A knowledge-integration framework for interpreting seismic facies

Abstract

In recent years, the size of seismic data volumes and the number of seismic attributes available have increased. As a result, the task of recognizing seismic anomalies for the prediction of stratigraphic features or reservoir properties can be overwhelming. One way to evaluate a large amount of data and understand potential geologic trends is to automate seismic facies classification. However, the interpretation of seismic facies remains an elusive issue. Interpreters are confronted with the selection of the clustering technique and the optimal number of seismic facies that best uncover the spatial distribution of seismic facies. An interpretation framework combining data visualization with the results from various clustering techniques was evaluated. The framework allows interpreters to be directly involved in the seismic facies classification process. Because of the active participation, interpreters (1) gain insight into the detected seismic facies, (2) verify hypotheses with respect to the spatial distribution of seismic facies, (3) compare different seismic facies classification, and (4) gain more confidence with the seismic facies interpretation.