Fracture Analysis for Carbonate Reservoirs Using 3D Seismic P-Wave Data: Middle East Case Study

Seismic reflection amplitude at the top of a fractured rock layer varies as a function of both offset and azimuth. Travel time and attenuation along ray paths through a fractured layer also vary with offset and azimuth. An integrated set of procedures and a flexible workflow were developed to extract these measures of azimuthal seismic anisotropy. The procedures include screening tools to evaluate the feasibility of applying seismic inversion methods to extract azimuthal attributes, and quick-look tools to rapidly visualize seismic anisotropy that may indicate open fractures. Azimuthal anisotropy is recognized in seismic data from a giant Middle East oil field by identifying differences between sector stacks formed from different azimuth bins. This screening analysis shows an azimuthal AVO response that seems to vary in a meaningful way across the field. Further analysis utilized a proprietary, azimuthal AVO, inversion process that estimates two attributes. The first quantifies AVO anisotropy that may be related to fracture intensity, differential stress, significant stratal dips, or other factors. The second attribute is related to fracture orientation when fractures are the cause of anomalies. A goodness-of-fit estimate quantifies the error between the fitting function and the observed data. Inversion results show areas with strong azimuthal variations that correspond well with coherent energy observed in the sector stack difference, but with improved spatial resolution and contrast. The key to using azimuthal seismic anomalies wisely in reservoir development is their integration with subsurface data to discriminate among possible sources of the anomalies. The validation phase of the study used subsurface data on fracture intensity, in-situ stress state, stratal geometry, erosion surfaces, and production data and tests. Maps of seismic anisotropy show coherent anomalies related to known geologic features, giving confidence that the method detects meaningful geologic information that can be used to constrain reservoir quality models.