During the past two decades or so, significant progress has been made in our theoretical understanding of the mechanical (elastic) and hydraulic (transport) behavior of fractures.  On an empirical level, we have made enough observations, both in the laboratory and on real data, to be able to categorize pretty well how seismic waves behave in fractured rock.  For instance, we know that the size, shape, contacts and connectivity of the surfaces of fractures are important, as are the number of fracture sets and their orientations.  Geophysicists may argue that we now know enough about the seismic response to fractures and may be able to predict their presence and spatial distribution in reservoirs.  However, we cannot yet claim success in obtaining these fracture parameters from seismic data since we have not yet convinced reservoir engineers to routinely use seismically derived ‘fracture’ parameters in their flow simulations.  

In this presentation, I will review the concept of fracture prediction from seismic data, focusing on fracture modeling, and will highlight our current understanding of its limitations and the challenges ahead.