Uncertainty theories applied to the analysis of CO2 plume extension during geological storage

Abstract

We use the IRS (Independent Random Set) method to combine both aleatory and epistemic uncertainties in the prediction, by means of an analytical model, of CO2 plume extension during geological storage. While some model parameters can be handled within a classical probability framework, due to the availability of data, other parameters are dealt with using possibility theory, due to the imprecise/incomplete nature of the available information. Results of calculated CO2 plume extension are presented in terms of distributions of the upper and lower probability that plume migration distance lies below a certain value. In a decision-making framework, these results could be combined into a single distribution, referred to here as a “confidence index”, such as a weighted average of upper (optimistic) and lower (pessimistic) probabilities that migration distance is below a certain value. The selected weight reflects the decision-maker’s degree of “risk aversion”.