Statistical and hydrological evaluation of the latest Integrated Multi-satellitE Retrievals for GPM (IMERG) over a midlatitude humid basin in South China
Statistical and hydrological evaluation of the latest Integrated Multi-satellitE Retrievals for GPM (IMERG) over a midlatitude humid basin in South China
复制标题
华南中纬度湿润盆地最新多星综合反演GPM（IMERG）的统计和水文评估
DOI：
10.1016/j.atmosres.2018.08.021
复制
发表时间：
2018-12-01
影响因子：
5.5
通讯作者：
Zeng, Xinmin
中科院分区：
文献类型：
--
作者：
Jiang, Shanhu;Ren, Liliang;Zeng, Xinmin
最近，具有高空间（0.1度x0.1度）和时间（半小时）分辨率的全球降水测量多卫星综合检索产品已投入使用。在这些新产品推广应用之前，对其进行综合评价是至关重要的。本研究重点关注最新IMERG的统计和水文评估（操作系统版本）产品：近实时的“早”跑和“晚”跑IMERG产品（分别为IMERG-E和IMERG-L），以及2014-2015年期间在中国南方中纬度潮湿的密水盆地的后实时“最终”运行IMERG产品（IMERG-F），与其前任相比，热带降雨测量使命多卫星降水分析产品（3B 42 RT和3B 42 V7）。实时后IMERG-F在五种卫星降水产品（SPP）中表现最好，在流域尺度上具有最高的日相关系数（CC）为0.85，最低的日均方根误差为5.58 mm。近实时IMERG-E和IMERG-L的性能与3B 42 V7相当。此外，泰勒图直观地证明了IMERG产品优于3B 42产品。对于情景I（基于雨量计观测的模型校准）下的水文模拟，后实时IMERG-F的性能明显优于3B 42 V7，其CC值为0.81，Nash-Sutcliffe系数为0.63，相对偏差为-3.98%。IMERG-E和IMERG-L的性能均优于3B 42 RT。对于情景二（基于每个卫星数据集的模型重新校准），IMERG和3B 42产品的水文性能得到改善。特别是IMERG-E，无论是在整个模拟期还是在汛期，其径流模拟性能都与3B 42 V7相当，这表明最新的近实时IMERG-E产品在洪水模拟和预测方面具有巨大的潜力。总的来说，这一系统的评价强调，最新的IMERG产品在研究区域具有理想的水文效用。该研究将为新一代SPP的水文应用以及IMERG算法开发提供有用的指导方针。
Recently, the Integrated Multi-satellitE Retrievals for Global Precipitation Measurement (IMERG) products with high spatial (0.1 degrees x 0.1 degrees) and temporal (half-hourly) resolution have become operationally available. It is of crucial importance to comprehensively evaluate these new products before they are applied extensively. This study focuses on statistical and hydrological evaluations of the latest IMERG (Version OS) products: the nearreal-time "Early" run and "Late" run IMERG products (IMERG-E and IMERG-L, respectively), and the post-realtime "Final" run IMERG product (IMERG-F) over the mid-latitude humid Mishui basin in South China during 2014-2015, in comparison with their predecessors, the Tropical Rainfall Measuring Mission Multi-satellite Precipitation Analysis (TMPA) products (3B42RT and 3B42V7). The post-real-time IMERG-F presents the best performance among the five satellite precipitation products (SPPs), with the highest daily correlation coefficient (CC) of 0.85 and the lowest daily root-mean square error of 5.58 mm at the basin scale. The near-real-time IMERG-E and IMERG-L demonstrate comparable performance with 3B42V7. In addition, the Taylor diagrams visually demonstrate that the IMERG products are better than the 3B42 products. For hydrological simulations under scenario I (model calibration based on rain gauge observations), the post-real-time IMERG-F performs obviously better than the 3B42V7 does, with a relatively high CC of 0.81, a good Nash-Sutcliffe coefficient of 0.63, and an acceptable relative bias of - 3.98%. Both the IMERG-E and IMERG-L demonstrate a better performance than the 3B42RT does. For scenario II (model recalibration based on each satellite dataset), the hydrological performances of both the IMERG and 3B42 products are improved. The IMERG-E, especially, demonstrates streamflow simulation performance comparable to that of the 3B42V7, both for the whole simulation period and flood season, indicating the great potential of the latest near-real-time IMERG-E product for flood simulation and prediction. Overall, this systematic evaluation highlights that the latest IMERG products have desirable hydrological utility in the study region. This study will provide useful guidelines for hydrological applications of the new generation SPPs, as well as IMERG algorithm development.