Riemannian Wavefield Extrapolation

Offshore oilfields in the Gulf of Mexico and elsewhere around the earth are embedded in complex formations of the earth's crust. It is absolutely necessary to understand these formations in terms of their physical structure and geologic contents to assure that the fields may be fully exploited to supply the world's energy needs.

Since it is not possible to directly observe complex subsurface structures, the structures and their contents must be inferred from seismic experiments. A high-powered seismic wave is directed into the earth's crust, and scientists examine the field of waves that bounce back in the form of a seismic recording. Far fainter than an X-ray picture, the seismic images are, to the untrained eye, messes composed of wavy lines.

Dr. Fomel is a specialist in developing seismic imaging algorithms. In particular, he and colleagues have developed an algorithm called Riemannian wavefield extrapolation (RWE) that is uniquely adapted to interpreting extremely intricate seismic returns. Under the Gulf of Mexico, for example, there are large blocks of crust—salt formations-––and numerous "subsalt sediment structures" that have been rearranged over geologic time. The seismic reflections of this crustal jumble must be untangled computationally to reveal the steeply dipping and overturning structures. Fomel's RWE can handle the large velocity contrasts in the seismic returns and the divisions of a single signal into multiple pathways, as well as other wave-propagation effects that are expressed over a broad band of frequencies. RWE runs efficiently on the Lonestar cluster at TACC, and Fomel is currently assemblling some of the most accurate picture to date of undersea oilfields in the Gulf.

Fomel notes that the project is sponsored by ExxonMobil and that he works in collaboration with ExxonMobil scientists.