HR: 10:50h
AN: S31H-03    [PDF]
TI: Surface-Wave Test of the Upper Mantle Low Velocity Zone
AU: * Pontevivo, A
EM: antonella@geo.geol.ku.dk
AF: Geological Inst., Univ. of Copenhagen Oster Voldgade 10, Copenhagen, DK 1350 Denmark
AU: Thybo, H
EM: thybo@geo.geol.ku.dk
AF: Geological Inst., Univ. of Copenhagen Oster Voldgade 10, Copenhagen, DK 1350 Denmark
AU: Panza, G F
EM: panza@dst.units.it
AF: Dpt. of Earth Sciences, Univ. of Trieste Via Weiss 4, Trieste, TS 34127 Italy
AB: Surface Waves hold the potential to detect the mantle low velocity zone (LVZ). We test the influence of a LVZ in the depth range of 100-180 km on the group and phase velocity dispersion curves of the fundamental and the first few higher modes as well as on the waveforms of both Rayleigh and Love waves. We invert calculated synthetic dispersion curves and real data from the Siberian Kraton area by the non-linear inversion Hedgehog method.\\ Our reference structure is the average 1D-model obtained from body wave tomography along profile Kraton. S-wave velocity is calculated for typical Vp/Vs ratios and density is estimated from the Nafe-Drake relation. We test several models: the reference model itself and the reference model with velocity reductions of 2% and 5% in the LVZ for different thicknesses of the LVZ. The group and phase velocity dispersion curves of the fundamental and higher modes are calculated for an elastic flat Earth model. Average errors ($\pm$ 0.06 and $\pm$ 0.04 km/s for group and phase velocity, respectively) are superimposed before the inversion. In order to check if the thickness of the LVZ can be resolved, synthetic seismograms are calculated for several depths of the sources, both for Rayleigh and Love waves, for the same source parameters and epicentral distances as for real events with travel paths across Siberia.\\ The differences between the seismograms are small, even when the source is located in or around the LVZ, which is the only situation where a LVZ can be detected. The LVZ cannot be identified in the ambient noise from the phase or group velocities of the fundamental mode. However, for the higher modes at short periods, the difference exceeds the observational error for a pronounced LVZ.\\ We use non-linear inversion to check the set of solutions for the dispersion curves of the different models. As expected, the data is more sensitive to Vs than Vp. The solution space requires a low velocity only when the LVZ is pronounced ($>$ 80 km (40 km) thick for a velocity reduction of 2% (5%)). However, the calculated thickness and velocity contrast are smaller than in the original model.
DE: 7218 Lithosphere and upper mantle
DE: 7255 Surface waves and free oscillations
SC: Seismology [S]
MN: 2003 Fall Meeting