摘要

This paper presents an effective decomposition method for multicomponent seismic wavefields. This method assumes that the changes of amplitude and phase for the seismic wavefields in the adjacent traces are small. In this case the vertical and the horizontal components for both longitudinal and shear waves are determined by their different polarization and slowness during propagation. Under the least square constraint, the polarization and the slowness for each wave can be estimated from the input multicomponent data. Therefore, the longitudinal and the shear waves can be reconstructed and then separated. This method is accurate. Not only the multicomoponent wavefields of primaries can be decomposed effectively with this method, but also the longitudinal and converted waves which are reflected many times can be decomposed as well. Straigraphical information such as velocities is not needed in this method and therefore effect of artificial errors is reduced. The example of synthetic data confirms that this method is feasible and correct.