Fault imaging of Hibernia three-dimensional seismic data using edge detection and coherency measures

Abstract

Seismic methods play a key role in exploration and reservoir characterization of the offshore Newfoundland Hibernia oil field. Seismic data is traditionally processed to image continuous reflections rather than to image discontinuities such as faults. Thus, when interpreting such 3D seismic data it is often difficult to obtain a clear and unbiased view of faults. Due to the complex structure of offshore Newfoundland fields, fault imaging is extremely important since field production may be affected by sealing faults, and may effect hydrocarbon production and development strategies. -- In recent years, companies have produced algorithms for enhanced fault detection. The edge detection differencing method, as presented by Lou et al., (1996), measures changes in the subsurface such as faults using differencing of seismic traces. The C1 coherency algorithm, as described by Marfurt et al., (1998), uses cross-correlation between seismic traces. Another method used for fault detection is second derivative maps which take the second derivative of the seismic traces. -- The main objective of the this study is to compare different fault detection algorithms using both synthetic (Kelly, 1998) and real seismic data from the Hibernia fields (1991 and 1997 surveys). For each data set, a representative of the data volume containing the Murre fault was chosen for comparison purposes. Results indicating that the coherency method produces best results with the second derivative produce good results for the model data. The use of EDGE and Poststack/PAL software proved to image more subtle faulting than large scale faulting such as the Murre fault.