摘要Important information about small-scale heterogeneities is hardly accessible by the traditional deterministic seismic tomography. Fluctuations of the phase and the logarithmic amplitude of direct teleseismic plane P waves can be used to characterize small-scale heterogeneities. Seismic data recorded by the Hi-CLIMB array are used to analyze the average power spectrum of the small-scale velocity heterogeneities under Tibet. Coherence functions of the logarithmic amplitude and the phase fluctuations due to different earthquakes from different back azimuths show consistent characteristics, indicating that fluctuations are due to heterogeneities under the stations. Assuming that the heterogeneities are statistically stationary and distributed within a layer, we invert for the average heterogeneity spectrum by fitting both the logarithmic amplitude and the phase coherence data. Multiscale nature of the heterogeneity is evident. The inverted power spectrum is “red” at the large-scale end, meaning that the power spectrum decreases as the length scale decreases. Such a decreasing trend stops at smaller scales ~20–50 km and 10 km. This may indicate that mantle convection is not effective in destroying smaller heterogeneities.
Abstract：Important information about small-scale heterogeneities is hardly accessible by the traditional deterministic seismic tomography. Fluctuations of the phase and the logarithmic amplitude of direct teleseismic plane P waves can be used to characterize small-scale heterogeneities. Seismic data recorded by the Hi-CLIMB array are used to analyze the average power spectrum of the small-scale velocity heterogeneities under Tibet. Coherence functions of the logarithmic amplitude and the phase fluctuations due to different earthquakes from different back azimuths show consistent characteristics, indicating that fluctuations are due to heterogeneities under the stations. Assuming that the heterogeneities are statistically stationary and distributed within a layer, we invert for the average heterogeneity spectrum by fitting both the logarithmic amplitude and the phase coherence data. Multiscale nature of the heterogeneity is evident. The inverted power spectrum is “red” at the large-scale end, meaning that the power spectrum decreases as the length scale decreases. Such a decreasing trend stops at smaller scales ~20–50 km and 10 km. This may indicate that mantle convection is not effective in destroying smaller heterogeneities.
