Understanding bias: An example of 'concept uncertainty' in seismic interpretation

To maximize the commercial value of new ventures we need to increase our own understanding of the geology of petroleum systems and to deal with this information strategically within industry workflows. We have investigated the initial stages of the geological workflow and designed experiments to quantify how professionals develop a geological interpretation from seismic data. Our experiments consider both the process of developing an interpretation and the final model or concept produced. In a number of "blind" trials we have demonstrated a less than 21% interpretational success rate in identifying a known tectonic setting and a less than 2% success at identifying 90% of the key features. The range of concepts produced and their variance from the known solution results in a high level of uncertainty in a critical part of the industry workflow. This uncertainty has significant implications for decision making and risk management. Our results quantify the uncertainty within geological interpretation and show that prior knowledge has an effect on the type of geological interpretations and concepts produced.