Brine migration resulting from CO2 injection into saline aquifers - An approach to risk estimation including various levels of uncertainty

Comprehensive risk assessment is a major task for large-scale projects such as geological storage of CO2. Basic hazards are damage to the integrity of caprocks, leakage of CO2, or reduction of groundwater quality due to intrusion of fluids. This study focuses on salinization of freshwater aquifers resulting from displaced brine. Quantifying risk on the basis of numerical simulations requires consideration of different kinds of uncertainties and this study considers both, scenario uncertainty and statistical uncertainty. Addressing scenario uncertainty involves expert opinion on relevant geological features such as caprock properties, faults, and distinct geological layers. This is considered in this work by 6 different scenarios having different characteristic geological features. On the other hand, Monte Carlo methods are a classical approach to address statistical uncertainty. This is not feasible for large-scale 3D models including complex physics. Therefore, we apply a model reduction based on arbitrary polynomial chaos expansion combined with probabilistic collocation method. It is shown that, dependent on data availability, both types of uncertainty can be equally significant. The presented study provides estimates of the risk of brine discharge into freshwater aquifers due to CO2 injection into geological formations and resultant salt concentrations in the overlying drinking water aquifers. (C) 2012 Elsevier Ltd. All rights reserved.