Feasibility analysis of searching for the Slichter triplet in superconducting gravimeter records

The search for the elusive Slichter triplet requires elaborate analysis of the elastic-gravi- tational mode characters and the non-stationary behavior of noisy time-series. A typical question is that it is difficult to characterize the excitations with attenuation by diffusion when their intensity is low compared to noise. Thus the theory for deriving the modes＇ frequencies is still controversial, and various scholars tried to search for the Slichter triplet in superconducting gravimeter （SG） records, but failed. One of the main causes might be due to the inappropriate use of datasets. We present in this paper synthetic experiments on the selection of record length, sampling rate and number of SG records under the Global Geodynamics Project （GGP） to detect the damped harmonic signals hidden in noises based on the optimal sequence estimation （OSE） method. Moreover, our results show that the existing observation conditions arouse restrictions and it might be impossible to detect the Slichter triplet excited by single excitation source based on Fourier spectrum analysis. Thus we suggest a stacking way of combining several seismic events in the case that the excitation mechanism has so far been unclear.

Detection of the Inner Core Translational Triplet Using Superconducting Gravimetric Observations

Based upon seven superconducting gravimeter （SG） records of 20 000 h length after the 2004 Sumatra earthquake, four methods, namely the ensemble empirical mode decomposition （EEMD）, the multi-station experiment （MSE） technique, the autoregressive （AR） method and the product spec- trum analysis （PSA） method, are chosen jointly together to detect the inner core translational modes （1S1）. After the conventional pretreatment, each of the seven simultaneous residual gravity series is di- vided into five segments with an 80% overlap, and then EEMD is applied to all the 35 residual SG se- ries as a dyadic filter bank to get 35 filtered series. After then, according to different stations and dif- ferent time windows, five new simultaneous gravity datasets are obtained. After using MSE for each of the five new datasets, the AR method is used to demodulate some known harmonic signals from the new sequences that obtained by using MSE, and three demodulated product spectra are obtained. Then, according to two criterions, two clear spectral peaks at periods of 4.548 9±2.3×10^-5 and 3.802 3±3.2×10^-5 h corresponding respectively to the singlets m=-1 and m=＋l are identified from various spectral peaks, and they are close to the predictions of the 1066A model given by Rieutord （2002）, but no spectral peak corresponding to the singlet m=0 is found. We conclude that the selected two peaks might be the ob- served singlets of the Slichter triplet.