MSSA在GPS坐标序列分析中的应用

本文选取CMONOC位于云南省域的4个测站坐标序列作为试验对象，利用多通道奇异谱分析（MSSA）对序列进行缺失数据的插补及季节性信号的提取。试验结果表明，MSSA迭代插补精度在水平方向优于1.5 mm，在垂直方向上优于4 mm，表现均好于三次样条插值，而且当序列的缺失率升高、连续性减弱时依然不失其优越性；对比MSSA及最小二乘拟合两种方法提取季节性信号的效果并对其残差序列及功率谱进行分析，发现MSSA能够有效提取不同周期信号且更加适用于振幅不恒定的时变季节信号。

Application of MSSA in GPS coodinate series analysis

The coordinate series of four CMONOC stations located in Yunnan province are selected as the experimental objects, and the multichannel singular spectrum analysis(MSSA) is used to interpolate its missing data and extract its seasonal signal. The results show that the precision of iterative interpolation of MSSA is better than 1.5 mm in the horizontal direction and 4 mm in the vertical direction, both better than cubic spline interpolation and the superiority is still not be lost when the missing rate of the sequence increases and the continuity decreases. Comparing MSSA and leastsquare fit methods to extra ctseasonal signal extraction effect and analyze its residual sequence and power spectrum, it is found that MSSA can effectively extract different periodic signals and is more suitable for time-varying seasons with inconstant amplitude signal.