ALBERT

Description: Short-period, three-component recordings of the seismic wave field of peaceful nuclear explosions recorded on deep seismic sounding profiles (Quartz and Ruby-I, collected in 1984 and 1988, respectively) in northern Russia are used to constrain the nature of the high-frequency teleseismic Pn phase, which can be observed for receiver distance of 〉3000 km. We suggest that this phase is caused by velocity fluctuations in the upper mantle acting as scatterers. To test this hypothesis and to determine the properties of the upper mantles scatterers, the elastic reflectivity method (one-dimensional isotropic models) was used to model the coda of the high-frequency teleseismic Pn phase. Both the observed data and the synthetic record sections were analyzed by examining the coda decay rates of the teleseismic Pn phase at 5 Hz. Synthetic seismograms are computed for different models based on the global model IASP91 with an added zone of randomly distributed velocity fluctuations (lamellae) just below the crust-mantle boundary. These models may be characterized by the thickness of the scattering layer L, the vertical heterogeneity correlation length a, and the heterogeneity standard deviation sigma. The numerical simulation of the wave propagation in these models generated a high-frequency teleseismic Pn. Comparing the coda decay rates of synthetic seismograms with the observation, we tried to constrain the properties of the velocity model containing fluctuations in the upper mantle, L, a and sigma. In the distance range of the uninterfered high-frequency teleseismic Pn phase beyond 1300 km, the coda decay rates for our best fitting model are similar to the observed ones. This model has a 75 km thick zone of scatterers below the Moho, containing lamellae with an average thickness of 2 km and a RMS velocity perturbation of 5%. The modeling results show also that two other possible models, the whispering gallery phase along the crust-mantle boundary and scattering in the lower crust, taken alone are not able to explain the coda properties of the high-frequency teleseismic Pn phase.