Simultaneous Determination of Average Thickness and P-wave Speed of the Crust by Virtual Deep Seismic Sounding (VDSS)

Dou Kang1, Chunquan Yu2, Jieyuan Ning1, Kai TAO3 and Wang-Ping Chen4, (1)School of Earth and Space Sciences, Peking University, Beijing, China, (2)Department of Earth, Atmospheric and Planetary Sciences, MIT, Cambridge, MA, United States, (3)State Key Laboratory of Petroleum Resource and Prospecting, and Unconventional Natural Gas Institute, China University of Petroleum at Beijing, Beijing, China, (4)Now at Ocean College, Zhejiang University, Hangzhou, China
Abstract:
Using teleseismic S-waves, VDSS treats the SV-to-P conversion under the free surface (on the station-side) as a virtual source to generate strong, post-critical reflection off the Moho (SsPmp phase). With just a single, good-quality earthquake, arrival-time difference between SsPmp and the direct S-phase (TSsPmp-Ss) can effectively determine the crustal thickness (H) near the receiver. However, there is a strong trade-off between H and P-wave speed (Vp) in the crust.

Here we extend VDSS to constrain both H and Vp by taking advantage of the variation in ray-parameters, or incident angles, as a function of epicentral distance. Note that in conventional receiver functions, information contained in data of different ray-parameters is usually lost, because stacking over move-out corrected data is required to get a clear signal.

At a given station, we collect data from many events, each with a different ray-parameter of the direct S-phase (ps­). For each event, we 1) estimate the source wavelet of the direct S-wave through particle motion analysis; 2) deconvolve this wavelet from the vertical- and radial-component seismograms (Yu et al., GJI, 2013); and then 3) determine TSsPmp-Ss through waveform modeling. Finally, we analyze data pairs (ps2, T2SsPmp-Ss) to find the best-fitting values of H and Vp.

Synthetic tests verify the robustness of the method even with 15% of white noise. Moreover, we applied the method to public domain data from Forrest (FORT), located in the Eucla basin of western Australia. Based on 30 earthquakes from a narrow back-azimuth range (105±15°) but with ps changing from 0.1221 to 0.1349 s/km, we estimate that near FORT, H and Vp are about 44±2 km and 6.67±0.35 km/s, respectively. This crustal thickness is consistent with previous reports – a surprisingly high value for a region where the elevation is less than 200 m. Together with the high Vp, our results imply that the crust has a dense, mafic component.

Back to: Imaging the Earth III Passive and Active Source Studies of the Crust Posters