Pseudo 2D Elastic Waveform Inversion for Velocities and Q Factors in the Near Surface

Abstract:

Seismic reflection wave is significantly affected by the complexity in the near surface area. Therefore, it is important for subsurface imaging to obtain the near surface information as much as possible in exploration seismology. Seismic attenuation, described by the quality or Q factor, has great effect on the seismic waveform. But it is rarely estimated for the near surface. It is very important to determine a reliable Q structure for accurate full waveform simulation.

 

Elastic full waveform data includes a variety of wave modes: P-waves, converted waves and surface waves. Early arrivals mostly contain information of V_P and Q_P. The S waves contain information of V_S and Q_S. We propose a method to determine velocities and Q factors simultaneously via joint early arrivals and surface wave inversions for V_P, Q_P, V_S and Q_S. Many methods for elastic full waveform inversion have been developed in both 2D and 3D. However, these methods need too much computation time when inverting massive data. We develop a technique to perform 2D structure inversion using 1D forward modeling which is named pseudo 2D elastic waveform inversion. This method can only deal with approximate 1D structure which may have small anomalies. However, it saves computational time compared to conventional 2D full waveform inversion. This inversion method can help image useful information about approximately layered near surface.

 

We generate some synthetic structures and do inversion tests. The inversion results are satisfying. Then we apply this method to real data from an oil field in China. The inversion results show us this area has strong attenuation in the near surface and low Q values have great influence on waveforms.