] 王磊之 商守卫 YUN Zhaode;HU Qingfang;WANG Yintang;WU Haiyan;WANG Leizhi;SHANG Shouwei(Nanjing Hydraulic Research Institute,State Key Laboratory of Hydrology,Water Resources&Hydraulic Engineering Science,Nanjing 210029,China;Yangtze Institute for Conservation and Development,Nanjing 210098,China;Huzhou Municipal Bureau of Water Resources,Huzhou 313000,China)

机构地区：[1]南京水利科学研究院水文水资源与水利工程科学国家重点实验室,江苏南京210029 [2]长江保护与绿色发展研究院,江苏南京210098 [3]湖州市水利局,浙江湖州313000

出　　处：《水利学报》2023年第1期117-127,共11页Journal of Hydraulic Engineering

基　　金：国家自然科学基金面上项目(51479118);国家重点研发计划课题(2019YFC0408903)。

摘　　要：采用GRACE陆地水储量变化数据和两参数月水量平衡模型(WBM-DP)分别构建了流域实际蒸散发月序列ET GRACE和ET WBM,以此为基准评价了两种遥感蒸散发数据MOD16、SSEbop和一种大气再分析蒸散发数据GLDAS-Noah在汉江流域上游的性能,解析了基准序列不同对蒸散发数据精度的影响。结果表明,GRACE数据和WBM-DP得到的2006—2014年汉江流域上游蒸散发基准序列在年尺度上较接近,但在月尺度上差异明显。WBM-DP可较好地模拟降水和潜在蒸散发作用下流域水储量的季节性变化特征,构建的蒸散发基准序列也更合理;而GRACE数据反映的流域水储量变化存在不合理现象,由此根据水量平衡方程推算的月蒸散发基准序列也具有异常现象。在月尺度上,基准序列的不同会导致蒸散发数据精度指标的非一致性变化,其精度评估结果也可能不同,在非汛期各月,以ET WBM为基准时,MOD16综合精度最优;而以ET GRACE为基准时,SSEbop最优。且相对于ET GRACE,以ET WBM为基准时,MOD16、Noah和SSEbop的定量精度差异更加显著。本文深化了对GRACE数据适用性的认识,可为卫星遥感和大气再分析蒸散发数据的精度评价和数据优选提供参考。The actual evapotranspiration monthly series ET GRACE and ET WBM were constructed by using the GRACE terrestrial water storage change data and monthly Water Balance Model with Double Parameters(WBM-DP)respectively.Taking this as the benchmark,the performance of two remote sensing evapotranspiration data(MOD16,SSEbop)and an atmospheric reanalysis data(GLDAS-Noah)in the upper Hanjiang River basin were evaluated.The results show that the evapotranspiration reference series obtained by GRACE data and WBM-DP in the upper Hanjiang River basin from 2006 to 2014 are close at the annual scale,but there are significant differences at the monthly scale.WBM-DP can simulate the seasonal variation characteristics of terrestrial water storage better under the influence of precipitation and potential evapotranspiration,and WBM-DP-based evapotranspiration benchmark sequence is also more reasonable.However,the variation of water storage in the basin reflected by the GRACE data is unreasonable,so the benchmark sequence of monthly evapotranspiration calculated according to the water balance equation also has anomalies.On the monthly scale,the difference of the benchmark series will lead to non-uniform changes in the precision index of evapotranspiration data,so its impact cannot be ignored.At the same time,the results of accuracy evaluation may also be different.In each month of non-flood period,when ET WBM is used as the benchmark,the comprehensive accuracy of MOD16 is the best;when ET GRACE is taken as the benchmark,SSEbop is the best.Compared with ET GRACE,the quantitative accuracy differences of MOD16,Noah and SSEbop are more significant when ET WBM is used as the benchmark.The paper deepens the understanding of the applicability of GRACE data,and provides a reference for the accuracy evaluation and data optimization of satellite remote sensing and atmospheric reanalysis evapotranspiration data.

关 键 词：GRACE 水量平衡模型 遥感蒸散发 大气再分析 

分 类 号：P407[天文地球—大气科学及气象学] P426.2