Effects of Gaussian filter in processing GRACE data: Gravity rate of change at Lhasa,southern Tibet

作     者：SUN WenKe HASEGAWA Takashi ZHANG XinLin FUKUDA Yoichi SHUM C. K. WANG Lei 

作者机构：Key Laboratory of Computational Geodynamics Graduate University of Chinese Academy of Sciences Beijing 100049 China Department of Geophysics Kyoto University Japan Earthquake Research Institute The University of Tokyo Japan Division of Geodetic Science School of Earth Sciences The Ohio State University OH 43210-1398 USA 

出 版 物：《Science China Earth Sciences》 (中国科学（地球科学英文版）)

年 卷 期：2011年第54卷第9期

页      面：1378-1385页

核心收录：

学科分类：070801[理学-固体地球物理学] 07[理学] 08[工学] 0708[理学-地球物理学] 0835[工学-软件工程] 0802[工学-机械工程] 080201[工学-机械制造及其自动化] 

基　　金：study was supported by NASA’s Interdisciplinary Science Program (Grant No. NNG04GN19G) the Ohio State University Climate, Water, and Carbon Program 

主　　题：gravity change GRACE Gaussian filter Tibetan Plateau Lhasa 

摘      要：In this paper, the spatial gravity distribution over Tibetan Plateau and the gravity rate of change at Lhasa for different Gaussian filter radii are computed using GRACE data. Results show that the estimate of the gravity rate of change is spatialradius-dependent of the Ganssian filter. The GRACE-estimated gravity rate of change agrees well with the surface measured one. In other words, the GRACE-estimated gravity rate of change has a limited value as that obtained by surface measurement when the spatial filter radius reaches zero. Then numerical simulations are made for different spatial radii of the Gaussian filter to investigate its behaviors when applied to surface signals. Results show that the estimate of a physical signal is filter-radius dependent. If the computing area is equal to or less than the mass area, especially for a uniformly distributed mass, the estimate gives an almost correct result, no matter what filter radius is used. The estimate has large error because of the signal leakage caused by harmonic truncation if the computing area is much bigger than the mass distribution （or inverse for a small mass anomaly）. If a mass anomaly is too small, it is difficult to recover it from space observation unless the filter radius is extremely small. If the computing point （or area） is outside the mass distribution, the estimated result is almost zero, particularly for small filter radii. These properties of the Gaussian filter are helpful in applying GRACE data in different geophysical problems with different spatial position and geometrical size. We further discuss physical sources causing the scalar gravity change at Lhasa. Discussions indicate that the gravity rate of change at Lhasa is not caused by the present-day ice melting （PDIM） （or Little Ice Age, LIA） effect because no ice melting occurs in Lhasa city and nearby. The gravity rate of change is attributable mainly to tectonic deformation associated with the Indian Plate collision. Simultaneous surf