Broadband Finite Frequency Ambient Noise Tomography

Ambient noise tomography (ANT) has become a popular method to study the crustal and uppermost mantle structure of the earth in recent years due to its exclusive capability to extract short-period surface wave signals. Most of ANT are based on ray theory that assumes interstation surface waves from ambient noise are mainly sensitive to a narrow zone alone the ray path from one station to the other. Recently, many studies have demonstrated that long-period Rayleigh wave signal with high SNR can be obtained from cross-correlation of ambient noise data and could be used to do long period surface ware tomography. In order to obtain accurate phase velocity maps using long period surface waves from ambient noise, frequency effects must be considered in tomography. In this study, we investigate the feasibility of finite frequency ANT by calculating 2-D phase sensitivity kernel based on Born approximation. In calculating 2D sensitivity kernels for empirical Green's functions extracted from cross-correlations between a pair of stations, one station is regarded as receiver and the other as virtual source. Based on the 2D finite frequency sensitivity kennels, we develop a finite frequency ambient noise tomography method to construct Rayleigh wave phase velocity maps. To demonstrate the feasibility of our developed method, we apply the method to empirical Green's functions extracted from cross-correlations of USArray noise data to construct phase velocity maps at 20-150 sec periods. Our resulting phase velocity maps are very similar to earthquake-based phase velocity maps with almost zero means and 20-30 m/s stand deviations of differences. Major tectonic features in USA are well revealed in our phase velocity maps.