Quantifying Uncertainty in Statistical Regression Techniques for Reservoir Mapping

C032 QUANTIFYING UNCERTAINTY IN STATISTICAL REGRESSION TECHNIQUES FOR RESERVOIR MAPPING Abstract 1 Reservoir characterization requires a good understanding of the spatial distribution of reservoir properties. Statistical regression techniques such as multiple regression and neural networks have been used extensively for this purpose. This paper presents a novel technique to understand the limitations of the reservoir data and to quantify the uncertainty of the subsequent mapping results. We use the Pinedale Anticline dataset in Wyoming to illustrate the workflow and the usefulness of the technique. A set of structural and seismic attributes is to map the cumulative gas production a proxy