USE OF SPECTRAL DECOMPOSITION AND OTHER SEISMIC ATTRIBUTES TO PREDICT SAND DISTRIBUTION IN SOUTHERN PATTANI BASIN, THAILAND

Abstract

The study area has reservoirs that are thickening and thinning along pay zones and restricted in lateral distribution. Therefore it is necessary to have a better understanding of the distribution of sand to optimize the hydrocarbon recovery in this area.  The objective of this study is to predict the sand distribution by using spectral decomposition and other seismic attributes such as Structurally-Oriented Filtering (SOF) and Similarity to have a better prediction of the hydrocarbon zones. Spectral decomposition techniques typically generate a continuous volume of instantaneous spectral attributes from broadband seismic data, to provide useful information for reservoir characterization and direct hydrocarbon detection (Partyka et al., 1999; Castagna et al., 2003; Liu and Marfuit, 2007). The filters in a filter bank for spectral decomposition are usually Gabor (a linear scale) or Morlet (an octave scale) wavelets, which have the property of minimum uncertainty. The property of the octave scale is to have the higher central frequency with a higher bandwidth. The linear scale has every central frequency with the same bandwidth. By using multiple frequency bands and comparing the result between octave and linear scale there may be additional insights about the thickness of the sand. Based on the amplitude characteristics of both sands, the results showed that the octave scale can be used to identify the sand distribution, but in order to identify the thickness of the sand and also distinguished between the two thin sands, the linear scale is the best method. Horizon slices, extracted from high-frequency volumes of the spectral decomposition show the spatial distribution of hydrocarbon which matches with existing well data. Hence, this technique is useful for identifying the sand distribution, sand thicknesses and hydrocarbon occurrence.