HOW DOES RESOLUTION OF SEDIMENTARY ARCHITECTURE DATA AFFECT PLUME DISPERSION IN MULTISCALE AND HIERARCHICAL SYSTEMS?

Hierarchical Lagrangian-based transport model is an effective approach to address the impact of sedimentary texture on solute transport in heterogeneous porous formations. This model is based on an expression of global covariance function, which represents a hierarchical organization and combine lithofacies proportions, covariances of log-permeability as well as facies transition probabilities. Incorporating more sedimentary structure data is beneficial for representing the spatial correlation structure. However, the extent to how the resolution of sedimentary architecture (i.e., indicator) data affect the model estimations remains unclear. In this study, we specifically investigate their effects on the dispersivity and the particle displacement variance by comparing models with different resolution data sets collected on an outcrop exposure in Espanola Basin. Polynomials chaos expansion based Sobol’ indices global sensitivity analysis was conducted here to explore how the data resolution affect plume dispersion in multiscale systems. Our results show that the dispersivity and particle displacement variance are underestimated if less resolution lithofacies data are used, even though it can provide reasonable estimation on spatial correlation structure. Global sensitivity analysis indicates that longitudinal dispersivity is most sensitive to indictor scale and the mean log-permeability. As for the transverse dispersivity, anisotropy ratio and mean log- permeability are the most sensitive parameter.

Key: Solute transport; Dispersion; Sedimentary architecture; Lagrangian-based model; Sensitivity analysis