Azimuthal variations in attributes for induced fracture detection in the Barnett Shale

Abstract

Wide azimuth seismic data acquisition has become standard in many plays where azimuthal characterization of fractures or stress is important for fully understanding reservoir behavior. Usually, the desire is to image preexisting fracture networks, but, in artificially stimulated reservoirs, it is also desirable to understand the fracture state after completion. To characterize the induced fractures in a heavily drilled area of the Barnett Shale, a wide azimuth survey was acquired and processed to generate four separate azimuthal stacks. Attributes exploiting variations in amplitude and waveform were independently calculated for each azimuthal stack. Azimuthal variations in the attributes were then quantified through ellipse fitting. The attributes with the most statistical significance in azimuthal variability were those related to frequency such as the wavelet dominant frequency. These attributes were consistent with each other but show a wide distribution over azimuths across the survey. In addition, random distribution of orientations within the Barnett shale compared to the over and underlying Marble Falls and Viola Limestones indicate a difference in orientation for the induced as opposed to preexisting fractures. With the ability to map the induced fractures, it becomes feasible to take an intelligent approach to planning infill locations to avoid the previously fractured areas.