We propose an approach to assess the importance ranking of uncertainty sources, with regards to the behavior of the mobile CO2 during the post-injection period: a variance-based global sensitivity analysis is performed on three output parameters characterizing the location and the quantity of mobile CO2, considering residual and dissolution trapping. The use of advanced meta-modeling techniques of ACOSSO-type allows circumventing two major difficulties: 1. The large number of computationally intensive reservoir-scale flow simulations; 2. The different nature of uncertainties whether linked to parameters (continuous variables) or to modeling assumptions (scenario-like variables). The feasibility of the approach is demonstrated using a potential site for CO2 storage in the Paris basin (France), for which the amount, nature and quality of the data available at disposal and the associated uncertainties can be seen as representative to those of a storage project at the post-screening stage. A special attention has been paid to confront the results of the sensitivity analysis with the physical interpretation of the processes. (C) 2014 The Authors. Published by Elsevier Ltd.
- Residual trapping
- dissolution trapping
- variance based sensitivity analysis
- ACOSSO metamodeling
- IMBIBITION RELATIVE PERMEABILITY
- SUPERCRITICAL CO2/BRINE
- RESERVOIR CONDITIONS
- GEOLOGICAL STORAGE
Manceau, JC (reprint author), Bur Rech Geol & Minieres, 3 Ave C Guillemin, F-45060 Orleans 2, France.