Publications & Presentations


Enhancement of the Three-Dimensional Seismic Method Through the use of Well Control and Seismic-Modeling Techniques in the Resolution of Fourth-Order Depositional Sequences within the Grayburg Reservoir, Foster Field, Ector County, Texas; Part 1

Document Type


Publication Date



The study area is located in sections 31 and 32 of T&PRR block 42, Township 2S, and sections 35,36,37 and 38 of T&PRR block 43, which is about 3 miles southwest of downtown Odessa. Pre-existing well data shows the Grayburg Formation can be divided into fourth-order shallowing upward sequences, that are 60 to 130 ft (18.3 to 39.6 m) thick. Previous work has also shown the Grayburg dips very gently to the east. The spatial distribution of flow parameters important in reservoir characterization is controlled, in part, by fourth-order depositional sequences or sets of sequences forming the reservoir. Within the study area, the Grayburg Formation occurs at a depth of roughly 4000 feet, and is composed of dolomitic limestones with a sonic velocity of roughly 17,000 feet/second. Seismic modeling was used to investigate the conventional frequency band near (50 Hz) and bandwidth (20 to 75 Hz) necessary to image fourth-order sequences within the Grayburg Formation. The thickness resolution of the model is roughly 85 feet, which is within the range of thicknesses of the fourth-order sequences. The modeling indicates that thickening and thinning of thin bed sequences cause tuning effects that need to be understood when picking sequence boundaries. For thin beds, geometrical considerations appear to predominate, making acoustic differences in lithology of secondary importance. Conclusions from the modeling are: 1) 3-D seismic data need to be wavelet processed to zero phase, and relative amplitudes preserved, before comparison with modeling results to make interpretive stratigraphic decisions. 2) Modeling algorithms may not include real-earth effects, such as dispersion, source-generated noise, attenuation, etc. These differences provide an element of uncertainty in the comparison of a modeled response to real data.