Amplitude Anomalies of S Waves Caused by Low Shear Velocity Structures at the Base of the Mantle

Abstract:

Previous studies have shown that the direct S and Sdiff waveforms of earthquakes in Papua New Guinea region recorded by seismographs in Northern America are distorted due to sampling slow shear velocity anomalies at the base of the mantle. The emergence of postcursours to the S/Sdiff waves and the travel time anomalies have been reasonably explained by placing a ultra low velocity zone (ULVZ) in southwest of Hawaii. In this study, we focused on the amplitude anomalies of the S/Sdiff waveforms. The direct S phase show very low amplitude at stations in Southern California, at the distance and azimuth around 90 and 55 degrees from the earthquake. The amplitude is as low as 10% of the synthetic amplitude of a standard 1D model, especially at higher frequency range above 0.025 Hz. We first checked and confirmed that the anomalies are not due to errors in the focal mechanism, which is used to calculate the reference synthetic waveforms. Also we checked that the amplitude anomalies are unlikely to be caused by the structures near the earthquake or near the stations, by looking at the amplitude of the depth phases or waveforms of other earthquakes. We assumed that the anomalies are produced by the focusing and defocusing effect of sampling 3D heterogeneous at the base of the mantle, and searched for the causal structures. Full 3D synthetic waveforms are calculated down to 8 seconds for tens of structural models with slow anomalies of different size and velocity reduction placed on the core-mantle boundary (CMB). The result shows that existing tomographic models do not fully explain the observed amplitude anomalies. Stronger shear velocity anomalies are required. The previously proposed thin large ULVZ placed on the CMB southwest of Hawaii partly explains the observed amplitude reduction, even at the distance as short as 90 degrees from the earthquake. This result indicates the significance of finite frequency effect of the ULVZ structure to the S waves, since the ray theoretical bottoming point of the phase of the interest is predicted to be more than 200km above the ULVZ. The observed S wave amplitudes are better explained by lowering the shear velocity of the slow regions, which are imaged in tomographic models, close to the event region in Western Pacific at the base of the mantle.