Getting the Best out of Up/down Deconvolution for 3D and 4D Imaging

Sparse Ocean Bottom Node (OBN) surveys can be difficult to process for 4D purposes, due to the strongly aliased common shot domain. However, where the survey records pressure and particle velocity, options exist for up/down wavefield separation and then deconvolution. This method offers a lot in principle, including attenuation of free-surface multiples, and removal of the source wavelet. However, the method is notoriously difficult to perform in practice, and results are often poor. In this contribution we discuss practical data preparation steps that mitigate problems in the up/down deconvolution method. Using sparse OBN data from the North Sea we illustrate the quality of result that is possible to achieve with the method, and show how the result can be further utilized to attenuate free surface multiples with alternative methods that have reduced reliance on adaptive subtraction. Reduced need for adaptive subtraction is an attractive proposition that can help protect weak 4D signal. Even with optimised data preparation, the weakness of up/down deconvolution lies in an increased noise level in the output. The method is suitable for fast-track processing, however, producing results in this case that compare favourably with alternative workflows using target-generator multiple modelling and adaptive subtraction.