Seismic attributes in imaging challenging reservoirs: A case study of Boonsville Field, North Central Texas

In this study, we investigate the Caddo sequence from the Boonsville gas field in the Fort Worth Basin of North Central Texas. Two Middle Pennsylvanian thin reservoirs are closely separated by a thin Caddo limestone unit. Seismic attributes, namely, instantaneous frequency and amplitude attributes did predict the distribution of reservoir facies in some productive wells. Also, the attribute maps implied that the Caddo facies would exist at well locations where the reservoir was not encountered and were absent where their presence was confirmed from well control. To remove this ambiguity, more advanced techniques of filtering, tracking and information extraction have been invoked and integrated. The data were firstly subject to dip-steered filtering and dense tracking processes. Next, the cleaned data set was spectrally decomposed using a Time Frequency Continuous Wavelet Transform. This decomposition successfully resolved both reservoirs. However, some non-reservoir areas were characterized by frequency responses similar to those shown in reservoir areas. Spectral examination of individual traces from producing and non-producing areas inferred that producing zones are characterized by frequency features different than those of the nonproducing zones. Next, poststack seismic inversion was performed to incorporate well data with seismic data and to produce an acoustic impedance cube. Interestingly, the acoustic impedance sections also suggested that the productive sandstones are characterized by different and higher impedance character relative to limestone formations and their surroundings. This study demonstrates that incorporating information from different sources (amplitude, frequency, spectral decomposition, well data, etc.) can assist significantly in overcoming challenging formations in the subsurface.