Publications

Thin layer hydrocarbon bearing reservoirs with thicknesses below seismic resolution are often encountered while drilling. To complicate matters further, seismic signal strength decays with depth and data often have significant levels of random noise. In this study, the authors have attempted to enhance vertical seismic resolution using a combination of a ‘blueing’ method and a wavelet transform method. The ‘blueing’ method gives weight to higher frequencies thus improving resolution. The method provides a single filter operator applied to the entire seismic volume in the time domain. The seismic image will become crisper, but the vertical resolution is still constrained by the bandwidth of the recorded data. A second method, the Stationary Wavelet Transform (SWT), is applied to increase resolution on a more targeted volume. SWT decomposes signal into sub-bands of frequencies with each sub-band containing redundancy from the adjacent lower sub-band. The seismic signal has an inherent finite bandwidth, meaning that it always has a smooth transition in time and frequency (it cannot have an abrupt change of amplitude in time and frequency). Unfortunately, due to noise interfering with the seismic signal combined with the band limited source (deficient in high frequencies), sudden changes in frequency amplitude can occur. We attempted to recreate some missing higher frequencies using redundancy from lower frequencies sub-bands. Using the redundancy characteristics of SWT, we can predict higher frequencies by extrapolation (leaking method). Application of this method to seismic data around “well-X” improved vertical resolution revealing some thin layers, which were validated by correlation to well-log data.