Abstract
Processing of ocean-bottom seismic (OBS) data by deconvolution of up-down separated wavefields simplifies and improves 3D and 4D seismic processing workflows because it attenuates surface-related multiples, compensates for spatio-temporal variations of the water velocity, and performs source designature and deghosting. The assumptions of a 1D earth's interior, which are inherent in an implementation of the technology known as up/down deconvolution (UDD), limit its potential.
We show a successful application of multidimensional deconvolution (MDD), which is a generalization of UDD in complex media, to an OBS dataset acquired in the Caspian Sea that shows a rugose seabed varying from approximately 150 to 450 m with steep slopes. Moreover, the overburden has geological features such as shallow gas, mud volcanoes and crestal faulting and requires a technology that does not make simplifying assumptions.