【摘要】 地球自转速率的研究是天文地球动力学和自然灾害的主要研究内容,其变化影响日长（LOD）及人类的生产生活,而利用不同频谱分析方法均可以提取LOD的主要周期成分。为了研究非等间距时间序列的有效观测数据量及样条插值分别对不同频谱方法分析结果的影响程度,本文采用最大熵法（Maximum Entropy Method,简称MEM）、Lomb算法（简称LOMB）、相位弥散法（Phase Dispersion Minimization,简称PDM）对LOD进行频谱分析。分别随机扣除月平均日长数据序列（LODM）样本数的1/4、1/3和1/2,可构成非等间距日长变化数据序列（LODMD）,再进行LOMB和PDM的频谱分析,与样条插值后的新序列（LODMDN）进行MEM、LOMB、PDM频谱分析结果的比较。结果表明:1)对于LODM的频率结果,MEM的双峰总维持在0.1660和0.0840月^-1,而LOMB和PDM的峰在0.166和0.083月^-1;2)对于LOD的两个主要周期,有效观测数据量（超过等间隔应有样本数的一半）或样条插值对于三种频谱分析方法的结果几乎没有影响。更多还原

【Abstract】 The variations in the Earth’s rotation are important to space dynamic theory and natural disasters because they affect the length-of-day（LOD） and consequently human lives. We use maximum entropy method（MEM）, Lomb method（LOMB）, and phase dispersion minimization（PDM） to determine the natural periods of equally spaced LOD time series. We transform the observational monthly LOD time series（LODM） to unequally sampled series（LODMD） by removing every fourth, third, and half of the total samples. We also apply spline interpolation to LODMD to yield equally spaced time series（LODMDN）. The results suggest that regardless of the time series, the MEM frequency is 0.1660 month^-1 and 0.0840 month^-1, whereas LOMB and PDM yield 0.166 month^-1 and 0.083 month^-1, respectively. Furthermore, missing data that are less than half of the total data or spline interpolation do not affect the analysis. For the amplitude, neither missing data nor spline interpolation affect the analysis.更多还原