In pure carbonate systems, the combined effect of variations in depositional facies and diagenetic alterations plays a key role in controlling variations in sonic velocities and thus in acoustic impedance. As seismic facies are delineated by acoustic impedance contrasts, depositional facies and geometries may be rather poorly defined for various carbonates environments (e.g., shallow-water platform carbonates). Accurate three-dimensional imaging of seismic facies and geometries is critical to construct a realistic, seismically constrained reservoir model. Conventional two- and three-dimensional (2-D and 3-D, respectively) seismic mapping is not an ideal predictive method when attempting to characterize carbonate reservoirs mainly because of the complexity and heterogeneity of carbonate systems.

Three-dimensional image-processing techniques of stacked and migrated data incorporate all three dimensions, which when combined help to identify and highlight events of significance in the data. The result is an attribute cube or volume that can be analyzed and interpreted more objectively by the interpreter than the conventional horizon-based interpretation. We have applied various 3-D image-processing techniques to produce filtered seismic reflectivity data and volume attributes to better visualize and delineate seismic facies, geometries, and the structure of heterogeneous carbonate reservoirs. Structure-oriented filtering was applied to improve signal-to-noise ratios and suppress random noise to obtain a better reflection definition. Combined volume dip and azimuth was calculated from the seismic cubes to detect subtle stratigraphic features, such as low-angle progradation units and shoal-type mounded seismic facies in the Permian Khuff and Upper Cretaceous Natih E reservoirs in Oman. Semblance volumes were used to highlight reflection terminations and distinguish between stratigraphic and structural features. In the Malampaya field (Philippines), neural network classification mapping was applied to the 3-D attribute-generated volumes to extract different seismic facies and properties, which can be related to potential good reservoir zones. Three-dimensional visualization tools were used to image both horizons and faults of a complex inverted structure of a deep Upper Cretaceous restricted marine-lacustrine carbonate reservoir in the Yacoraite Formation, northwest Argentina.

Seismic facies and geometries interpreted from the attribute analyses, combined with interpretation of the original seismic and core-log data, allowed us to construct robust structural and depositional models of carbonate environments that were used as input for static reservoir models.