Prediction of length-of-day using extreme learning machine

Traditional artificial neural networks （ANN） such as back-propagation neural networks （BPNN） provide good predictions of length-of-day （LOD）. However, the determination of network topology is difficult and time consuming. Therefore, we propose a new type of neural network, extreme learning machine （ELM）, to improve the efficiency of LOD predictions. Earth orientation parameters （EOP） C04 time-series provides daily values from International Earth Rotation and Reference Systems Service （IERS）, which serves as our database. First, the known predictable effects that can be described by functional models-such as the effects of solid earth, ocean tides, or seasonal atmospheric variations--are removed a priori from the C04 time-series. Only the residuals after the subtraction of a priori model from the observed LOD data （i.e., the irregular and quasi-periodic variations） are employed for training and predictions. The predicted LOD is the sum of a prior extrapolation model and the ELM predictions of the residuals. Different input patterns are discussed and compared to optimize the network solution. The prediction results are analyzed and compared with those obtained by other machine learning-based prediction methods, including BPNN, generalization regression neural networks （GRNN）, and adaptive network-based fuzzy inference systems （ANFIS）. It is shown that while achieving similar prediction accuracy, the developed method uses much less training time than other methods. Furthermore, to conduct a direct comparison with the existing prediction tech- niques, the mean-absolute-error （MAE） from the proposed method is compared with that from the EOP prediction comparison campaign （EOP PCC）. The results indicate that the accuracy of the proposed method is comparable with that of the former techniques. The implementation of the proposed method is simple.

New evidence for possible impact of solar activity on long-term fluctuation of the earth rotation

The long-term fluctuation of the Schwabe period (LSP) of sunspots number (SSN) has been found to have high correlation with the variation of the length-of-day (LOD) in low frequency by using the data of smoothed monthly mean SSN during 1818-1999 and the method of wavelet transform. Analyses indicate that the maximum correlation coefficient between the series of LSP and LOD during 1892-1997 is about 0.9, with a time lag of about 5 years for the LOD related to the LSP. Though the maximum correlation coefficients between the LSP and the other two LOD series (1818-1997) reduce to about 0.4, they remain over the thresholds of 95% confidence level. This suggests new evidence for possible impact of solar activity on the long-term fluctuation of the earth rotation.

Secular variations in geomagnetic Field caused by the fluctuations in the fluid flow in the outer-core

The theoretical spatial power spectrum of a dipole located at （ r0, θ0, φ0） can be fitted by a straight line in logarithmic scale when n is larger than 2. Based on the spherical harmonic coefficients of geomagnetic field during 1900—1995, the depth（r0） of source-layer of every 5a is calculated. The results show that r0 decreased from 1900 to 1960; abruptly changed from 1945 to 1950 related to some kind of disturbance; decreased again from 1960 to 1975; increased from 1975 to 1985; and kept stable after 1985. Then the mean energy density （MED） of each year is induced to its corresponding r0. We find that MED of dipole field kept nearly unchanged from 1900 to 1960. While, MED of non-dipole field increased. The change of r0 coinciding with the geomagnetic secular variation, impulse and length-of-day happened around 1970, suggesting that the change of r0 may be related to the impulse. The variation in the fluid flow in the outer-core caused by the core-mantle coupling is a plausible candidate