A comparative analysis of Vondrak and Gaussian smoothing techniques

Summary Smoothing time series of measurements is usually employed to remove high frequency noise when it is necessary to interpolate them for dates where no measurements are available, and also in an attempt to free the series from their short-term instability in order to provide an improved representation of the measured quantity. Two methods are currently used in connection with the Earth's rotation studies: Vondrak's algorithm and a moving average with a Gaussian weighting function Gaussian smoothing). This study describes their behaviour in terms of frequency filtering and of the response in the time domain. Examples of their application to the Earth Rotation Parameters are given.