Next: Discussion Up: Prucha and Biondi: STANFORD Previous: PS-DMO in the frequency-wavenumber

Real data

In order to further test our operator, we apply it to a real data line. The Alba Oil Field is located in the UK North Sea and elongates along a NW-SE axis. The oil reservoir is 9km long, 1.5km wide, and up to 90m thick at a depth of 1,900m sub sea (, ).

The producing reservoir consists of unconsolidated high-porosity turbiditic sandstone of Eocene age. The oil-bearing reservoir sand and the overlying shale have a very low P-wave acoustic impedance contrast but make a significant S-wave velocity contrast. Hence, the reservoir delineation using only normal PP-seismic data is very difficult. Therefore, the ocean bottom PS-data was acquired ().

Even though the Alba field is a 3D data set, we select one line by choosing a source line that overlaps a receiver line. There are several gaps in our selected data since we are taking a 2D line from a 3D data set.

Figure shows a CMP gather from the PZ-component and the PS-component of the data. Note the gaps and irregularities in the data.

Figure shows a comparison between the NMO stack and the DMO stack of the PP line. The DMO was performed by the same algorithm described here, specifying 65#65.

Figure shows the PS result. It compares the PS-NMO CCP binning stack and the PS-DMO stack. The PS-DMO stack was obtained using the filter described in equation (). Observe that, even though it is a flat events area, those small dips are better defined after doing PS-DMO.

Figure shows the PS-DMO result using the filter described in equation (), which better handles the amplitudes. Note that some strong dip energy appears. The rest of the section remains the same. Therefore, the filter in equation () produces more accurate results.