New computer solutions seem to crop up so often in oil and gas exploration that it’s sometimes hard to separate the good stuff from the marginally improved. But there are a few start-ups that are doing some very different things.

One of these is TerraSpark Geosciences. Based in Broomfield, Colo., the company has an unusual history. Geoff Dorn, president of TerraSpark Inc., the general partner, worked in ARCO’s research lab for 20 years, ultimately assembling a crack team to study the use of visualization technology in the industry.

When BP bought ARCO, it wished to do away with the research lab. Dorn convinced BP to donate his group to the University of Colorado at Boulder, along with US $5 million ($4 million from BP and $1 million from Landmark), establishing the BP Center for Visualization at the school.

“During the five years we were at CU, I think we did a lot of good industry-funded research and development,” Dorn said. “But we were always something of a square peg in a round hole.”

In January 2006 the groups decided to “amicably part ways,” and all of the center’s employees, including the graduate students, formed TerraSpark. While Dorn still is an Adjoint Professor in Computer Sciences and Geological Sciences at CU, TerraSpark is a separate entity. Its “incubation” at the university has been a tremendous help.

“We started the company with a business model that was based on minimal up-front capitalization, which we could do because we had been a ‘start-up’ at CU,” Dorn said. “We established the research portion of the business while we were there and brought that with us. We already have those revenue streams to support the core business.”

The core business is to develop a new approach to seismic interpretation. TerraSpark has formed two industry consortia, one focused on wellpath planning and one on interpretation, which help fund the research.

The interpretation group is developing new approaches to structural interpretation — interpreting horizons with surface draping, interpreting faults with automatic fault extraction (AFE), and interpreting the boundaries of canyons andsalt bodies with surface wrapping.

“There’s a completely new class of seismic attributes that we’ve created to help image these boundaries and datasets,” Dorn said.

The process is unique — once the structure is interpreted, the volume is transformed, removing structural deformation so that the stratigraphy can be interpreted separately.

“The interpreter looks at horizontal slices through the transformed stratal slice volume and sees the paleo-depositional systems — fluvial channels, deepwater channels, prograding fan complexes and other things that you simply couldn’t see in the original seismic volume because everything was faulted and dipping,” Dorn said.

The consortia members are so far very enthusiastic about the results — Dorn said that after the mid-year meeting in May, members asked for a hands-on workshop the following month. Two workshops were eventually scheduled to accommodate the demand from the sponsoring companies.

That enthusiasm is pushing the company to the next step, which includes a marketing strategy and staff expansion. Dorn said TerraSpark will soon be poised to offer software either directly from the company or as licensed modules and plug-ins to other types of software (already the company has developed an AFE plug-in for JOA’s Jewel Suite and has licensed well planning technology [IDP] to Transform Software and Services. Commercial versions of TerraSpark’s AFE and IDP are available from Paradigm).

“One of my graduate students was able to interpret around 17 channels in one hour in one stratal volume,” he said. “We’re working on technologies for interpretation that allow the interpreter to be much more productive and find things they couldn’t find before. We think that’s needed in the industry today.”

For more information, visit www.terraspark.com.