MPI Multiple Attenuation Method in the CMP Domain

The multiple prediction through inversion (MPI) method is a fully data driven multiple attenuation technique and has been implemented in the shot gathers. In this paper, we attempt to extend the MPI multiple attenuation method from the shot domain to the common mid-point (CMP) domain, so that it can be used to data sets that have poor signal-to-noise ratio in the shot gathers. For the CMP domain version, although it is still a pre-stack application, implemented on CMP gathers instead of shot gathers, the pre-stack multiple prediction process can be carried out on individual CMP gathers by auto-convolution in the frequency-space domain or multiplication in the frequency-wavenumber domain. Real data examples demonstrate the efficiency of this CMP domain MPI method, when the approximation of a low-relief earth is valid and the earth is laterally homogeneous.