December is one of my favorite issues of the year because it's the issue in which we showcase emerging technology.

Emerging technology is not, in my mind, to be confused with a tech trend. It's cutting-edge, leading-edge, dare we say bleeding-edge technology that has the potential to really blow the roof off of everyone's preconceived notions about the way things should be done. It puts the "Gee!" back in geoscience. And despite the constant grumblings about a lack of research and development funding for exploration and production technology, there is some way cool stuff going on behind the scenes.

I had the distinct pleasure of being introduced to Dr. Arthur Weglein at an IBM press briefing earlier this year. Weglein has a job title that practically fills the rest of this page, but effectively he's a distinguished physics professor at the University of Houston. He's also the director of a program called "Mission-Oriented Seismic Research Program" (M-OSRP). Not surprisingly, it was this latter title that piqued my interest.

M-OSRP addresses problems whose solutions will have the biggest positive impact on the location and production of hydrocarbons. Sponsors include all of the major publicly traded oil companies worldwide, four national oil companies and several service companies.

"We have difficulty exploring under salt, basalt, karsted sediments, volcanics and any other subsurface geology where there's rapid change in either a boundary between media or within the media, or both," Weglein said. "These things are causing havoc with our ability to define deepwater targets, and they're the underlying reason for the dry hole rate in the Gulf of Mexico."

The government is concerned that companies that are now attempting to exploit US reserves will eventually tire of these disappointments and go elsewhere for their bounty.

Here's the basic tenet of Weglein's research: In order to locate a target within the confines of current methods, it's necessary to estimate what's in between the surface and the target. "This has to be well enough estimated because the signal that returns cares about what it went through on the way down, what it hit and what it went through on the way back," Weglein said. "We try to remove the effect on the returning signal of what it went through on the way down and the way back up so that we extract the part that corresponds to where and what it hit."

But the data collected have more of a story to tell if only they're asked the right questions. What if, Weglein proposes, the data processor gives the data an idea, his or her opinion, of what the signal might have gone through on the way down and the way back up, and then lets the data "talk" amongst themselves to determine the accuracy of that estimate? "The imaged primary data overall have their own opinion," he said. "If you happen to be lucky and you give them the right information about what all the waves have experienced in their downward and upward propagation, then the communication that we've set up will send back a signal that everything's OK. If the processor's input about those experiences has been far from adequate, the data will start to talk, and the algorithm will get more complicated (and more expensive). The idea is to allow all of the primary reflections from all reflectors in the data to collectively express their view on the processor's treatment of each of the members in their group, and when they decide that one of their members needs help, they collectively assist in achieving the processing objective of target location and identification, without knowing or requiring the traditional actual downgoing and upgoing experience of any event.

"We develop algorithms that provide this more inclusive, participatory and empowering role for data in achieving precise processing objectives while removing the traditional need for knowing or determining adequate overburden information. That fundamentally new seismic vision and capability is aimed at making currently inaccessible reservoirs accessible. No accurate velocity model, no picking of events, no interpretive intervention is required."

In other words, what is considered a final output today is the input to Weglein's method. Given adequate compute power and the right algorithms, the data can be empowered into having almost a group therapy session, since as Weglein explains the more inaccurate the estimate is going in, the more agitated they become until they've settled on a consensus on the correct spatial location of each reflector.

This is a direct extension to the earlier work of Weglein and his colleagues on the removal of multiples, without the need or determination of subsurface properties: a fundamentally new comprehensive data processing vision and capability directly responding to our current most significant E&P challenges.

Comments