Issues Concerning the Aggregation of Estimates of Undiscovered Oil and Gas Resource Accumulation Units in the Circum-Arctic

John H. Schuenemeyer
Southwest Statistical Consulting, LLC, Cortez, CO.

U.S. Geological Survey recently completed an appraisal of possible future additions to oil and gas reserves from new fields in the Circum-Arctic. In that appraisal 48 assessment units (AU’s) were identified as having at least a 10-percent chance of one or more significant oil or gas accumulations. The AU’s are mappable units of rock with common geologic traits.

In the Circum-Arctic, the assessment team considered three types of association - geologic, regional and human factors. A geologic correlation was derived by considering three possible dependencies between AU’s, namely charge, rocks and timing. The association between each pair of AU’s was assessed and specified as an ordinal value representing a low, medium, or high correlation. These values were then assigned values of 0.3, 0.6 and 0.9 respectively. The resultant correlation used in the sampling of AU field size distributions was the average of the charge, rock and timing correlations.

Two additional possible causes of dependency are a regional factor and human factors. In this assessment, we judged it impossible to separate these factors and arbitrarily assigned a correlation of 0.2 between AU’s in situations where assessors believed correlation to be minimal, which is a correlation below 0.3.

As a specific example, consider the Western Siberian Basin (WSB) province. It was assessed as containing two AU’s, the Northern West Siberian Onshore Gas and the South Kara Sea Offshore. For each of charge, rock and trap, assessors assigned a high correlation. Then using the above metric, the average correlation was 0.9 before any adjustment was made to ensure a correlation matrix. However, the correlation between each of the WSB AU’s and the North Kara Basins and Platforms was judged to be 0.2, the level of background association. Generally AU’s in close spatial proximity to each other were judged to be most highly correlated unless some geologic factor such as faulting was present. After all pairwise associations were estimated, the check for positive definiteness was made. In this case, the resultant 48 x 48 matrix was not a correlation matrix and a bias factor was added. As a result, the correlation between the two WSB AU’s changed from 0.90 to 0.60. While a cause for some uneasiness, assessors judged that this correlation was still within reasonable bounds.

AAPG Search and Discover Article #90096©2009 AAPG 3-P Arctic Conference and Exhibition, Moscow, Russia