Risk management in a large-scale CO2 geosequestration pilot project, Illinois, USA
Like most large-scale infrastructure projects, carbon dioxide (CO 2) geological sequestration (GS) projects have multiple success criteria and multiple stakeholders. In this context "risk evaluation" encompasses multiple scales. Yet a risk management program aims to maximize the chance of project success by assessing, monitoring, minimizing all risks in a consistent framework. The 150,000-km2 Illinois Basin underlies much of the state of Illinois, USA, and parts of adjacent Kentucky and Indiana. Its potential for CO2 storage is first-rate among basins in North America, an impression that has been strengthened by early testing of the injection well of the Midwest Geological Sequestration Consortium's (MGSC's) Phase III large scale demonstration project, the Illinois Basin - Decatur Project (IBDP). The IBDP, funded by the U.S. Department of Energy's National Energy Technology Laboratory (NETL), represents a key trial of GS technologies and project-management techniques. Though risks are specific to each site and project, IBDP risk management methodologies provide valuable experience for future GS projects. IBDP views risk as the potential for negative impact to any of these five values: health and safety, environment, financial, advancing the viability and public acceptability of a GS industry, and research. Research goals include monitoring one million metric tonnes of injected CO2 in the subsurface. Risk management responds to the ways in which any values are at risk: for example, monitoring is designed to reduce uncertainties in parameter values that are important for research and system control, and is also designed to provide public assurance. Identified risks are the primary basis for risk-reduction measures: risks linked to uncertainty in geologic parameters guide further characterization work and guide simulations applied to performance evaluation. Formally, industry defines risk (more precisely risk criticality) as the product L*S, the Likelihood multiplied by the Severity of negative impact. L and S are each evaluated on five-point scales, yielding a theoretical spread in risk values of 1 through 25. So defined, these judgment-based values are categorical and ordinal - they do not represent physically measurable quantities, but are nonetheless useful for comparison and therefore decision support. The "risk entities" first evaluated are FEPs - conceptual Features, Events, and Processes based on the list published by Quintessa Ltd. After concrete scenarios are generated based on selected FEPs, scenarios become the critical entities whose associated risks are evaluated and tracked. In IBDP workshops, L and S values for 123 FEPs were generated through expert elicitation. About 30 experts in the project or in GS in general were assigned among six facilitated working groups, and each group was charged to envision risks within a sphere of project operations. Working groups covered FEPs with strong spatial characteristics - such as those related to the injection wellbore and simulated plume footprint - and "nonspatial" FEPs related to finance, regulations, legal, and stakeholder issues. Within these working groups, experts shared information, examined assumptions, refined and extended the FEP list, calibrated responses, and provided initial L and S values by consensus. Individual rankings were collected in a follow-up process via emailed spreadsheets. For each of L and S, three values were collected: Lower Bound, Best Guess, and Upper Bound. The Lower-Upper Bound ranges and the spreads among experts can be interpreted to yield rough confidence measures. Based on experts' responses, FEPs were ranked in terms of their L*S risk levels. FEP rankings were determined from individual (not consensus or averaged) results, thus no high-risk responses were damped out. The higher-risk FEPs were used to generate one or more concrete, well defined risk-bearing scenarios for each FEP. Any FEP scored by any expert as having associated risk of