Study of the background free oscillations and polar tide based on EEMD

A nonlinear and non-stationary time series analysis method, the ensemble empirical mode decomposition method (EEMD), is used to process superconducting gravimeter (SG) records to both extract the background earth free oscillations and estimate the tidal factors of the Chandler and annual wobbles. It decomposes gravity data into a finite number of intrinsic mode functions (IMFs), one or the combinations of which will correspond to geodynamic phenomena we are interested in. After removing local synthetic tides, local atmospheric pressure effects (time- and frequency-dependent atmospheric admittances derived from wavelet filter analysis) and instrumental drift from continuous gravity data (1997-2000) recorded at Syowa station, we decompose the residuals into a series of IMFs, among which the combination of the IMF1-5 contains the full information about the background free oscillations and especially the ₀S₂ mode is exactly extracted to the IMF5. That the EEMD method perfectly filters out all the noise outside the free oscillation frequency band and extracts the background free oscillations in a more narrow frequency band can be useful for the further detection of the source of this background excitation. In addition, the polar tide is also extracted (i.e. IMF6-7) based on the decomposition of the gravity residuals (with local synthetic tides subtracted, more than 5000 days) at Wuhan station, and then the tidal factors of the Chandler and annual wobbles are estimated by fitting a combination of two sinusoidal functions. The results about the polar tide are very close to that based on the previous proposed wavelet approach, which further confirms the correctness of our analysis based on EEMD.