SCK•CEN Day of the PhD's location:Mol date:6 October 2011
The small-scale spatial variability in hydraulic conductivity (K) is of high importance since it determines contaminant dispersion occurring as a result of flow field heterogeneity. In the current large-scale contaminant transport models with homogeneous K fields applied to the disposal site in Dessel, such small-scale variability in K is not incorporated. Accurate characterisation of the spatial structure of K and its incorporation in transport models will give indications whether the use of a macroscopic dispersion tensor in the homogeneous model is a sufficient approximation to accurately solve mass transport in the aquifer.
In this study, all secondary information (i.e. data that may be more easily obtained and used to estimate K) is transformed into K predictions. For CPT data, a literature-based hydraulic conductivity estimation methodology is used. Literature methods are also used to interpret dissipation tests. Air permeability measurements were converted to K values using validated equations from literature, and grain size data from two 90-m deep boreholes are converted to K values using a site-specific ensemble artificial neural network model. Since these different methods provide information on different spatial scales (i.e. cm to km scale), multi-scale variography was performed to characterise the spatial variability based on two-point statistics.