Ambient-noise cross-correlation is widely used to retrieve the Green function between two seismic stations. The Green functions of several body-wave phases between two near antipodal stations can be extracted by cross-correlating the coda waves of large earthquakes. In this study, we cross-correlate the data from Peking University temporary seismic arrays and China provincial seismic network in the Qinling and the data from the GSN (Global Seismographic Network) stations in South America to retrieve body-wave Green functions. The cross-correlation result clearly shows several body-wave phases (e.g., PKIKP, PP, PcPPKP, SKSP). Meanwhile, very similar Green-function waveforms can be retrieved by cross-correlating the synthetic seismograms generated from a 1D Earth model. This result suggests that free oscillations (normal modes) generated by large earthquakes are the most important energy source for the body-wave Green functions. The synthetic test also indicates that focal mechanism can influence the correlation result, e.g. the PKIKP phase cannot be observed if only strike-slip events are stacked. Our results suggest that by rejecting strike-slip events during the processing, we can increase the signal-to-noise ratio of the Green functions and reduce the computation time. Our study provides an example to demonstrate that the body-wave Green function may contain heterogeneity information of the Earth.
ASJC Scopus subject areas