Supervirtual Refraction Interferometry Based on Stacking of Neighboring Virtual-Traces and Its Application to Enhancing Wide-Angle OBS Refraction Waves

The first-arrival of far-offset refraction waves of wide-angle seismic data carries the information of structure in the deep subsurface. Enhancing the signal-to-noise ratio (SNR) of these waves can effectively improve the accuracy of first-arrival picking up, and would be very helpful for the tomographic velocity modeling of deep structures. The Supervirtual Refraction Interferometry (SRI) method can enhance refraction waves based on the principle of interference. It works well when the shots and receivers are densely distributed. However, when the SRI method is applied to the wide-angle Ocean Bottom Seismometer (OBS) observation with large station space and under high environmental noise, it cannot sufficiently enhance the refraction waves and could create artifacts in resulted waveforms, leading to low SNR of refraction waves. To cope with this problem, we propose a supervirtual refraction interferometry based on the stacking of neighboring virtual-traces (SRI-SNV). By stacking neighboring virtual-traces, our method can stably increase the accuracy of cross-correlation between the far- and near-offset traces, and achieve the high SNR of refraction waves. The synthetic tests and field-data applications both indicate that SRI-SNV can accurately construct virtual-traces under the situation of large station spacing and low SNR, and the resulted refraction waveforms have better continuity and higher SNR than the SRI method. Using the waveforms enhanced by the SRI-SNV, the first-arrival picking up can achieve higher accuracy than using the SRI method. The SRI-SNV method can also be used to enhance the seismic refraction waves of the wide-angle seismic data on land, which has large shot interval.