A double medium approach to simulate groundwater level variations in a fissured karst aquifer

Karst aquifers are known for their complex flow mechanisms produced by the high spatial heterogeneity of void distribution. This study aims at confirming and enhancing the understanding of these flow mechanisms in a fissured karst aquifer using flow simulation with a double medium approach. Groundwater level (GWL) variations of the studied site (a 17 km2 topographic hill located in Burgundy, France) are monitored hourly through 18 boreholes, 3 karstic resurgences and the outlet river discharge of the catchment basin (40 km2). The Gerke and Van Genuchten double porosity model is implemented in the METIS code and used to simulate GWL variations with a matrix and a fracture medium. Model parameters are calibrated by fitting a simulated fracture medium GWL to measured GWL for four boreholes that are representative of the topographic hill hydrodynamics. The double medium approach succeeds in reproducing measured fast GWL responses, after recharge events that are not reproduced by a single medium approach. Investigating the model parameters' impact on simulated GWL variations and the exchange flow between the two media allows the conceptualization of the physical structure and the hydrodynamic mechanisms of the fissured karst aquifer. The spatial organization of parameters is observed and indicates areas of higher fracture density producing rapid GWL responses in contrast to lower fracture density areas, which produce slower and higher GWL responses. The fracture medium hydraulic parameters and the exchange parameter controlling the matrix/fracture interface properties are the most influential parameters for the simulated fracture medium GWL and the exchange flow between the matrix and the fracture medium during a recharge event. Calibrated fracture medium hydraulic conductivity and effective porosity vary respectively from 6.3E to 05 m.s−1 and 0.3% to 3.2E-04 m.s−1 and 2.4%, and the exchange parameter varies from 2.9E-10 m−1.s−1 to 6.5E-10 m−1.s−1. Matrix medium hydraulic conductivity variations have little impact on the simulation results while matrix medium effective porosity affects slightly simulated GWL, values vary from 11% to 17%. Finally, this study shows the Gerke and Van Genuchten double porous medium validity for the simulation of groundwater level in a fissured karst aquifer.