Data-driven diffraction imaging of fractures using passive seismic data

We present a workflow of seismic imaging for passive seismic data. Unlike conventional diffraction imaging that often adopts reflection-type seismic imaging with known source, our approach relies on the data without the need for passive source information. We use two types of information from passive data: transmitted waves and scattering (coda) waves. The imaging formula states that direct waves should coincide with coda waves at scatterer points at the time of scattering. Instead of generating source wavefields in the conventional imaging method, we back propagate transmitted and scattered data from known surface or borehole receiver arrays. Then we apply zero-lag crosscorrelation imaging condition to produce an image. We can apply this processing for both P- and Swaves. In our numerical examples, we evaluate surface and borehole acquisition scenarios. We found that our approach performed well compared with conventional seismic imaging when assuming exact source information is known. When we perturbed source location, fractures were missing in the conventional seismic imaging map but our approach was not influenced.