A global model was constructed for shear wave velocity structure, expanded up to degree and order 36 in spherical harmonic expansion, and down in 500 km of the Earth's upper mantle. The data set included about 18,000 seismograms associated with 971 events with magnitude larger thanabout 5.5.
Assuming that the errors in data space and model space are both Gaussian random distributed, we used the Akaike Bayesian information criterion (ABIC) method and the Akaike information criterion (AIC) method to determine the optimal damping parameters and the optimal regularization of theunknown continuous model, respectively. Both the spherical harmonic expansion and the block parametrization approaches have been tested. The equal area block regularization approach gave a more accurate result than the spherical harmonic expansion approach, and it was adopted.
Age-Love wave phase velocity variations were retrieved for the Pacific, Atlantic and Indian Oceans, respectively. Simple forward modeling showed that seismic phase velocity variation with a continuous thickening of lithosphereup to about 150 Ma fits the present observation, disagreeing with the conventional model, which changes the age-depth curve at about 60-80 Ma. The age-Love wave phase velocity variations in different oceans showed systematicdifferences at younger ages, and convergence beyond 100 Ma. Age-seismic phase velocity relationships on each side of ridges were also examined and asymmetric velocity variations were found. The average age-phase velocityrelations were subtracted from Love wave phase velocity variation maps and "residual" maps were produced. The results indicated broad, low velocity regions in the south Pacific (super-swell region), the south and west IndianOcean, and high velocity regions east of the East Pacific Rise and in the north to northeast Indian Ocean.
In the S-wave velocity results under global mid-ocean ridges, low velocity anomalies were resolved but were limited in the 100 km below the surface. The minimum velocities were at depth about 50 km. The horizontal width of the low velocity anomaly, which crossed over ridges, increased with the spreading rate. The S-wave velocities under ridges were strongly correlated with spreading rates at shallow depth, but the correlation decreased for deeperresults and almost disappeared at 100 km. Several major hotspots were associated with low-velocity anomalies of about 1-2 percent and diameters larger than 1000 km. But the depths of the low velocity anomalies were between100-200 km, which were different from those of ridges and expressed the mechanism differences between ridges and hotspots. The velocity structures under the East African Rift Valley and the Baikal Rift Valley were inspected.An active mechanism was suggested with the East African Rift Valley, and a different mechanism was connected with the Baikal Rift Valley.
