Diffraction imaging has been used in fault imaging and direct fracture imaging for many years in surface seismic data processing. Surface seismic processing has generally focused on fault zones that are on the order of meters in thickness. According to theory and modeling results however, even reservoir fractures on the order of millimeters in thickness should radiate diffracted seismic energy when the fractures are stimulated by seismic energy from a surface seismic source. In this study the theory was tested by using a time lapse seismic dataset recorded with surface Vibroseis sources and downhole DAS data recording in the horizontal leg of a well. The baseline seismic survey was recorded before the well was hydraulically stimulated and a second survey after about 4 months during which hydraulic stimulation, flowback, and production occurred. The seismic data exhibited good repeatability and were processed in a surface consistent manner to retain source wavelet and amplitude consistency. A diffraction summing algorithm was applied to the seismic data under the assumption that if diffractions were present in the data then the algorithm would produce images with known characteristics of fractures. The resulting images showed near-vertical events with characteristics of fractures including a marked increase in the number of events and an increase in their amplitude between the pre- and post-stimulation datasets lending credibility to the idea that the images were indeed revealing fractures. Fracture images should have numerous applications in reservoir development.
Seismologists have known for decades that faults and large fractures act as discontinuities in the subsurface and generate diffraction energy from incident seismic waves. In fact, diffraction energy recorded in seismic data provides important signal for seismic migration programs to sum into images of faults (Yilmaz, 1987). Seismic data recorded at the earth's surface has also been used to successfully image fractures and dike systems using low-amplitude diffraction signals (Pedersen and Marcy, 2022; Tyiasning, et. al., 2016) that are generated at fault discontinuities. Processing data to extract diffraction images typically includes data processing methods that preferentially ignore primary reflections in favor of migrating only diffraction energy (Xue et. al., 2017).
