| [First Hit] |
|---|
Click to view
presentation in PDF format.
Does Scenario Modeling Really Lead to an Explosion in the Amount of 
Work We Have to Do?*
By
Alan D. Gibbs1, Clare Bond2, Roderick J. Muri2 ,and Zoe K. Shipton3
Search and Discovery Article #70041 (2008)
Posted August 5, 2008
*Adapted from oral presentation at AAPG Annual Convention, San Antonio, TX, April 20-23, 2008
1Midland Valley Exploration, Glasgow, United Kingdom ([email protected])
2Midland Valley Exploration, Glasgow, United Kingdom
3Geographical and Earth Sciences, University of Glasgow, Glasgow, United Kingdom
Traditional best practice leads us to make detailed syntheses of data to build a geological model. The model produced, we believe, is the best solution given the constraints of data and time. These models are often vigorously defended representing months of work that embody the distillation of our knowledge and experience.
Even with the best data and our best endeavors we find that on drilling, or the acquisition of additional data, the model is inadequate or even wrong. The model needs to be either modified or the modeling process must begin again. Many industry projects go through a cyclical workflow of data acquisition, model building and drilling, with projects evolving through several very different model paradigms in their lifetime. The authors have carried out some controlled studies to assess the level of uncertainty inherent in interpretation. This work has indicated that even when the best interpretational practices are deployed the creation of a single deterministic model will still lead to a significant level of uncertainty.
Recognition that geological datasets are massively unconstrained means that we need to adopt new workflows to define the range of "possible" models. Once the full range of models is acknowledged, they can be ranked for their impact on outcome and hence decision. Using current interpretation
and software methodologies, multiple complete models would need to be built prior to making decisions. This paper outlines revised workflows necessary to test model sensitivity to change, outlining the characteristics of software toolsets that enable interpreters to create a large number of scenario and sensitivity cartoons to support the decision process.