The semi-confined Ledo-Paniselian (Eocene) aquifer in Flanders is recharged in the areas with a higher topography, where it is covered by the Bartonian clay [1]. Recharge is thus occurring by downward groundwater flow through the Bartonian clay. This is demonstrated by piezometric levels. Flow modeling in the recharge area of Ursel, where many piezometers provide an excellent knowledge of the hydraulic heads, has indicated a vertical hydraulic conductivity for the Bartonian clay of 10(-9) m/s. However, laboratory measurements often provide values which are at least one order of magnitude lower. This discrepancy can be ascribed to the presence of preferential pathways in the clay, through which the flow is preferentially taking place. The considered Tertiary sediments have all been deposited in a marine environment. At the end of the Tertiary, the sea regressed from the area and the present-day topography developed due to fluvial erosion. Groundwater movement was induced, by recharge of precipitation in the higher regions. Calcite dissolved in the infiltrating water, and freshening of the sediments took place, consisting both in dilution of the marine pore solution and in cation exchange, the latter mainly in the Bartonian clay in the recharge areas [2]. The early formed groundwaters were pushed to the north, in the direction of groundwater movement in the aquifer, as more precipitation was infiltrating in the recharge areas. In the aquifer, in an upstream direction, progressively more freshened watertypes are found. The chromatographic sequence of cation exchange is expressed by the subsequent surplus of Na+, followed by K+ and finally Mg2+, resulting in NaHCO3 and MgHCO3 watertypes. The groundwater leaking out of the clay and entering the aquifer nowadays in the recharge area, contains only the Ca2+ cation in appreciable quantities. A cored boring in the recharge area at Ursel has provided the opportunity for determining CEC and exchangeable cations, both of the clay and of the underlying aquifer [3]. The Bartonian clay at Ursel shows a very clear depletion in adsorbed Na+, and to a lesser extent also in K+, when compared to a boring out of the recharge area. Adsorbed Ca2+ is higher at Ursel. However, the analyzed samples of the Bartonian clay at Ursel do not at all show a depletion in adsorbed Mg2+. This is a surprising result, as the Ledo-Paniselian aquifer in this recharge area is containing CaHCO3 water, indicating that the flushing of the overlying clay layer has practically been accomplished [4]. The geochemical/mixing cell model PHREEQM [5] has been used to simulate the freshening of the Bartonian clay [4] and the subsequent recharge to the underlying aquifer [6]. After the calculated flushing of the Bartonian clay by around 100 pore volumes, the distribution of the exchangeable cations in the clay is similar to the one found in the boring at Ursel. At that stage the pore water solution which is leaving the Bartonian contains Ca2+ and Mg2+ as main cations and no significant Na+ content. This water is strongly different from the one found at present in the considered recharge area of the Ledo-Paniselian aquifer, in which Mg2+ is almost absent. When the different solutions leaking out of the clay till this stage are used to flush the aquifer, it is not possible to model the observed distribution of the groundwater quality in the Ledo-Paniselian. The clay must be flushed near 400 times to obtain the present situation in the aquifer. This confirms the existence of preferential paths in the Bartonian clay, where the flow is faster. The flow line in the analyzed boring must be part of a slower pathway and the freshening here is still in progress.
