Seismic Models in the Public Domain - Impact on the Petroleum Resource Industry 

SEG Advanced Modeling (SEAM) Corp.

William L. Abriel - Chairman 

Forward seismic models in the public domain have significantly advanced both the art and the application of geophysics in the gas and oil industry. We should expect the same from current and future models. Forward modeling for private enterprise is most often used to confirm a specific interpretation from seismic data to match either the stacked image or the AVO response. In contrast, public domain models are generated primarily to test a process (such as migration), define a workflow, or for the purposes of education and illustration. 

Forward models are useful because they are a controlled experiment with a limited number of parameters. They are much less expensive than field acquisition and provide results with no interpretation ambiguity. The limits to modeling lie primarily in the design state. For instance statics models designed to test data processing solutions in karst terrains are not appropriate for permafrost. Further, the robustness of models is often limited by our lack of knowledge of geological complexity and computational limitations 

Models in the public domain have a purpose different than proprietary models. Instead of focusing primarily on a particular prospect, public domain models are used to provide a common challenge of specific importance to the industry. By maximizing the number of researchers focused on a common problem, the chances of generating a breakthrough technology are significantly increased. Further, a shared model provides a comparative benchmark and a common language for illustration.

Several historical examples of public domain models illustrate their benefits. In the mid 1970’s, the University of Houston produced a classic model by physical acquisition of marine seismic data in a water tank over a 3-D object. Known as “the French model” (for designer Bill French), this model proved to be a key vehicle to illustrate the fundamental differences in 2-D and 3-D migration of seismic data. Used for years by geophysicists as a benchmark to measure the quality of their migration algorithms, the French model significantly advanced the cause of 3-D seismic acquisition in exploration and development and nicely illustrated imaging theory to geologists and geophysicists alike. 

Other examples of public domain models show similar positive impact on the state of the industry’s seismic technology. The Marmousi 2-D numerical model generated in the mid 1980’s focused on the complexity of velocity in seismic imaging. For decades, this model has proven to be a robust challenge and an industry benchmark for advancing the development of 2-D seismic tomography and imaging. In the early 1990’s the generation and distribution of the 3-D numerical SEG/EAGE salt model proved to be another example of the high value of a public domain data set. This challenge significantly advanced the applications and art of 3-D seismic depth imaging, and is still sought after as a benchmark after more than 20 years.

Today, the SEG Advanced Modeling Corporation (SEAM) is leading the development of models and simulated datasets relevant to the current challenges faced by industry. SEAM was formed by SEG for the purpose of coordinating a continuous series of public domain models to advance the state of geophysical technology. In Phase I, twenty three company participants have designed an executed a very large and geologically complicated 3-D model consistent with challenges of the Gulf Coast and other offshore Tertiary basins. Key elements of this model include salt body complexity, a number of stacked and stratigraphicly complex reservoirs, and realistic noise characteristics. These data are held by the participants of SEAM for a two year proprietary period. Arrangements for access to the data is through these participants during this period, after which general public access will be available.

The Phase I model challenge is a significant contribution to the industry and a major step change in public model initiatives. It is anticipated that significant advances will result in field acquisition design, seismic reservoir characterization, velocity imaging and migration, and multiple elimination. Recognizing an opportunity from the USA Department of Energy, SEAM applied for and was awarded a grant by the Research Partnership to Secure Energy for America (RPSEA). This has doubled the SEAM budget and extended the scope of Phase I and is enabling the current acquisition of data sets for a noise free benchmark, 3-D gravity, magnetics, electromagnetics, seismic anisotropy (3-D TTI) and 3-D elastic waves. Thus the Phase I model will also serve as a highly robust benchmark for many additional technologies including the simultaneous inversion of multiple geophysical data sets.

The future for SEAM is to move into the next challenge of building Phase II. This new phase will concentrate on land seismic data including the impact of near surface variations, land noise elimination, acquisition strategies and fractured reservoir characterization. Participation for Phase II is scheduled to open on the first of January of 2011.