The Virtual Source Method

The virtual source method is a technique to image and monitor below a complex and time-varying overburden without the knowledge of overburden velocities and near-surface changes. It is based on extracting the response between a given pair of receivers by correlating the wavefields recorded by them and stacking over the physical sources. There are several challenges that arise while generating virtual source data, such as the maximum allowable spacing of the physical sources, limited acquisition aperture and reflections/multiples from the overburden, that need to be addressed. Some of these challenges can be overcome by separating the wavefield into upgoing and downgoing waves. Wavefield separation not only suppresses the artifacts associated with the limited acquisition aperture typically used in practice, but also helps reconstruct a response in the absence of downgoing reflections and multiples from the overburden. These improvements in removing the artifacts caused by limited acquisition aperture and overburden-related multiples are illustrated on a synthetic dataset of a complex layered model modeled after the Fahud field in Oman, and on OBC data acquired in the Mars field in the deepwater Gulf of Mexico. The virtual source method requires surface shooting recorded by subsurface receivers placed below the distorting or changing part of the overburden. Redatuming the recorded response to the receiver locations allows the reconstruction of a complete downhole survey as if the sources were also buried at the receiver locations. The ability to redatum the data independent of the knowledge of time-varying overburden velocities makes the virtual source method a valuable tool for time-lapse monitoring. The virtual source method is applied to the Mars field OBC data acquired in the deepwater Gulf of Mexico with 120 multicomponent sensors permanently placed on the seafloor. Applying a combination of wavefield separation and deconvolution of the correlation gather by the source power spectrum to the virtual source method suppresses the causes of non-repeatability in the overburden (sea water) and acquisition discrepancies, forming the basis for the improvement the virtual source method offers for time-lapse monitoring. When the two wavefields recorded by the receivers are correlated, along with the estimation of the Green’s function between the two receivers the correlation also contains the power spectrum of the source-time function. For a short duration source, the correlated data can be deconvolved by the source power spectrum. When, however, the source-time function is long and difficult to estimate (e.g., earthquake or other passive seismic recording), deconvolution as opposed to correlation becomes a preferred seismic interferometric tool because the deconvolved wavefield is independent of the source-time function. To demonstrate the usefulness of deconvolution as a tool for seismic interferometry, the method is applied to earthquake waves recorded in a borehole to extract near-surface properties such as the 1-D velocity profile. Furthermore, a connection is established between the deconi volved waveforms and the elements of the propagator matrix. Using the same earthquake recording, a P-to-S mode conversion is characterized by extending the application of the receiver function to downhole measurements.