Architecture of seepage zones combined with their residence time to constrain hydrogeological models

Abstract

<p>Hydrological predictions for ungauged basins at catchment and regional scales are still challenged by the lack of available data. Under the assumption that the perennial stream network is mostly fed by groundwaters, its spatial extent is controlled by the magnitude of the subsurface hydraulic conductivity (<em>K</em>) with respect to the actual recharge rate (<em>R</em>). In addition, the residence time of groundwater is directly controlled by the storage capacity of the aquifer system, i.e. the porosity (<em>Ө</em>). Here we propose a new inversion approach that jointly considers the spatial organization of observed hydrographic network and the residence times of groundwater measured at springs to infer the geometry of the aquifer system and its hydraulic properties.</p> <p>We used a dataset gathered in an alpine catchment observatory (Natural conservation area of the Massif of Saint-Barthélemy, Pyrenees, France). The extent of the stream network has been mapped using field observation. Residence times have been obtained from concentrations of dissolved CFCs and SF<sub>6</sub> gases measured at 6 spring locations distributed over the catchment. The average transit time is about 30 years for perennial springs with a significant variability across the watershed. The relatively high residence time is also confirmed by high Helium concentrations.</p> <p>In our inversion scheme, we evaluate the accuracy of an ensemble of 3D hydrogeological models with different aquifer geometries and hydraulic properties. We found that topography and aquifer compartmentalization, through the decreasing trend in hydraulic conductivity, are key parameters in setting the spatial pattern of seepage areas and the distribution of transit times across the catchment. In addition, by running transient simulations of the model ensemble we further explore the accuracy of the models by comparing results with measurements of stream discharge and the intermittency of the hydrographic network. We found that intermittence seems to be connected to high transmissive shallow flow structures with low storage capacities (mostly organized within shallow soils and rockslides). However, perennial springs are sustained by deep groundwater flow paths within the bedrock. In perspective, we discuss the potential evolution of the extent, discharge magnitude and the transit time of seeping groundwater under changing recharge scenarios.</p>