Présentations Colloques

Oral Presentation
8.10
Session 8.10: Recent developments in groundwater modeling and mathematical tools in Hydrogeology
Lotti
Geostatistics and numerical modelling in the characterization of contaminated sites
Groundwater characterisation in a contaminated site requires wells and piezometers to collect and analyse samples. By means of this information, the lateral and vertical extensions of the contaminant plume are depicted, constituting the starting point of any risk assessment or remediation procedures. This task requires all available knowledge on the site, in the effort to clarify the conceptual site model. Sometimes the pieces of the puzzle do not fit together, are contradictory or simply missing, especially when heterogeneous and or fractured aquifers are involved. Complexity of the data analysis is increased by the fact that contaminant plumes in groundwater are usually highly heterogeneous, anisotropic and non-stationary. Different geostatistical approaches have been proposed to define the optimal method for plume estimation. Nevertheless, even in the condition of the best plume estimate, the geostatistical analysis represent a “snapshot” of the plume not embedding any information about fate and transport of contaminant. This issue implies to extend the analysis through the inverse modelling approach- measured heads and concentrations in space and time are used as targets to estimate physical parameters, such as hydraulic conductivity, porosity and dispersivity. Most of the time the modelling phase come separately from any geostatistical approach, though any of the modelled parameters comes evidently with its spatial distribution. Incorporating different approaches (e.g. fate-and-transport models and geostatistical analyses) can greatly improve the ability to describe the shape, extent and temporal change of groundwater contaminant plumes. In addition, inverse modelling can handle multiple calibrated versions that satisfy geostatistical, historical data, and expert knowledge constraints. The whole process is necessarily completed by the quantification of geostatistical and model uncertainty to reasonably inform the decision makers. Explicative case studies are presented.
Italy