Quantifying the Effects of Prior Knowledge and Interpretational Methodology from the Creation of Structural Models

Clare Bond¹, Rebecca J. Lunn³, Zoe K. Shipton², and Alan D. Gibbs^1
1Midland Valley Exploration, Glasgow, United Kingdom
²Geographical and Earth Sciences, University of Glasgow, Glasgow, United Kingdom
³Civil Engineering, University of Strathclyde, Glasgow, United Kingdom

The prior knowledge that individuals apply to datasets has an impact on interpretation and model creation, but the nature of the impact is uncertain. A key question is what factors of a person’s prior knowledge are most influential? Here we have quantified some of the effects of prior knowledge and interpretational methodology on structural model creation for a dataset that supports multiple conceptual models.

Geoscientists often work with inherently uncertain data. In resource exploration datasets of different certainties are combined to build a picture of the sub-surface by data interpretation and model construction. The concepts used in making interpretation choices are based on prior knowledge. During the interpretation the geoscientist will attempt to validate each concept, often sub-consciously sometimes consciously against previous knowledge. Some concepts will be disregarded quickly, others considered more carefully.

In this paper we assess the factors which influenced concept choice during a seismic interpretation exercise by comparing expertise, prior experience, training and discipline with the concepts applied to the interpretation exercise. Rather than considering the impact of each variable in isolation, we have used polytomous regression analysis to assess relative impacts. Our statistical analysis of these data show that two key variables are statistically significant; level of expertise in structural geology and the type of interpretational technique employed. These results are the first to quantifiably demonstrate that interpretational methodologies and conscious use of validation techniques may have more impact on positive interpretational outcome and model creation than prior knowledge. The results show that focused training of staff in particular techniques would have a positive impact on reducing the risk inherent in seismic interpretation.

AAPG Search and Discovery Article #90078©2008 AAPG Annual Convention, San Antonio, Texas