Principal Component Analysis for stratigraphic imaging improvement and facies predictions

Summary Principal Component Analysis was used for stratigraphic imaging improvement and facies prediction using seismic data and multiple attribute volumes as input. It was determined that using a 3D moving time window operator it is possible to improve seismic resolution; specific horizontal and vertical dimensionalities for the operator will depend on the complexity of the area and quality of the input data; however, good results can be achieved using a 3x3 cell size and time window values ranging from one third up to three or four times the dominant period of the seismic wavelet. The calculated Principal Component volumes will usually have broader bandwidth and more uniform amplitude spectrum compared to the input seismic independently on the number of volumes used for the calculations. These results can be attributed to the fact that rotating the reference system compensates the effect of induced correlation generated by convolving the reflectivity series with the seismic wavelet.