3D PP/PS prestack depth migration on the Volve field

Abstract

Ocean-bottom seismic (OBS) methods record both compressional (PP) and converted shearwave (PS) data. PS data have successfully been used to image through gas clouds (XiangYang Li et al., 2001) and image reflectors that are weak on PP data ( MacLeod et al., 1999), and can help constrain rock property estimates (Ozdemir et al., 2001). Thus, in principle, the use of PS data can significantly mitigate risk in hydrocarbon exploration and production. However, imaging of PS data is reliant upon accurate compressional and shear velocity estimates, as both the moveout and offset of the imaged point depend upon these parameters. Velocity anisotropy is also a critical factor. The previously used processing flows have not adequately comprehended this, and the results have often been of poor quality and difficult to tie with the PP volumes. This can preclude their use for lithology prediction. We have implemented an imaging workflow that can resolve many of these issues, and here describe its application to OBS data from the Volve field, Norwegian Central North Sea. Imaging workflow The current workflow assumes polar anisotropy, in which case the moveout on PS prestack gathers depends upon compressional velocity (Vp), shear velocity (Vs) and Thomsen’s polar anisotropy parameters epsilon and delta. We build a layer-based depth model that contains these parameters and use this for 3D prestack depth imaging of both the PP and PS volumes. A top-down layer-stripping approach is used to create the velocity model, as shown schematically in Figure 1. For the current layer, Vp, epsilon and delta are updated from the PP data, constrained by depths, sonic logs, and checkshot information from wells, where available. The PS data are then used to derive Vs and to update epsilon and delta as necessary. Updates to Vs are constrained by the depth tie of the PP and PS volumes, ensuring consistency between the two. A variety of model update algorithms can be incorporated into the workflow, including common image point (CIP) tomography, layer-based tomography, and parameter scanning. This top-down approach means that the PS data for the next iteration can be quite accurately imaged using the current model and an assumed Vp/Vs ratio for the next layer, which greatly eases correlation of PP and PS reflectors.