Report Title: Multi-attribute Seismic/rock Physics Approach to Characterizing Fractured Reservoirs Type of Report: Final Report Reporting Period: for the Period of 10/01/1999 – 09/30/2004 Pi:

Most current seismic methods to seismically characterize fractures in tight reservoirs depend on a few anisotropic wave propagation signatures that can arise from aligned fractures. While seismic anisotropy can be a powerful fracture diagnostic, a number of situations can lessen its usefulness or introduce interpretation ambiguities. Fortunately, laboratory and theoretical work in rock physics indicates that a much broader spectrum of fracture seismic signatures can occur, including a decrease in Pand S-wave velocities, a change in Poisson’s ratio, an increase in velocity dispersion and wave attenuation, as well as well as indirect images of structural features that can control fracture occurrence. The goal of this project was to demonstrate a practical interpretation and integration strategy for detecting and characterizing natural fractures in rocks. The approach was to exploit as many sources of information as possible, and to use the principles of rock physics as the link among seismic, geologic, and log data. Since no single seismic attribute is a reliable fracture indicator in all situations, the focus was to develop a quantitative scheme for integrating the diverse sources of information. The integrated study incorporated three key elements: The first element was establishing prior constraints on fracture occurrence, based on laboratory data, previous field observations, and geologic patterns of fracturing. The geologic aspects include analysis of the stratigraphic, structural, and tectonic environments of the field sites. Field observations and geomechanical analysis indicates that fractures tend to occur in the more brittle facies, for example, in tight sands and carbonates. In contrast, strain in shale is more likely to be accommodated by ductile flow. Hence, prior knowledge of bed thickness and facies architecture, calibrated to outcrops, are powerful constraints on the interpreted fracture distribution. Another important constraint is that fracturing is likely to be more intense near faults – sometimes referred to as the damaged zone. Yet another constraint, based on world-wide observations, is that the maximum likely fracture density increases with depth in a welldefined way. Defining these prior constrains has several benefits: they lead to a priori probability distributions of fractures, that are important for objective statistical integration; they limit the number of geologic hypotheses that need to be theoretically modeled; they provide plausible models for fracture distributions below the seismic resolution. The second element was theoretical rock physics modeling of optimal seismic attributes, including offset and azimuth dependence of traveltime, amplitude, and impedance signatures of anisotropic fractured rocks. The suggested workflow is to begin with an elastic earth model, based on well logs, theoretically add fractures to the likely facies as defined by the geologic prior information, and then compute synthetic seismic traces and attributes, including variations in P and S-wave velocities, Poisson’s ratio, reflectivity, travel time, attenuation, and anisotropies of these parameters. This workflow is done in a Monte-Carlo fashion, yielding ranges of expected fracture signatures, and allowing realistic assessments of uncertainty to be honored. The third element was statistical integration of the geophysical data and prior constraints to map fracture intensity and orientations, along with uncertainties. A Bayesian framework was developed that allowed systematic integration of the prior constraints, the theoretical relations between fractures and their seismic signatures, and the various observed seismic observations. The integration scheme was successfully applied on an East Texas field site. The primary benefit from the study was the optimization and refinement of practical workflows for improved geophysical characterization of natural fractures and for quantifying the uncertainty of these interpretations. By presenting a methodology for integrating various types of information, the workflow will help to reduce the risk (and therefore the cost) of exploring for and recovering natural gas and oil reserves in fractured reservoirs.