Computation of short periodical variations of pole coordinates using maximum entropy spectral analysis and an ormsby filter

Abstract

The main purpose of this paper is the search for the optimum method of detecting weak short periodical variations of pole coordinates determined by different techniques in theMERIT Campaign. The optimum filter, length of Maximum Entropy Spectral Analysis (MESA) for these analyses was investigated on the basis of the Rovelli-Vulpiani formula. The unbiased autocovariance estimation multiplied by different lag windows was introduced into this formula for a better estimation of the optimum filter length. The optimum filter length in theMESA is discussed on the basis of the model data similar to the observed data. The model data was disturbed by white and red noises with standard deviations greater than the average amplitude of an oscillation in the model. Each short periodical oscillation in pole coordinates was calculated by a properly, defined Ormsby band pass filter. Their sum creates a short periodical signal part which subtracted from smoothed pole coordinates diminishes their standard deviations and their autocovariance estimations.