Datapages, Inc.Print this page

Inference to the Best Geologic Explanation and Error Types in Reservoir Characterization and Modeling

Ma, Yuan Z.1; Gomez, Ernest 1
1 DCS, Schlumberger, greenwood Village, CO.

Reservoir characterization and modeling almost always faces a problem of lack of data, as core and well-log data are limited. This problem, while serious, can be often mitigated by using a geologic conceptual model interpreted from 3D seismic data, outcrop analogues, comparison with contemporary sedimentation and choosing appropriate methods to reconcile the inconsistencies in the available data and to realistically predict reservoir properties. This is called inference to the best explanation or abductive reasoning by science philosophers. However, with the lack of the direct information of the subsurface, even the best geologic inference contains uncertainty, which explains why uncertainty and risk analysis has received significant attention of geoscientists in the last decade or so.

Uncertainty analysis and reservoir modeling involves many disciplines, including geology, geophysics, petrophysics and reservoir engineering. In this paper, error types associated with uncertainty analysis in reservoir characterization and modeling will be discussed, including Type I error or false positives (such as drilling dry holes), Type II error or false negatives (such as underestimations of subsurface resources, and farming out a prolific reservoir), Type III error or correct positives for the wrong reason (such as correct estimate of resources by overestimating pore space and underestimating hydrocarbon saturation) and Type IV error or correct negatives for the wrong reason. Examples for these error types will be given, and their analysis will provide insights for better uncertainty evaluation in reservoir characterization and modeling.


AAPG Search and Discovery Article #90090©2009 AAPG Annual Convention and Exhibition, Denver, Colorado, June 7-10, 2009