A multi-stage inversion method for simultaneous source deblending of field data

Simultaneous source technology has significant impacts on modern seismic acquisition. It allows us to acquire seismic data in a much more efficient way and can play a key role in areas where the acquisition window is rather short, such as the northern seas. A wide range of methods—from processing type to inversion type—have been proposed and studied in order to handle the source interference. Many of those work well on synthetic data but may suffer on field data. Here we propose a multi-stage inversion method, which combines traditional seismic processing with sparse inversion and provides a more target-oriented approach. Each stage of this method focuses on a particular portion of energy in the blended record and separates that from the rest. This method also has a simple and flexible framework, which could potentially combine with other state-of-the-art deblending algorithms. A field trial in the North Sea shows this hybrid approach to be highly successful.