The Gorleben salt dome, a potential repository site for all types of radioactive waste, is located near the community of Gorleben in the north-eastern part of Lower Saxony. The investigation of the suitability of the site has been suspended due to a moratorium by the Federal Government and the electricity utilities. Before this moratorium, hydrogeological investigations were conducted in an area of more than 380km2 around the salt dome to study the aquifer system in the sediments above it. During the extensive field investigations and data evaluation, the existence of heterogeneities on various scales has been verified. These heterogeneities mainly reflect differences in lithology. On scales of decameters to kilometers, the site is characterized by a predominance of aquifers with intercalated aquitards. The salt dome is crossed by a subglacial erosion channel in which the lowermost aquifer is partly in contact with the cap rock or the salt itself. The density of the groundwater, which is a function of salt concentration, varies spatially. It generally increases with depth, often up to that of saturated brine at considerable depth. The change in groundwater density has to be considered together with spatially variable hydrogeological parameters because it has a pronounced influence on groundwater movement and contaminant transport.The groundwater modelling was carried out stepwise. Three-dimensional, porous media, freshwater models represent the overall hydrogeological structure and neglect the small-scale heterogeneities. These models are used to evaluate regional groundwater movement. They provide the characteristic features of the main flow paths under freshwater conditions. Preliminary uncertainty and sensitivity analyses have been performed using such freshwater models. Two-dimensional, porous media, variable salinity models are used to calculate changes in the flow field, density distribution, flow paths and travel times compared to freshwater simulations. They have demonstrated the sensitivity of the salt/fresh water system to major changes in the hydrogeological structure, as well as other important factors like initial density distribution and the period of time being modelled. Comparison of measured and calculated hydrochemical data shows the importance and validity of incorporating hydrogeological structures in these flow and transport models. Simple three-dimensional, variable salinity models representing only the main hydrogeological structures on a large scale confirm the importance of the third dimension for the flow field. They show that the flow of the salt water corresponds to that derived from density data analyses. Recent model development will soon enable more sophisticated three-dimensional sensitivity studies that take into account the spatial variability of the hydrogeological parameters and the variability in groundwater density.