基于多通道奇异谱分析的北京房山站2007—2015年坐标时变分析  

Displacement analysis of BJFS station with MSSA from 2007to 2015

作　　者：高文宗 郭金运[1] 沈毅[1] 赵春梅[2] GAO Wenzong;GUO Jinyun;SHEN Yi;ZHAO Chunmei(College of Geometrics,Shandong University of Science and Technology,Qingdao,Shandong 266590,China;Chinese Academy of Surveying and Mapping,Beijing 100830,China)

机构地区：[1]山东科技大学测绘科学与工程学院,山东青岛266590 [2]中国测绘科学研究院,北京100830

出　　处：《中国科技论文》2018年第3期358-364,共7页China Sciencepaper

基　　金：国家自然科学基金资助项目(41374009;41774001);国家科技基础性工作专项(2015FY310200);山东省自然科学基金资助项目(ZR2013DM009)

摘　　要：为了得到国际全球导航卫星系统服务-北京房山站(international GNSS service BJFS,IGS BJFS)站的台站位移速率和周期变化特性,选取欧洲定轨中心(center for orbit determination in Europe,CODE)发布的BJFS站2007—2015年每周的参考坐标作为基础数据,剔除粗差并利用奇异谱迭代插值补全缺失数据;利用最小二乘法对处理后的时间序列进行线性趋势拟合,进而对扣除线性趋势项的时间序列进行多通道奇异谱分析(multichannel singular spectrum analysis,MSSA),发现BJFS站在水平方向具有向东南运动的趋势,E方向运动的位移速率为30.6mm/a,N方向运动的位移速率为-11.3mm/a,在U方向有抬升趋势,速率为2.0mm/a,并且N、E、U这3个方向均有1、0.5、4.9、2.5a的周期成分;将多通道奇异谱分析与最小二乘谱分析(least squares spectral analysis,LSSA)进行对比,发现MSSA拟合得到的3个方向的残差时间序列的RMS值均小于LSSA拟合的结果,说明MSSA相对于LSSA能够更加准确地提取出IGS站坐标时间序列中的信息。To obtain the station displacement rate and period variation characteristics of IGS BJFS station,weekly coordinates of BJFS station in IGS08 released by CODE were chosen as the basic data series from 2007 to 2015.The gross errors were removed and the missing data was completed with singular spectrum analysis(SSA)iterative interpolation.The least squares method was employed to estimate the linear trend of BJFS movement.After that,the coordinate series which has subtracted the trend are analyzed with the multichannel singular spectrum analysis(MSSA)to study the periodic variations of station BJFS in the north,east and up directions,respectively.BJFS has a tendency of southeastward movement in the horizontal direction with the eastward movement rate of 30.6 mm/a,and the southward movement rate of 11.3 mm/a.The vertical movement rate is 2.0 mm/a.There are 1 year,0.5-year,4.9-year and 2.5-year periodic variations in the north,east and up directions.Comparing the MSSA with least squares spectral analysis(LSSA),it is found that the RMS values of the residual time series in the three directions obtained by MSSA are less than those of LSSA fitting.The result of comparing shows that MSSA can extract information more accurately from IGSstation coordinates time series than LSSA.

关 键 词：IGS基准站 时间序列 多通道奇异谱分析 线性趋势 周期变化 

分 类 号：P228.4[天文地球—大地测量学与测量工程]