Developing a specific data processing flow in areas of shallow carbonates, as in other seismically difficult areas, requires a clear understanding of the problems associated with such environments. Once the initial problems are defined, a specific processing procedure is designed to achieve the objectives of seismic data processing. This design properly starts from the end, at the final required results, and proceeds backward to identify the processing tools needed to achieve these results. A number of severe data problems are associated with the shallow carbonate environment, the most obvious one being the strong multiples generated by the surface and carbonate interfaces. Water-trapped multiples propagate with water velocity, hence they are more easily discernible (due to their NMO differences) from the deeper, higher velocity primary reflections. The surface-connected and interbed multiples generated by carbonate interfaces, however, propagate with higher velocities, hence they are not as easily identified as the low-velocity multiples. As a result, a variety of multiple suppressing methods must be included in the processing stream.
Carbonate layers also cause the generation of horizontally traveling waves, such as the refracted events on marine data and large-amplitude ground roll on land data. These must be suppressed before proceeding with processes such as deconvolutions and wavelet processing. On land data, carbonates are normally covered by weathered low-velocity layers, and in some areas, carbonate layers are cut by old riverbeds. These conditions create short-wavelength disturbances in travel-times and thus adversely affect the imaging process. The processing flow must therefore include computation and removal of these time disturbances as well. Large-velocity differences between carbonates and clastic rocks cause most of the pressure wave energy to be reflected, with little transmitted to the layers below. This results in low signal-to-noise ratios in the primary reflections below carbonate layers. The velocity difference also causes the transversely isotropic layer velocity anisotropy, which in turn affects the apparent NMO velocities of longer offset data.