Quantitative seismic reservoir characterization of tight sands (granite wash) play at Stiles Ranch field in the Anadarko Basin, Texas (USA)

The main objective of this study is to conduct quantitative seismic reservoir characterization study of the Granite Wash (Marmaton-tight sand) play at Stiles Ranch field in the Anadarko Basin, Texas (USA). The proposed methodology incorporates seismic petrophysics, rock physics, Amplitude Variation with Offset (AVO) analysis and seismic pre-stack simultaneous elastic impedance inversion. In addition, it utilizes geostatistical technique to improve the reservoir property estimation and quantify uncertainty in seismic lithology and fluid prediction.;The general objective encompasses several more specific goals to study: well data conditioning and prediction of essential petrophysical properties (e.g., porosity, permeability and saturation), and their relationship to the elastic properties. Due to the multidisciplinary nature of seismic petrophysics, only three core aspects are focused on that cover the desired objectives: 1) porosity modeling, 2) shear wave prediction, and (3) fluid substitution. The rock types are characterized by Rock Physics Diagnostic (RPD) approach conducted on well log data calibrated with core data and thin sections.;The Granite Wash reservoir elastic properties are upscaled from log to seismic scale using Backus averaging to obtain a more coarsely (upscaled) sampled data set equivalent to the seismic scale. Anisotropy parametric (epsilon, gamma and delta) log curves are estimated consistent with seismic measurements using rock properties, seismic velocity and clay volume (Vsh) as a function of depth. The reservoir elastic properties are related to both the depositional environment and burial history through rock physics depth trends as function of depth.;Furthermore, based on the practical aspects two separate inversion approaches; AVO and Elastic Impedance (EI) are evaluated prior to their application to real seismic. Various AVO derived attribute volumes such as intercept (A), gradient (B) and reflection coefficients (scaled Poisson's ratio, fluid factor, and sum of reflection coefficients) are quantitatively used to help identify prospective hydrocarbon zones. The accuracy of various EI formulas is discussed and the most optimal one is used for pre-stack simultaneous elastic inversion.;Finally, post-stack and pre-stack simultaneous elastic inversions are comparatively analyzed on AVO compliant re-processed data for fluids prediction. Pre-stack simultaneous elastic inversion is preferred over post-stack inversion which provides more promising results. The net-to-gross (N/G) ratio is estimated using a quantitative method calibrated with log derived seismic rock properties. Finally, a probabilistic approach is applied to derive lithofacies information from inverted seismic attributes and respective well log data to quantify the uncertainty in the seismic lithology prediction (hydrocarbon saturation). Lithofacies probability maps derived from log and seismic attributes provide a better prediction of reservoir producing and non-producing zones.