Tiwary, Vladimir K. (Institute for Theoretical Geophysics, University of Oklahoma, Norman, OK 73019) | Chesnokov, Evgeni M. (Institute for Theoretical Geophysics, University of Oklahoma, Norman, OK 73019) | Bayuk, Irina O. (Institute of Physics of the Earth, Russian Academy of Sciences) | Ammerman, Mike (Institute of Physics of the Earth, Russian Academy of Sciences, Devon Energy Inc., Oklahoma City)
Linear prediction theory plays a pivotal role in seismic data processing. Time domain and
Several decimation tests were run on the Valhall Life of Field Seismic (LoFS) survey 1 (November 2003) and survey 2 (March 2004) to determine which decimations may be acceptable while still maintaining a good 4D signal quality. Three different analysis methods were studied: NRMS analysis, 4D resolution and cross-correlation. An automated process to measure 4D resolution was developed using statistical comparisons of 4D attributes. This 4D resolution method combines both 4D signal and background noise and provides a better measure of 4D signal quality than the NRMS method that only looks at background noise. Results from this study indicate that reducing the receiver sampling or reducing the shot effort, both have a large impact on the 4D resolution of the 4D signal. For a particular decimation factor, shot decimated datasets have better 4D quality signal than cable or node decimation datasets. This should be understood in the context that the survey design is already biased towards shots (shot/receiver ratio 20:1) and so a larger 4D signal quality degradation is expected by increasing the receiver distance further. Furthermore, node decimated datasets produce a better 4D quality signal than cable decimated datasets with the same decimation factor. These observations were corroborated by the cross-correlation method on the amplitude difference maps.