From seismic data to core data: an integrated approach to enhance reservoir characterization

Integrated structural analyses of seismic and various well data are necessary to optimize hydrocarbon reservoir characterization. However, there are many published examples from the oil and gas industry where single data types are analysed but not integrated. This may lead to erroneous interpretations and drainage strategies. As illustrated by an example from the area around well 34/10-B-12 in the North Sea Gullfaks Field, integrated structural interpretation should typically utilize all available seismic surveys, well log correlation data, dipmeter data and core data. Interpretation of seismic data helps in the understanding of large-scale structural and stratigraphic geometries. Time-lapse (4D) seismic helps to identify changes in reservoir properties caused by injection and production. Well log correlation data are used to document variations in zonation thickness caused by sedimentological or structural changes. Dipmeter data tie observations of bedding orientation from seismic data to subseismic scale. Core data represent the most detailed (millimetre to metre scale) data available and can yield information on rock properties as well as sedimentological and (micro)structural features. Small-scale deformation structures such as deformation bands and fractures can typically be identified and characterized. In addition, it is possible from unorientated cores to find the orientation of bedding and deformation structures. This information is compared to observations from dipmeter data, well log correlation data and seismic data to improve the interpretation. Well 34/10-B-12 is a hanging wall injector near one of the large-scale faults in the Gullfaks Field. Several 3D seismic surveys are available from the area, as are standard well log data, dipmeter information and cores. Together, the data range from millimetre to kilometre with some overlap between the data types. Through integrated analysis, pitfalls such as interpreting any linear feature on timedip attribute maps as faults has been avoided. Also, a geometric relation between core-scale and seismic-scale faults has been established, and it has been possible to relate small-scale and large-scale structures in a model which is consistent with all the available data.