利用GPS和GRACE分析四川地表垂向位移变化

陆地水储量的季节性变化是导致地表周期性负荷形变位移的主要因素，有效地剔除地表位移中的陆地水储量影响，是获取地壳构造垂向运动的必要过程.四川地处青藏高原东边缘，地形分区明显，境内以长江水系为主，水资源丰富，研究四川地区地表负荷形变位移，有助于分析陆地水储量的时空分布特性及地壳构造形变信息.本文利用研究区域内59个CORS站的GPS观测数据，计算了CORS站点的垂向位移，并将其与GRACE所得相应结果进行对比分析.结果显示，GPS和GRACE所得垂向位移时间序列的振幅大小整体相符，但存在明显的相位差.GPS站点振幅最大值为12.7 mm，对应HANY站，最小值为1.5 mm，对应SCMX站.GRACE所得的地表垂向位移振幅大小均为3~4 mm，且最大位移集中出现在7-9月份；而GPS站点出现最大位移的月份和地形相关，东部盆地、西北部高原和南部山地分别出现在7-8月份、10-11月份和10月份.GPS站点时间序列中的周年项与陆地水的季节性变化强相关，为了讨论陆地水储量对GPS站点位移的影响，本文利用改进的总体经验模态分解方法（MEEMD：Modified Ensemble Empirical Mode Decomposition），从GPS垂向位移时间序列中提取出周年项及约2年的年际变化项.发现利用MEEMD获取的周年项改正原始GPS时间序列时可使其WRMS（Weight Root Mean Square）减少量减小约26%，结果优于最小二乘拟合方法提取的GPS周年项改正效果，验证了MEEMD方法在GPS坐标时间序列处理中的可行性及有效性.

Determination of vertical surface displacements in Sichuan using GPS and GRACE measurements

Seasonal variations of terrestrial water storage (TWS) mainly result in periodical fluctuations of crustal deformation, eliminating the effects from TWS on surface displacement is accordingly conducive to highlight regional tectonic movements. Sichuan Province lies close to the eastern margin of the Qinghai-Tibetan Plateau and has a rising terrain feature from its east to the west. Moreover, abundant river systems, such as Yangtze River system, provide Sichuan with adequate water resource. We combine GPS-derived and GRACE-derived vertical displacement time series to explore the spatial-temporal distribution characteristics of TWS and the crustal tectonic movements in Sichuan. The result shows that the amplitudes of vertical displacements derived by those two ways are accorded very well, but there exist phase delays. The GPS-determined amplitudes show a maximum of 12.7 mm at station HANY and minimum of 1.5 mm at station SCMX, respectively. Due to a low spatial resolution, GRACE-derived amplitudes present comparable magnitude of 3~4 mm and the largest displacements are in between July and September. However, the time when the largest displacements of GPS happen correspond with topography, the eastern basin and hills, northwestern plateau and southern mountains possess its own largest displacement in July to August, October to November and October, separately. Seasonal TWS changes are relevant to GPS-derived vertical displacement variations. So, in order to illustrate the effect of TWS changes on GPS sites' movements, we use the Modified Ensemble Empirical Mode Decomposition (MEEMD) to extract annual term and a 2-year periodic signal from GPS vertical displacement time series. Consequently, we find that when using MEEMD-derived annual terms to correct GPS time series, the WRMS (Weight Root Mean Square) reduction can decrease roughly 26%, which is superior to least square method, suggesting that the MEEMD method is feasible in dealing with GPS time series.