Researchers Apply Quantum Physics Idea to Seek Improvements in Seismic Imaging

May 15, 2007

AUSTIN, Texas—Despite decades of improvements, seismic imaging—one of the geosciences' key methods for visualizing the subterranean world—remains a tad blurry.

Sergey Fomel, a research scientist at the Bureau of Economic Geology at the Jackson School of Geosciences, and colleagues Evgeny Landa of Organisme Pétrolier de Recherche Appliquée and Tijmen-Jan Moser of Zeehelden Geoservices set out to to improve this critical tool used for oil and gas exploration, examination of soil and water contamination, and study of Earth’s internal structures.

Scientists interpreting seismic data still have to guess what kinds of rocks the emitted sound waves are passing through and the velocities of those waves. They make their best guesses using computer models but uncertainties lead to grainy images. Some geologic features are obscured and phantom artifacts appear that don’t actually exist.

“The key idea in this paper, which was actually the idea of the first author, Landa, was that one should not try to find the unique velocity because it is uncertain,” said Fomel. “So instead, you try many different velocities at once.”

It’s an idea they borrowed from quantum physics.

In classical physics, when something moves—say, a ball is dropped from the top of a skyscraper or a spaceship blasts off into space—there is only one path that the object takes. You can draw an arc in the air, point to it and say, “That’s the path it took.”

In the strange world of quantum physics, objects don't follow a single path from point A to point B. Instead, they take all the possible paths between A and B. Image: Matt McIrvin.

But in quantum physics (which is better for describing the actions of very small and very fast things, such as electrons or photons of light), things aren’t so cut and dry. An object moving from point A to point B actually takes multiple paths to get there, all simultaneously. Not only that, but there are different degrees of taking a path, so that it might mostly take one or two paths, but only slightly take a few other paths.

It might all sound counterintuitive, but quantum physicists have used these ideas to develop a mathematical way to manage uncertainty. In this case, uncertainty about the path an object took. And it works. (Richard Feynman, a brilliant and eccentric physicist, won the 1965 Nobel prize for his creation of this “path integral formulation.”)

Fomel and his colleagues applied the path integral technique to uncertainties in seismic velocities. And the images they produce of the subsurface are clearer than those made with conventional seismic imaging.

Comparison of (a) Conventional (optimized) time-migrated image of North Sea seismic data, obtained by velocity continuation; (b) path-integral time migrated image of North Sea data.
Click image to expand.

The researchers won the European Association of Geoscientists and Engineers’ 2006 Loránd Eötvös Award for their paper, “Path-Integral Seismic Imaging.” The award is presented to the authors of the year’s best paper published in the journal “Geophysical Prospecting.” The collaboration was sponsored by Total, a French oil and gas company.

Fomel said that their proposed method is not yet fully practical, so don’t expect to see it used directly.

“I think the value of the paper is in provoking new ideas and emphasizing the idea of uncertainty in our knowledge about the subsurface,” he said. “I think that’s why the paper was recognized. Not because it proposed a new practical method, but because it opened a new area of ideas around the uncertainty.”

by Marc Airhart

For more information about the Jackson School, contact J.B. Bird at jbird@jsg.utexas.edu, 512-232-9623.