Reduced Order Models for Prediction of Groundwater Quality Impacts from CO2 and Brine Leakage

Abstract

Abstract A careful assessment of the risk associated with geologic CO2 storage is critical to the deployment of large-scale storage projects. A potential risk is the deterioration of groundwater quality caused by the leakage of CO2 and brine leakage from deep subsurface reservoirs. In probabilistic risk assessment studies, numerical modeling is the primary tool employed to assess risk. However, the application of traditional numerical models to fully evaluate the impact of CO2 leakage on groundwater can be computationally complex, demanding large processing times and resources, and involving large uncertainties. As an alternative, reduced order models (ROMs) can be used as highly efficient surrogates for the complex process-based numerical models. In this study, we represent the complex hydrogeological and geochemical conditions in a heterogeneous aquifer and subsequent risk by developing and using two separate ROMs. The first ROM is derived from a model that accounts for the heterogeneous flow and transport conditions in the presence of complex leakage functions for CO2 and brine. The second ROM is obtained from models that feature similar, but simplified flow and transport conditions, and allow for a more complex representation of all relevant geochemical reactions. To quantify possible impacts to groundwater aquifers, the basic risk metric is taken as the aquifer volume in which the water quality of the aquifer may be affected by an underlying CO2 storage project. The integration of the two ROMs provides an estimate of the impacted aquifer volume taking into account uncertainties in flow, transport and chemical conditions. These two ROMs can be linked in a comprehensive system level model for quantitative risk assessment of the deep storage reservoir, wellbore leakage, and shallow aquifer impacts to assess the collective risk of CO2 storage projects.