Recharge mechanisms in fractured crystalline-rock aquifers

People involved:Joaquín JIMENEZ-MARTINEZ (CNRS), Laurent LONGUEVERGNE (CNRS), Tanguy LE BORGNE (UR1), Pascal GODERNIAUX (UR1), Philippe DAVY (CNRS), Olivier BOUR (UR1), Pierre GAVRILENKO (CNRS), Yves MEHEUST (UR1), Nicolas Lavenant (CNRS), Christophe PETTON (CNRS)

Published articles:

Jiménez-Martínez, J., Longuevergne, L., Le Borgne, T., Davy, P., Russian, A., and Bour, O. (2013), Temporal and spatial scaling of hydraulic response to recharge in fractured aquifers: Insights from a frequency domain analysis, Water Resources Research, 49(5), 3007-3023.

A frequency domain analysis to characterize heterogeneity and recharge mechanisms

We investigate aquifer behavior and recharge mechanisms in fractured media using a frequency domain approach. The main objectives are the quantification of aquifer physical characteristics (time response, storage capacity, transmissivity), and estimation of heterogeneity and connectivity impact on frequency domain response of piezometric level, in a range of temporal scales from 1 day up to a few years. Transfer Functions are calculated for Ploemeur experimental site (S Brittany, France). Recharge, first calculated as effective rainfall and groundwater level fluctuations are used as input and response functions, respectively. Classical behavior models for interpreting transfer functions are the linear reservoir and Dupuit model, and the combination with fast flow component. Some of the transfer functions computed can be reproduced successfully by linear reservoir model; on the contrary, some of them do not follow this scaling behavior. This suggests that the heterogeneity at different scales involves a variety of transfer processes that cannot be represented by classical models. Heterogeneity is constrained by analyzing the variability of the response (characteristic time, amplitude at low-frequency and asymptotic log-log slope) for monitoring wells. We discuss the relevance of scale effects on hydraulic properties by comparison with hydraulic property estimates. 


Figure .(a) Effective rainfall and groundwater level fluctuations (h) at monitoring well F07 as input and output functions, respectively. (b) Empiric TF (grey line) and regularized TF (black dots) for monitoring well F07 as function of frequency f.

Figure.Empirical regularized TFs (symbols) and linear reservoir model fits at monitoring wells: F07, and MF2. Linear regression fits and confidence intervals (95%).



Figure.Storage coefficient (a) and transmissivity (b) estimates from field techniques (single-borehole flowmeter, cross-borehole flowmeter and long-term pumping tests by Le Borgne et al. [2006a]), from ground deformation by GPS measurements [Moreau et al., 2006], and frequency domain analysis. Plots are organized according to increasing observation scale from left to right. An aquifer length L=2700 m has been used to compute transmissivity values.


Unsaturated zone monitoring

People involved: Joaquin JIMENEZ-MARTINEZ (CNRS), Elise COULON (UR1), Laurent LONGUEVERGNE (CNRS), Olivier Bour (UR1), Tanguy LE BORGNE (UR1), Nicolas LAVENANT (CNRS), Christophe PETTON (CNRS)

A complete monitoring of the unsaturated zone, including soil and weathered rock, was carried out at Ploemeur experimental site. The monitoring includes the installation at different depths (0.15, 0.25, 0.5, 0.9, 1.4, 2 m depth) of different type of sensors: TDRs (time domain reflectrometry) to measure soil water content; tensiometers to measure pressure head; and thermometers.


Figure.Detail of the TDR sensor (left), tensiometer (middle), and thermometer (right).

Figure.Field experiment photo and scheme of the sensors configuration.

Commonly, for crystalline-rocks aquifers two recharge components are considered: regional recharge through a wide network of fractures and direct infiltration through the upper soil and weathered rock zone. The main objective of the current experience is to quantify the second mentioned component, in order to reduce the uncertainty about aquifer recharge in this type of aquifers. The installation of thermometers will allow using the temperature as passive tracer, since the area is protected as a drinking water pumping site.


Hydro-mechanical modeling

People involved:Laurent LONGUEVERGNE (CNRS), Pierre GAVRILENKO (CNRS), Clément ROQUES (UR1)


Observation activities

1.       Ploemeur Hydrogeological observatory

Several actions have been carried out. For example, numerical modeling studies underlined the high sensitivity of tilt observation to local structures. As a consequence, detailed tilt-meter position and environment have been determined for integration in models. Moreover, several interesting deformation signals associated with Ploemeur pumping station maintenance have been observed and will be further interpreted in the upcoming year.

2.       Saint Brice en Coglès experiment

Saint Brice en Coglès can be considered as equivalent of Ploemeur hydrogeological observatory but at a pre-development stage. Indeed, RHAPSODI project highlighted here very high water yields in fractured context. Geologically speaking, permeable structures are located at the contact zone between granite and micaschist. Hydrogeologically speaking, the site is a natural discharge area for a deep fractured aquifer system (see figure below). We benefited from a 3-month pumping experiment within the framework of CASPAR project to set up a geodetic monitoring and observe gravity and ground deformation associated with pumping. Main interest is related here to the relationship between deep and sub-surface aquifer system. Large vertical deformations (~1 cm) have been observed in relation with the pumping experiment. Interestingly, the largest deformation is not located at the pumping well (F3), but follows a North-South structure fitting exactly with the mapped active fracture. Vertical deformation might be an interesting tool for mapping active underground structures.

Figure.Left: Interpreted structure and behavior of the deep fractured system in ambient (natural) conditions. Right: Surface vertical deformation after 1 month of pumping. Maximum deformation associated with pumping is 1 cm.

Modeling activities

Accurate interpretation of deformation patterns as the impact of underground permeable structures requires modeling tools focusing on poroelasticity. Pierre Gavrilenko has been developing a 3D code to model both pressure (water flow) and associated elastic deformation in heterogeneous context.

Main difficulties are related to

1.       High heterogeneity: modeling a wide range of spatial structures. Specific averaging methods are required to consider the impact of sub-grid permeability structures.

2.       High sensitivity of tilt deformation to local structures, requiring a thorough investigation of high permeable contrasts (well behavior for example).

Results are encouraging and we will begin parametric studies to identify the interest of tilt and vertical displacement measurements to recover underground permeable structures, before application to real data.

Figure. Left: Permeability field, main fracture (black rectangle) and pumping well (red). Fracture is 4 orders of magnitude more permeable as compared to rock. Middle: Pressure field (vertical cut) after 2h pumping with a rate of 1000m3/day. Right: Associated vertical deformation (max=7 mm). Specific storage is 2. 10-5 m-1.. Note the dissymmetry associated with the fracture.