Uncertainty quantification for the CO2 storage efficiency in heterogeneous saline aquifers: The impact of variability in the lithology distribution

Abstract

Geological storage of anthropogenic CO${}_2$ can require a preliminary evaluation and ranking of subsurface reservoirs characterized by limited geophysical data. On the basis of regional concepts and trends, only some geostatistical variables can be estimated for the potential storage reservoirs near a CO${}_2$ emitter. Consequently, the storage efficiency may be subject to a substantial uncertainty caused by the variability in the heterogeneous distribution of geological parameters, which details are not exactly known without extensive geophysical surveys. We evaluate that uncertainty by the numerical modeling of the CO${}_2$ storage efficiency in open heterogeneous aquifers. Using geostatistical modeling, we generate a large dataset of synthetic heterogeneous reservoir models characterized by various lithofacies ratios and distributions. We then simulate the injection of CO${}_2$ through a standalone vertical well to propagate the uncertainty forward to the storage efficiency. We show that the efficiency is characterized by a significant variance. Through an extensive parametric study of the statistical characteristics, we propose several recommendations for the CO${}_2$ storage site selection in the regional surveys. We show that the storage is more efficient in aquifers built of thin and extended reservoir interlayers at a specific net-to-gross ratio. The heterogeneity should also be isotropic in the lateral directions. We propose useful scaling relationships for the efficiency against the variogram ranges and the net-to-gross ratio. Finally, we show that the uncertainty in the efficiency cannot significantly be reduced using the well logging data.