一种数据驱动的GRACE泄漏误差改正方法  

A data-driven method for GRACE leakage error correction

作　　者：戴纤蕴 钟波[1,2] 李贤炮 DAI Xianyun;ZHONG Bo;LI Xianpao(School of Geodesy and Geomatics,Wuhan University,Wuhan 430079,China;Key Laboratory of Geospatial Environment and Geodesy,Ministry of Education,Wuhan University,Wuhan 430079,China)

机构地区：[1]武汉大学测绘学院,武汉430079 [2]武汉大学地球空间环境与大地测量教育部重点实验室,武汉430079

出　　处：《测绘科学》2022年第7期43-52,共10页Science of Surveying and Mapping

基　　金：国家自然科学基金项目(41974015,42061134007)

摘　　要：针对重力恢复与气候试验时变重力场南北条带误差和高频误差滤波处理导致的信号泄漏问题,该文讨论了一种数据驱动的泄漏误差改正方法。通过闭合数值模拟进行有效性检验,将其用于反演12个典型流域在2004—2010年的陆地水储量变化,并与传统的尺度因子法进行比较。结果表明,这两种方法的改正结果具有很好的一致性,并且数据驱动法在多数流域的改正结果与Mascon模型更为接近。同时,它们改正得到的陆地水储量变化趋势及相位与Mascon模型较为一致,但在个别流域与Mascon模型的振幅差异较大。与尺度因子法相比,数据驱动法并不需要水文模型作为参考,其应用范围及适应性更好。Aiming at the problem of reducing the leakage error introduced by filtering longitudinal stripes and high-frequency noise in the gravity recovery and climate experiment(GRACE) data, the paper discussed a data-driven method used for correcting leakage error. The effectiveness of the algorithm was tested using a closed-loop numerical simulation from January 2004 to December 2010 in twelve typical river basins, compared with results from the traditional scale-factor method. The results demonstrated that the water storage changes inferred by the two methods showed significant agreement, and the results corrected by the data-driven method agreed better with Mascon solutions in most of the basins. Furthermore, the annual trends and phases of water storage changes corrected by the two methods agreed well with Mascon models, but the annual amplitudes for individual river basins differed significantly. The data-driven method did not require the use of hydrological model as a reference, so it had a better application and adaptability than the scale-factor method.

关 键 词：数据驱动方法 泄漏误差改正 尺度因子方法 陆地水储量变化 重力恢复与气候试验 

分 类 号：P228[天文地球—大地测量学与测量工程]