Time lapse seismic is a proven technology for reservoir monitoring. It allows accessing useful information about the fluid movement, pressure changes and so on. This technology, however, has still some limitations in term of area of application. Shallow marine Middle East carbonate reservoirs are one of them.
In order to test this limit, ADMA-OPCO performed a 4D pilot (OBC/OBC) over a giant carbonate field, where the 2013 monitor has repeated the geometry of the 1994 base (parallel shooting). In this paper, we describe the challenges of the 4D processing of this data, the applied solutions and some results.
The water bottom over the pilot area ranges from 15 to 25 m. The main challenges of this 4D processing are:
Geometry errors in the legacy OBC data: they have negligible impact in 3D but they increase the 4D noise. Significant time was spent correcting these errors.
Noise contamination: the 2C component over the area is highly contaminated by Scholte waves and guided waves. Additionally, the geophone of the monitor data was highly impacted by current noise. It was necessary to bring the hydrophone and geophone to a satisfactory S/N before summation, in order to derive reliable operator for PZ summation. One important point for this 4D project was to favor the less statistical and less adaptive process.
Multiple attenuation: in 4D, the multiple can be highly non repeatable because of changing sea water conditions. The area is known for its severe contamination of short period multiples due to shallow-marine, hard sea bottom environment. After PZ summation, an optimized 3D deconvolution was performed.
4D time and amplitude destriping: it was important to remove the distortion due to different acquisition conditions, coupling etc.
The data integrity analysis and correction, the optimal denoise and demultiple, the careful attention paid to the acquisition-related distortion removal as well as adapted 4D QCS allowed to obtain an interpretable 4D signal which gives useful information about the reservoirs.
These successful results were made possible by a careful control of the noise and multiple attenuation, favoring the use of the deterministic and constrained processes over the adaptive processes even if these latter may be superior in 3D processing; and a robust S/N separation in the 4D crossequalisation/parallel processing were designed and applied. These elements enabled to get an interpretable 4D signal under the very challenging conditions of the considered area and reservoirs.