Land surface-related multiple attenuation based on wave-equation deconvolution

Free-surface multiple prediction of land data is challenging due to the unknown downward reflecting free-surface combined with poorly sampled recording of the shallow multiple generators. We use a wave-equation deconvolution approach to derive an image of the shallow multiple generators using the multiple periodicity. The image is then used to model multiples which are subsequently subtracted from the input data. The approach adjusts to variations in the shallow subsurface and the multiple model requires minimal adaptation.