Comparison of seismic interferometry for surface wave extraction from traffic noise

This study compares and analyzes four seismic interferometry methods to extract empirical Green's functions of surface waves from traffic noise generated by vehicles traveling on roads. The cross-correlation, deconvolution, and cross-coherency methods were evaluated using synthetic seismic records and field data. All methods effectively extracted surface waves from short-duration seismic records. Pre-whitened cross-correlation and cross-coherency methods provided clearer empirical Green's functions than raw cross-correlation. The deconvolution method exhibited non-physical crosstalk relatively. Dispersion analysis confirmed the fundamental and higher modes of surface waves above 10 Hz, corresponding to shear wave velocities of 400 m/s and 700 m/s in the numerical model. Field experiments showed consistent phase velocities in the range of 200- 300 m/s, reflecting the shear wave velocity of the surficial layers in the study area. The findings confirm that seismic interferometry using traffic noise is effective for determining the shear wave velocity structure of shallow subsurface layers.