Présentations Colloques

Oral Presentation
8.10
Session 8.10: Recent developments in groundwater modeling and mathematical tools in Hydrogeology
Giese Markus
Karst conduit geometry- reduction of ambuigity by using a multiple signal approach for numerical hybrid models
Karst aquifers exhibit complex flow patterns as a result of the large contrast in hydraulic parameters between the conduits and the permeable fractured matrix. Especially the presence of highly conductive karst conduit systems embedded in the fractured matrix with substantial storage poses a challenge for the assessment and prediction of water resources. In mixed flow karst systems flow pattern are dominated by conduit flow. The differentiation and consideration of flow patterns at different scales are required for the investigation of complex processes, for example dual flow, heat and or solute transport. Therefore distributed numerical models have major advantage above lumped parameter models. Normally the distribution of the karst conduit network is unknown and its implementation in distributed numerical models is difficult. Hybrid numerical modeling approaches such as MODFLOW-2005 Conduit Flow Process Mode1 (CFPM1), which couples a discrete conduit network to a groundwater continuum, allows the analysis of the effect of different conduit networks on the evolution of large-scale flow patterns.**The application of pattern matching combined with inverse calibration of multiple signals (e.g. drawdown of a large scale pumping test, 1D- heat solute transport) can reduce the ambiguity resulting from the unknown location of the highly conductive features. The study demonstrates the implementation of different karst network geometries generated with a random walk approach. The geometry is supported by (hydro)geological information and covers a wide range of different networks (from single conduit up to meshed conduit network). Each realization can be classified in terms of hydraulic (e.g. conduit diameter, roughness, anisotropy) and geometrical properties (e.g. ‘Strahler’-approach). Finally, the stepwise addition of processes and objective functions helps to reduce the number of possible network realizations and uncertainty of the parameter set.**The approach is validated by high resolution measurements of multiple signals during a pumping test at the mixed flow karst regime of Cent Fonts (Languedoc, France).**
Germany