Two-dimensional (2D) and 3D numerical simulations of the dispersive Henry problem show that heterogeneity affects seawater intrusion differently in 2D and 3D. When the variance of a multi-Gaussian isotropic hydraulic conductivity field increases, the penetration of the saltwater wedge decreases in 2D while it increases in 3D. This is due to the combined influence of advective and dispersive processes which are affected differently by heterogeneity and problem dimensionality. First, the equivalent hydraulic conductivity controls the mean head gradient and therefore the position of the wedge. For an isotropic medium, increasing the variance increases the equivalent conductivity in 3D but not in 2D. Second, the macrodispersion controls the rotation of the saltwater wedge by affecting the magnitude of the density contrasts along the saltwater wedge. An increased dispersion due to heterogeneity leads to a decreasing density contrast and therefore a smaller penetration of the wedge. The relative magnitude of these two opposite effects depends on the degree of heterogeneity, anisotropy of the medium, and dimension. Investigating these effects in 3D is very heavy numerically; as an alternative, one can simulate 2D heterogeneous media that approximate the behaviour of the 3D ones, provided that their statistical distribution is rescaled.

The Korba aquifer is located in the east of the Cape Bon peninsula in Tunisia. A large groundwater depression has been created in the central part of the aquifer since the 1980s, due to intense groundwater pumping for irrigation. The data collected show that the situation continues to deteriorate. Consequently, seawater is continuing to invade a large part of the aquifer. To better understand the situation and try to forecast its evolution, a three-dimensional (3D) transient density-dependent groundwater model has been developed. The model building process was difficult because of data required on groundwater discharge from thousands of unmonitored private wells. To circumvent that difficulty, indirect exhaustive information including remote sensing data and the physical parameters of the aquifer have been used in a multi-linear regression framework. The resulting 3D model shows that the aquifer is over-exploited. It also shows that after 50 years of exploitation, the time needed to turn back to the natural situation would be about 150 years if the authorities would ban all exploitation now. Such an asymmetry in the time scales required to contaminate or remediate an aquifer is an important characteristic of coastal aquifers that must be taken into account in their management.

The performances of kriging, stochastic simulations and sequential self-calibration inversion are assessed when characterizing a nonmultiGaussian synthetic 2D braided channel aquifer. The comparison is based on a series of criteria such as the reproduction of the original reference transmissivity or head fields, but also in terms of accuracy of flow and transport (capture zone) forecasts when the flow conditions are modified. We observe that the errors remain large even for a dense data network. In addition some unexpected behaviours are observed when large transmissivity datasets are used. In particular, we observe an increase of the bias with the number of transmissivity data and an increasing uncertainty with the number of head data. This is interpreted as a consequence of the use of an inadequate multiGaussian stochastic model that is not able to reproduce the connectivity of the original field.

The study makes use of polynomial chaos expansions to compute Sobol' indices within the frame of a global sensitivity analysis of hydrodispersive parameters in a simplified vertical cross-section of a segment of the subsurface of the Paris Basin. Applying conservative ranges, the uncertainty in 78 input variables is propagated upon the mean lifetime expectancy of water molecules departing from a specific location within a highly confining layer situated in the middle of the model domain. Lifetime expectancy is a hydrogeological performance measure pertinent to safety analysis with respect to subsurface contaminants, such as radionuclides. The sensitivity analysis indicates that the variability in the mean lifetime expectancy can be sufficiently explained by the uncertainty in the petrofacies, i.e. the sets of porosity and hydraulic conductivity, of only a few layers of the model. The obtained results provide guidance regarding the uncertainty modeling in future investigations employing detailed numerical models of the subsurface * Corresponding author, e-mail: konakli@ibk.baug.ethz.ch of the Paris Basin. Moreover, the study demonstrates the high efficiency of sparse polynomial chaos expansions in computing Sobol' indices for high-dimensional models.

No abstract